research paper decision making

• Browse All Articles
• Newsletter Sign-Up

DecisionMaking →

No results found in working knowledge.

• Were any results found in one of the other content buckets on the left?
• Try removing some search filters.
• Use different search filters.

The effects of visualization on judgment and decision-making: a systematic literature review

• Open access
• Published: 25 August 2021
• Volume 73 , pages 167–214, ( 2023 )

Cite this article

You have full access to this open access article

• Karin Eberhard   ORCID: orcid.org/0000-0001-6676-8889 1  

28k Accesses

14 Citations

Explore all metrics

The visualization of information is a widely used tool to improve comprehension and, ultimately, decision-making in strategic management decisions as well as in a diverse array of other domains. Across social science research, many findings have supported this rationale. However, empirical results vary significantly in terms of the variables and mechanisms studied as well as their resulting conclusion. Despite the ubiquity of information visualization with modern software, there is little effort to create a comprehensive understanding of the powers and limitations of its use. The purpose of this article is therefore to review, systematize, and integrate extant research on the effects of information visualization on decision-making and to provide a future research agenda with a particular focus on the context of strategic management decisions. The study shows that information visualization can improve decision quality as well as speed, with more mixed effects on other variables, for instance, decision confidence. Several moderators such as user and task characteristics have been investigated as part of this interaction, along with cognitive aspects as mediating processes. The article presents integrative insights based on research spanning multiple domains across the social and information sciences and provides impulses for prospective applications in the realm of managerial decision-making.

Similar content being viewed by others

What is Qualitative in Qualitative Research

Patrik Aspers & Ugo Corte

Literature reviews as independent studies: guidelines for academic practice

Sascha Kraus, Matthias Breier, … João J. Ferreira

Research Methodology: An Introduction

Avoid common mistakes on your manuscript.

1 Introduction

A visualization is defined as a visual representation of information or concepts designed to effectively communicate the content or message (Padilla et al. 2018 ) and improve understanding in the audience (Alhadad 2018 ). This representation can manifest in a range of imagery, from quantitative graphs (Tang et al. 2014 ) to qualitative diagrams (Yildiz and Boehme 2017 ), to abstract visual metaphors (Eppler and Aeschimann 2009 ) or artistic imagery. Visualization design may also intend to promote a specific behavior in the audience (Correll and Gleicher 2014 ). The visualization of information is associated with effective communication in terms of clarity (Suwa and Tversky 2002 ), speed (Perdana et al. 2018 ), and the understanding of complex concepts (Wang et al. 2017 ). Research shows, for example, that visualized risk data require less cognitive effort in interpretation than textual alternatives and are therefore comprehended more easily (Smerecnik et al. 2010 ), and complex sentiment data visualized in a scatterplot improve the accuracy in law enforcement decisions compared to raw data (Cassenti et al. 2019 ).

Visual experiences are the dominant sensory input for cognitive reasoning in everyday life, business, and science (Gooding 2006 ). As Davis ( 1986 ) points out, image creation and perception are part of the “unique and quintessential competencies of homo sapiens sapiens”. Hence, the visualization of information is an integral research subject in the domains of cognitive psychology, education (Alfred and Kraemer 2017 ), management (Tang et al. 2014 ) including financial reporting, strategic management, and controlling, marketing (Hutchinson et al. 2010 ), as well as information science (Correll and Gleicher 2014 ).

Management researchers study visualizations from a business perspective. First, the field of financial reporting considers the effect of financial graphs on investor perception (Beattie and Jones 2008 ; Pennington and Tuttle 2009 ). Second, the potential consequences of visualizations on decision-making are examined in the area of managerial decision support, with a focus on judgments based on quantitative data such as financial decisions (Tang et al. 2014 ) and performance controlling (Ballard 2020 ). Finally, a small number of works investigate more complex decision-making based on qualitative, multivariate, and relational information (Platts and Tan 2004 ). Altogether visualizations fulfill a variety of functions, from focusing attention to sharing thoughts to identifying data structures, trends, and patterns (Platts and Tan 2004 ).

The vast majority of existing research in visualization, however, arises from the two domains of information science and cognitive psychology. Information science research on how to design visualizations for effective user cognition stretches back almost one century (Washburne 1927 ). While early research focuses on comparing tables and simple graphs, newer research on human–computer interfaces covers advanced data visualizations facilitated by computing power (Conati et al. 2014 ). For example, interactive visualization software enables users to manipulate data directly. While promising in terms of analytic capability, the potential for biases and overconfidence is suggested as a downside (Ajayi 2014 ). Equally, cognitive psychology research notes that visual information may be superior over verbal alternatives in certain cognitive tasks since they can be encoded in their original form, where spatial and relational data is preserved. Thereby, visual input is inherently richer than verbal and symbolic information, which is automatically reductionistic (Meyer 1991 ), but more suited for discrete information retrieval due to its simplicity (Vessey and Galletta 1991 ). However, the processes behind visual cognition remain largely unclear (Vila and Gomez 2016 ).

Despite the ubiquity of visualizations in research and practice, there is no comprehensive understanding of the potential and limits of information visualization for decision-making. Although at times converging, insights from research of different areas are seldom synthesized (Padilla et al. 2018 ), and there has been no effort for a systematic review or overarching framework (Zabukovec and Jaklič 2015 ). However, a synthesis of existing research is essential and timely due to three reasons. First, information visualization is ubiquitous both in the scientific and business community, yet there are conflicting findings on its powers and limits in support of judgment and decision-making. Second, cognitive psychology research provides several promising suggestions to explain observable effects of visualizations, yet these are rarely integrated into research in other domains, including strategic decision-making. Third, the barriers to using information visualization software have fallen to a minimum, making it available to a wide range of producers and users. This raises the issue of the validity of positive effects for various task and user configurations. The goal of this paper is therefore to provide an overview of the fragmented existing research on visualizations across the social and information sciences and generate insights and a timely research agenda for its applicability to strategic management decisions.

My study advances visualization research on three paths. First, I establish a framework to summarize the numerous effects and variable interactions surrounding the use of visualizations. Second, I conduct a systematic literature review across the social and information sciences and summarize and discuss this plethora of findings along with the aforementioned structure. Third, I utilize this work as a basis for identifying and debating gaps in existing research and resulting potential avenues for future research, with a focus on the area of strategic management decisions.

The structure of the article is as follows. The next chapter briefly describes the research field, followed by the methodology of my literature search. Next, I analyze the results of my search and discuss common insights. In the ensuing chapter, I develop an agenda for management research by building on particularly relevant ideas with conflicting or incomplete evidence. Finally, I conclude my review and discuss contributions and implications for practice.

2 Definition of the research field

2.1 definition of key terms.

Information visualizations support the exploration, judgment, and communication of ideas and messages (Yildiz and Boehme 2017 ). The term “graph” is often used as a synonym for information visualization in general (Meyer 1991 ) as well as describing quantitative data presentation specifically (Washburne 1927 ). As my review exhibits, these graphs constitute the prevalent form of information visualization. Common quantitative visualizations are line and bar charts, often showcasing a development over time and regularly used in financial reporting (Cardoso et al. 2018 ) and controlling (Hutchinson et al. 2010 ). In scientific literature, probabilistic charts such as scatterplots, boxplots, and probability distribution charts (Allen et al. 2014 ) frequently depict risk and uncertainty. More specialized charts include decision trees to depict conditional logic (Subramanian et al. 1992 ), radar charts to display complex multivariate information (Peebles 2008 ), or cluster charts and perceptual maps for marketing decision support (Cornelius et al. 2010 ).

Despite the breadth of existing visualization research, its application to strategic decisions is narrow and there is an abundance of research limited to elementary tasks and choices. To provide a clear distinction, I focus my search on decisions, judgments, and inferential reasoning as more advanced forms of cognitive processing. Decision-making can be broadly defined as choosing between several alternative courses of action (Padilla et al. 2018 ). On the other hand, reasoning and judgment refer to the evaluation of a set of alternatives (Reani et al. 2019 ), without actions necessarily being attached as for decision-making. Such efforts are cognitively demanding and complex when compared to more elementary tasks, such as a choice between options (Tuttle and Kershaw 1998 ), and include the rigorous evaluation of alternatives across a range of attributes, which is characteristic for strategic decisions (Bajracharya et al. 2014 ). For this reason, I include studies that examine the influence of visualizations on some form of decision or judgment outcome. Mason and Mitroff ( 1981 ) highlight that strategic decisions, in management and elsewhere, involve complex and ambiguous information environments. Information visualization may relate to decision quality in this context since one critical factor in the effectiveness of strategic decisions is the objective and comprehensive acquisition and analysis of relevant information to define and evaluate alternatives (Dean and Sharfman 1996 ).

2.2 Perspectives in literature

Visualization research exists within a range of domains in the social and information sciences, which reflects the diversity of the empirical application. I identify psychology (cognitive and educational), management (financial reporting, strategic management decisions, and controlling), marketing, and information science as the primary areas of research. This heterogeneity in terms of application area provides the first dimension in my literature review. Second, I classify existing studies along the type of variable interaction they primarily investigate. Based on the framework first introduced by DeSanctis ( 1984 ), I hereby differentiate four categories: Works principally focused on (1) the effects of visualizations on comprehension and decisions as dependent variables provide the basis of all research. This relationship is then investigated through: (2) User characteristics as moderators; (3) task and format characteristics as moderators; and (4) cognitive processing as mediator. An overview of this classification, including the prevalence of extant findings across domains, is given in Fig.  1 .

Visualization research structured by domain and variables primarily investigated

First, the investigation of visualization effects on decisions and judgments is established across all research areas mentioned, and primarily studies outcome variables such as decision accuracy (Sen and Boe 1991 ), speed (Falschlunger et al. 2015a ), and confidence (Correll and Gleicher 2014 ). While these studies contribute examples for graphs influencing observable decision effectiveness and efficiency across a range of contexts, they do not investigate moderating or mediating factors.

Second, psychology research pushes this investigation further towards including moderating effects of user characteristics , such as domain expertise and training (Hegarty 2013 ), and measures of cognitive ability such as numeracy (Honda et al. 2015 ) or literacy (Okan et al. 2018a ). The relevance of these moderating factors is validated both in studies focusing on cognition as well as experiments in educational research, for example by providing evidence that the quality of a judgment made based on a graph may depend more on the user than the format itself (Mayer and Gallini 1990 ).

Similarly, human–computer interface research spearheads further insights into moderating factors of task and format characteristics, such as task type (Porat et al. 2009 ), task complexity (Meyer et al. 1997 ), data structure (Meyer et al. 1999 ), and the graphical saliency of features (Fabrikant et al. 2010 ) through rigorous user testing. At the same time, Vessey ( 1991 ) developed the theory of cognitive fit as a concept bridging cognitive and information systems research, stating that positive effects of graphs depend on a fit between task type and format type, differentiating between symbolic and spatial archetypes.

Finally, cognitive psychology research aims at explaining the observable effects of visualization in terms of mediating cognitive mechanisms . Here, cognitive load theory provides the foundation, stating that an individual’s working memory capacity is limited, and performance in a task or judgment depends on the cognitive load they experience while assessing information. According to this logic, cognitive load that is too high damages performance (Chandler and Sweller 1991 ). Reducing cognitive load by providing visualizations in complex environments is therefore often stated as a key goal of graph design (Smerecnik et al. 2010 ).

Importantly, the boundaries between these variable categories are fluid. Many studies investigate more than one relationship and the inclusion of moderating variables has become common. Various application areas covering these interdependencies attest to the heterogeneous nature of visualization research. However, previous reviews highlight that insights are seldom shared across fields and call for the integration of findings into new studies (Padilla et al. 2018 ). In particular, strategic management research does not yet follow such a holistic approach.

3 Method of literature search

3.1 search design.

The methodological basis of this paper is a systematic literature search as a means to collect and evaluate the existing findings in a systematic, transparent, and reproducible way on the specified topic (Fisch and Block 2018 ) in order to produce a more complete and objective knowledge presentation than in traditional reviews (Clark et al. 2021 ). I conduct a keyword search on the online search engines EBSCOhost and ProQuest, limited to English-language works that have been peer-reviewed, in order to ensure the quality of the sources. Gusenbauer and Haddaway ( 2020 ) identify both search engines as principal academic search systems as they fulfill all essential performance requirements for systematic reviews. On EBSCOhost, I use the databases Business Source Premier , Education Research Complete , EconLit , APA PsycInfo , APA PsycArticles , and OpenDissertations to search for empirical works; on ProQuest, I use the databases British Periodicals , International Bibliography of the Social Sciences (IBSS) , Periodicals Archive Online , and Periodicals Index Online with a filter on articles to cover the social sciences comprehensively. The keyword used is the concatenated term “(visualization OR graph OR chart) AND (decision OR judgment OR reasoning)”, searched for in abstracts. Footnote 1 The terms were chosen as “visualization” is commonly used as a category name for visualized information (Brodlie et al. 2012 ), and the “graph” is the focus of traditional visualization research (Vessey 1991 ). The term “chart” is a synonym for both quantitative and qualitative graphs which has seen increasing use particularly in the 2000s (Semmler and Brewer 2002 ). The terms “judgment OR decision OR reasoning” were added to ensure that studies examining observable outcomes of visualization use, as opposed to cognitive processes such as comprehension only, were highlighted. After a review of the evolution of visualization research over time, I focus my search to articles published from the year 1990 in order to capture the recent advancements covering modern modes of information visualization. Footnote 2 This search results in 1658 articles combined, after removing duplicates 1505 articles remain.

Next, I review all article abstracts based on the three content criteria defined in the following. I include all articles rooted in the (1) social sciences or information sciences , where the focus of the study lies on (2) how a visualization per se or a variation within related visualizations affects a user's or audience's decision or judgment in a given task , and the topic is studied through (3) original empirical works. Most articles are excluded in this process and 116 studies remain due to the prevalence of graphs as auxiliary means, not the subject of research, in various domains, particularly in medical research. I repeat this exclusion process by reading the full texts of all articles and narrow down the selection further to 81 papers.

Building on this systematic search, I conducted a supplementary search through citation and reference tracking, as well as supplementary search engines, such as JSTOR (Gusenbauer and Haddaway 2020 ). Footnote 3 This includes gray literature such as conference proceedings or dissertations, which lie outside of traditional academic publishing. In addition, I limit the inclusion of gray literature to studies by researchers included in my systematic search and completed within the last 10 years in order to gather a comprehensive and up-to-date overview of the findings of working groups particularly relevant to visualization research. Thereby I identify 52 additional articles, resulting in a total of 133 articles included.

3.2 Limitations of search

Due to the plethora of existing literature mentioning the topic of visualization in various contexts and degrees of quality, I subject my search to well-defined limitations. First, I only include peer-reviewed articles in my systematic search. These are studies that have been thoroughly validated and represent the major theories within a field (Podsakoff et al. 2005 ). However, I incorporate gray literature of comparable quality as part of my additional exploratory search.

Second, I limit the search to information and social sciences to deliberately omit results from the broad areas of medicine and natural sciences. In these, various specific concepts are visualized as a means within research, yet not investigating the visualization itself. For the same reason, I only apply the search terms to article abstracts, since the terms “graph” and “chart” in particular will result in a high number of results when searched for in the full text, due to the common use of graphs in presenting concepts and results.

Third, I only include original empirical work in order to enable the synthesis and critical validation of empirical findings across research areas. At the same time, I acknowledge the existence of several highly relevant theoretical works, which inform my search design and structure while being excluded from the systematic literature search and analysis.

4.1 Overview of results

I identify a total of 133 articles, published between 1990 and 2020. Interest in visualization research gained initial momentum in the early 1990s (Fig.  2 ). More recently, the number of studies rises starting around 2008, with the continued publication of five to ten papers per year since and a visible peak in interest around 2014/15. A significant share of recent works stems from the information science literature, and the wealth of publications around 2014 coincides with the advent of mainstream interest in big data (Arunachalam et al. 2018 ), which is closely linked to information visualization for subsequent analysis and decision-making (Keahey 2013 ). In addition, a cluster of publications by one group of authors (Falschlunger et al. 2014 , 2015a , c, b) in the financial reporting domain enhances the observed peak in publications, which is therefore not indicative of a larger trend. Instead, the continued wealth of publications in the last decade shows the contemporary relevance of and interest in visualization research.

Articles included in systematic search by publication year and area of research

Next to the information sciences, the largest share of the studies identified originates in cognitive psychology research. Furthermore, management literature discusses visualization and graphs continuously throughout the last three decades, with notable peaks in interest around the year 2000 in the domain of annual reporting (Beattie and Jones 2000 , 2002a , b ; Arunachalam et al. 2002 ; Amer 2005 ; Xu 2005 ) and internal management reporting with classic bar and line graphs around the year 2015 (Falschlunger et al. 2014 , 2015a , c ; Tang et al. 2014 ; Hirsch et al. 2015 ; Zabukovec and Jaklič 2015 ). Consumer research in marketing constitutes a further domain regularly discussing visualizations and their effect on decisions and judgment (Symmank 2019 ), albeit to a smaller extent. This heterogeneity in research areas is reflected by the journals identified in my search, where the 133 articles spread across 83 different journals, complemented by ten studies from conference proceedings and three papers included in doctoral dissertations (Table 1 ). Apart from the articles in conference proceedings added through the supplementary exploratory search, the studies were published in journals with a SCIMAGO Journal Rank indicator ranging from 0.253 (Informing Science) to 8.916 (Journal of Consumer Research). All but four journals received Q1 and Q2 ratings, which equals the top half of all SCImago rated journals. The h-index ranges from 6 (Journal of Education for Library and Information Science) to 332 (PLoS ONE) (Scimago Lab 2021 ).

In the 133 articles identified, experiments are by far the most common method for data collection, with 113 (85%) of studies conducting a total of 182 controlled experiments with over 28,000 participants (Fig.  3 ). In addition, I find seven instances of archival research covering over 600 companies, six instances of surveys with almost 1000 participants in total, four quasi experiments, two natural experiments, and one field experiment to complete the picture.

Articles included in systematic search by methodology

Of the 182 experiments conducted, the majority works with students as subjects (125 or 69%). The largest remaining share investigates a sample of the general (online) population (32 or 18%) and only 13% study the effect of visualization with practitioners in their respective domain (24). In contrast, four out of the six surveys were conducted with practitioners that were addressed explicitly. Besides, one survey each was conducted with students and subjects from the general population.

Following the advice by Fisch and Block ( 2018 ), I categorize the results from literature in a concept-centric manner, based on the primary variable interaction investigated. I further distinguish by the four application domains and seven subdomains discussed and present a structured overview at the end of each subchapter. The independent variable in all cases is the use of a visual representation designed for a specific use case, either as opposed to non-visual representation methods such as verbal descriptions [e.g. Vessey and Galletta ( 1991 )], or traditional visualizations that the research aims to improve on [e.g. Dull and Tegarden ( 1999 )].

4.2 Effects of visualizations on decisions and judgments

4.2.1 judgment/decision accuracy.

The most common dependent variable investigated in visualization research is the accuracy of the subjects on a given comprehension, judgment, or decision task. Most studies are in psychology research, with positive effects dominating. In cognitive psychology, experiments show that well-designed visualizations can improve problem comprehension (Chandler and Sweller 1991 ; Huang and Eades 2005 ; Nadav-Greenberg et al. 2008 ; Okan et al. 2018b ). For example, Dong and Hayes ( 2012 ) show in their experiment with 22 practitioners that a decision support system visualizing uncertainty improves the identification and understanding of ambiguous decision situations. Likewise, visualizations improve decision (Pfaff et al. 2013 ) and judgment accuracy (Semmler and Brewer 2002 ; Tak et al. 2015 ; Wu et al. 2017 ) and improve the quality of inferences made from data (Sato et al. 2019 ). Findings in educational psychology support this claim. In teaching, visual materials improve understanding and retention (Dori and Belcher 2005 ; Brusilovsky et al. 2010 ; Binder et al. 2015 ; Chen et al. 2018 ) in students, and support the judgment accuracy of educators when analyzing learning progress quantitatively (Lefebre et al. 2008 ; Van Norman et al. 2013 ; Géryk 2017 ; Nelson et al. 2017 ). Furthermore, Yoon’s longitudinal classroom intervention (2011) using social network graphs enables students to make more reflected and information-driven strategic decisions. However, other studies arrive at more mixed or opposing findings. In their experiment, Rebotier et al. ( 2003 ) find that visual cues do not improve judgment accuracy over verbal cues in imagery processing. Other experiments even demonstrate verbal information to be superior over graphs in comprehension (Parrott et al. 2005 ) as well as judgment accuracy (Sanfey and Hastie 1998 ). Some graphs appear unsuitable for specific content, such as bar graphs depicting probabilities (Newman and Scholl 2012 ) and bubble charts encoding information in circle area size (Raidvee et al. 2020 ). In addition, more complex charts like boxplots, histograms (Lem et al. 2013 ), and tree charts (Bruckmaier et al. 2019 ) appear less effective for the accurate interpretation of statistical data in some experiments, presumably as they elicit errors and confusion in insufficiently trained students.

Studies in management and business research arrive at further, more pessimistic results. While Dull and Tegarden ( 1999 ) find in their experiment with students that three-dimensional visuals can improve the prediction accuracy in financial reporting contexts, and Yildiz and Boehme ( 2017 ) observe in their practitioner survey that a graphical model of a corporate security decision problem improves risk perception when compared to a textual description, most other studies present a less positive picture. Several studies do not find graphs superior over tables in financial judgments (Chan 2001 ; Tang et al. 2014 ; Volkov and Laing 2012 ), and in consumer research (Artacho-Ramírez et al. 2008 ). In financial reporting, a dedicated school of research investigates the effect of distorted graphs lowering financial judgment accuracy (Arunachalam et al. 2002 ; Beattie and Jones 2002a , b ; Amer 2005 ; Xu 2005 ; Pennington and Tuttle 2009 ; Falschlunger et al. 2014 ), irrespective of whether the distortion is intended by the designer. Chandar et al. ( 2012 ) elaborate on the positive effect of the introduction of graphs and statistics in performance management for AT&T in the 1920s, but more recent case study examples are rare.

By contrast, several experimental studies from human–computer interaction research largely contribute evidence for a positive effect. Targeted visual designs lead to higher judgment accuracy in specific tasks (Subramanian et al. 1992 ; Butavicius and Lee 2007 ; Van der Linden et al. 2014 ; Perdana et al. 2018 ) and improve decision-making (Peng et al. 2019 ). For example, probabilistic gradient plots and violin plots enable higher accuracy in statistical inference judgments in the online experiment by Correll and Gleicher ( 2014 ) than traditional bar charts. However, experiments by Sen and Boe ( 1991 ) and Hutchinson et al. ( 2010 ) equally lack a significant effect on data-based decision-making quality. Amer and Ravindran ( 2010 ) find a potential for visual illusions degrading judgment accuracy similar to results from financial reporting, and McBride and Caldara ( 2013 ) find that visuals lower accuracy in law enforcement judgments when compared to raw data presentation (Table 2 ).

4.2.2 Response time

The next most common outcome variable investigated in visualization research is response time , often referred to as efficiency. Across the board, experimenters observe that information visualization lowers response time in various judgment and decision tasks. In psychology, this includes decision-making in complex information environments (Sun et al. 2016 ; Géryk 2017 ). The opposite effect emerges from only one study, where Pfaff et al. ( 2013 ) find that a decision support system visualizing complex uncertainty information requires a longer time to use than one omitting this graphical information. In management research, Falschlunger et al. ( 2015a ) find that visually optimized financial reports can speed up judgment both for students and practitioners. Studies originating in information science validate this picture, observing that well-designed visualizations reduce response time in quantitative (Perdana et al. 2018 ) as well as geospatial judgment tasks (MacEachren 1992 ). Furthermore, McBride and Caldara ( 2013 ) observe that students in their experiments arrive at faster judgments when provided with a network graph as opposed to a table (Table 3 ).

4.2.3 Decision confidence

Next to these directly observable metrics, experimenters regularly elicit measures of decision confidence in visualization research based on subjects’ self-assessment. From a cognitive psychology perspective, Andrade ( 2011 ) finds that subjects display excessive confidence in estimates based on visualizations, which biases subsequent decision-making. On the other hand, Dong and Hayes ( 2012 ) show that a visual decision support system depicting uncertainty in engineering design leads to marginally lower decision confidence, compared to traditional methods omitting uncertainty information. In management research, Tang et al. ( 2014 ) present an increase in confidence in the context of financial decision-making, and Yildiz and Böhme (2017) find in their practitioner survey that an appealing visual increases decision confidence in a managerial setting without changing the actual decision outcome. Similarly, further experiments in information science provide evidence for increased confidence with a link to increased judgment accuracy (Butavicius and Lee 2007 ) or without (Sen and Boe 1991 ; Wesslen et al. 2019 ). In the context of uncertainty, Arshad et al. ( 2015 ) once again report novice subjects having lower confidence in the use of graphs with uncertainty visualized, however, this effect does not occur for practitioners (Table 4 ).

4.2.4 Prevalence of biases

Several studies investigate the prevalence of biases by searching for distinct patterns of deviations in judgment and decision accuracy with largely mixed results. Through a total of seven cognitive psychology experiments, Sun et al. ( 2010 , 2016 ) and Radley et al. ( 2018 ) find that varying scale proportions in graphs change the resulting decision-making since data points are evaluated in a cognitively biased manner based on their distance to other chart elements. Furthermore, Padilla et al. ( 2015 ) demonstrate that uncertainty is understood to a disparate extent when it is encoded through spatial glyphs, color, or brightness. In human–computer interaction research, experiments observe similar framing biases through salient graphical features (Diamond and Lerch 1992 ) such as color schemes (Klockow-McClain et al. 2020 ). Lawrence and O’Connor ( 1993 ) also show that graph scaling affects judgment and relate this to the anchoring heuristic. Finally, financial reporting research extensively dedicates its field of impression management on the observation that such biases are prevalent and possibly intended in annual report graphics, including through distorted graph axes (Falschlunger et al. 2015b ) and an intentional selection of information to visualize (Beattie and Jones 1992 , 2000 ; Dilla and Janvrin 2010 ; Jones 2011 ; Cho et al. 2012a , b ). Two further experiments compare the prevalence of cognitive biases with graphs compared to text directly and find no difference for the recency bias in financial reporting (Hellmann et al. 2017 ) as well as for other heuristics in data-based managerial decision-making (Hutchinson et al. 2010 ) (Table 5 ).

4.2.5 Attitude change and willingness to act

Observations on attitude change and the willingness to act on information constitute the final category of outcome variables found in visualization research. Cognitive psychology research observes an effect of visualizations on risk attitude, where salient graphs can either enhance risk aversion (Dambacher et al. 2016 ) or risk-seeking (Okan et al. 2018b ), depending on the information that is highlighted most saliently. Similarly, varied financial graphs change investors’ risk perception and subsequent investment recommendations (Diacon and Hasseldine 2007 ). In the area of performance management, the visualization of KPIs motivates managers’ intention to act on the information when compared to text (Ballard 2020 ). Consumer research investigates such phenomena commonly, where brand attitude and the intention to purchase a product represent specific cases of judgment and decision-making. Miniard et al. ( 1991 ) were among the first to show that different pictures can result in different attitudes, while Gkiouzepas and Hogg ( 2011 ) extend this investigation to visual metaphors. Finally, information science research provides further insights. King Jr et al. (1991) find that visualizations are more persuasive in attitude change than text, and Perdana et al. ( 2018 ) increase student subjects’ willingness to invest in their experimental setting through visualization software. On the other hand, Phillips et al. ( 2014 ) find their subjects to be less willing to seek out additional information in ambiguous decision settings (Table 6 ).

4.3 User characteristics as moderating variables

4.3.1 expertise and training.

Common moderating variables investigated both in psychological and information science research are the users’ expertise or training experience in a given domain. Experimenters widely encounter a positive impact of experience on the influence of visualizations on judgment accuracy and efficiency. In cognitive psychology, Hilton et al. ( 2017 ) find that graphs of statistical risk improve decision quality for more experienced practitioners alone. On the other hand, some results from educational psychology point towards the opposite effect of experience. Mayer and Gallini ( 1990 ) find in their student experiments that learners with higher pre-test performance benefit less from visual aids than learners on a lower level. In the information sciences, Conati et al. ( 2014 ) find in their testing of computer interfaces that experience with visualizations leads to a pronounced advantage in judgment accuracy. Training sessions (Raschke and Steinbart 2008 ) and experience through task repetition (Meyer 2000 ) enhance the positive effects of graphs (Table 7 ).

4.3.2 Cognitive ability

Another user characteristic regularly investigated in the social sciences is the measurement of cognitive ability . In psychology studies, Honda et al. ( 2015 ) and Cardoso et al. ( 2018 ) find that reflective ability determines in part how well subjects translate visualizations into accurate judgments. Visual working memory (Tintarev and Masthoff 2016 ) and numeracy (Honda et al. 2015 ) are further traits related to cognitive ability in dealing with visualizations and found to enhance the benefits of visualizations on judgment effectiveness and efficiency. The only study presenting contrary results consists of three experiments by Okan et al. ( 2018a ), where subjects with higher graph literacy are more prone to specific biases when shown bar graphs of health risk data, and thereby make less accurate judgments. On the other hand, experiments in financial reporting (Cardoso et al. 2018 ) confirm the positive effect of the reflective ability. Conati and Maclaren ( 2008 ) and Conati et al. ( 2014 ) extend this idea to perceptual speed in the area of consumer research (Table 8 ).

4.3.3 User preferences

Finally, experimenters investigate user preferences at times. In the adjacent field of musical education, for example, Korenman and Peynircioglu ( 2007 ) demonstrate that the visual presentation of learning material is only helpful to students with the respective learning style. In cognitive psychology, Daron et al. ( 2015 ) observe a variation in user preferences when presented with visualization options, however without a significant effect on decision performance. This result is replicated in an online survey on human–computer interaction by Lorenz et al. ( 2015 ). O’Keefe and Pitt ( 1991 ) operationalize cognitive style from the MBTI framework and find a weak association with the subjects’ reported preferences for text or specific chart types. However, no relation to actual judgment accuracy or efficiency is found (Table 9 ).

4.4 Task and format characteristics as moderating variables

4.4.1 task type.

One common task characteristic identified as a moderating variable is the task type , originally defined in the information sciences. In her seminal theoretical paper, Vessey ( 1991 ) identifies spatial and symbolic tasks as the two archetypes, which correspond to spatial and symbolic types of cognitive processing and spatial (graphical) and symbolic (textual/numerical) representations. She hypothesizes that visualizations improve judgment effectiveness where these three manifestations align, which she defines as cognitive fit and validates through experiments (Vessey and Galletta 1991 ), including in the sphere of multiattribute management decisions (Umanath and Vessey 1994 ). Further research in information science widely supports this moderating effect by comparing tables and standard quantitative graphs in judgment tasks of increasing complexity (Coll et al. 1994 ; Tuttle and Kershaw 1998 ; Speier 2006 ; Porat et al. 2009 ). On the other hand, experiments in managerial forecasting (Carey and White 1991 ) and financial reporting (Hirsch et al. 2015 ) present the effectiveness of graphical displays in spatial decisions, based on cognitive fit theory. Fischer et al. ( 2005 ) provide further evidence from the domain of cognitive psychology, showing that bar graphs support spatial-numerical judgments particularly well when the chart orientation equals the cognitive processing by following a left-to-right direction (Table 10 ).

4.4.2 Level of data structure

I identify two other task characteristics investigated in the literature, albeit infrequently. First, the level of data structure has been investigated only once in the information science domain. Meyer et al. ( 1999 ) find line charts superior over tables in judgment tasks when the underlying data is structured, with the opposite effect for unstructured data (Table 11 ).

4.4.3 Task complexity

Second, two further experiments observe task complexity as a moderating effect. Meyer et al. ( 1997 ) demonstrate that the speed advantage they find for tables over bar graphs in their computer interface tasks becomes more pronounced with increasing task complexity. However, the same effect does not occur for line graphs. On the other hand, Falschlunger et al. ( 2015c ) find task complexity to be the main factor in predicting task efficiency and effectiveness in handling financial reports but do not observe interaction effects with the visualization (Table 12 ).

4.4.4 Graphical saliency of relevant data

Finally, various studies investigate modifications in the graph format as a variable, with a focus on the graphical saliency of relevant data . This area of research is bridging the two domains of cognitive psychology and information science with widely overlapping results. For example, Verovszek et al. (2013) observe in their information science experiment that colored visualizations are less effective in supporting laypeople’s judgments on urban planning than simple black-and-white line drawings since colorful, irrelevant features distract from the core information. Van den Berg et al. ( 2007 ) identify color as a more powerful feature to highlight salient information in graphs than other variables, such as size. Spence et al. ( 1999 ) find that variations in brightness lead to faster response times in comparison tasks than variations in color. Breslow et al ( 2009 ) demonstrate that the moderating effect of the use of color on judgment speed depends on the task type, with multicolored visuals ideal for identification tasks and black-and-white brightness scales preferable for comparison tasks. Finally, MacEachran et al. (2012) find colorless suited to represent uncertainty when compared to features such as fuzziness or transparency in their surveys with students and practitioners.

Next to color, three-dimensional depth cues have received attention in research. Several psychology experiments find that three-dimensional depth cues irrelevant to the information visualized lower judgment accuracy (Zacks et al. 1998 ; Edwards et al. 2012 ) as well as speed (Fischer 2000 ). Negative effects occur equally for other irrelevant visual cues lowering the saliency of actually relevant information (Fischer 2000 ). Further studies show that increasing the saliency of relevant features can enhance the tendency to make compensatory choices (Dilla and Steinbart 2005 ) and shorten response time (Fabrikant et al. 2010 ), while visual clutter decreases judgment accuracy and boosts response times (Ognjanovic et al. 2019 ). Several other studies test the suitability of a specific set of graphs for unique judgment areas such as uncertainty simulation in urban development (Aerts et al. 2003 ), risk communication (Stone et al. 2017 ; Stone et al. 2018 ), and performance management (Peebles 2008 ) (Table 13 ).

4.5 Cognitive aspects as mediating variables

4.5.1 cognitive load.

Cognitive psychology research introduces the idea of cognitive processes mediating the influence of visualizations on judgment performance, with a focus on cognitive load . Jolicœur and Dell’Acqua ( 1999 ) show in their experiment that the perception of visualizations is subject to structural constraints in working memory capacity, and Allen et al. ( 2014 ) manipulate cognitive load as a dependent variable to demonstrate that judgment accuracy and speed using visualizations decrease under higher cognitive load. Subsequently, psychology experiments provide evidence that visualizations improve decision performance by reducing cognitive load as a mediating factor, operationalized and measured either through pupil size and dilation (Smerecnik et al. 2010 ; Toker and Conati 2017 ) or self-reported load (Cassenti et al. 2019 ). In management research, Ajayi ( 2014 ) investigates this relationship in the context of a proprietary visualization tool for financial data but finds no effect of the visualization component on cognitive load or judgment accuracy. Two further experiments in human–computer interface research operationalize cognitive load based on subjective reporting (Anderson et al. 2011 ) and performance in a secondary task (Block 2013 ) and demonstrate that cognitive load mediates the relationship between visualization use and judgment accuracy and speed, with some types of graphics better suited than others (Table 14 ).

4.5.2 Gazing behavior

Another concept frequently operationalized to represent working memory capacity is gazing behavior , which more recent experiments observe through the use of eye-tracking technology, pioneered by the information sciences. Reani et al. ( 2019 ) observe in their experiment with 49 students that gazing behavior is associated with judgment accuracy, where subjects that pay more attention to relevant visual areas deliver more accurate answers. Similarly, Lohse ( 1997 ) finds that in the more complex decision environment of a budget allocation simulation, decision accuracy is related to efficient gazing behavior and can be improved through the use of colors to reduce the time subjects spend looking at the chart legend. Psychology experiments validate that well-designed graphs enable subjects to focus their attention on relevant information and subsequently improve decision accuracy (Huestegge and Pötzsch 2018 ) and response time (Vila and Gomez 2016 ) (Table 15 ).

4.5.3 Attention

Another variable operationalized at times in eye-tracking experiments is attention, which is elicited through metrics such as the average gazing duration on a specific visual element (Pieters et al. 2010 ). In their cognitive psychology experiment, Smerecnik et al. ( 2010 ) observe that graphs attract more attention in risk communication compared to tables and text and are associated with more accurate judgments. Applying this idea to consumer research, Pieters et al. ( 2010 ) study the consumer’s attention towards visual advertisements and observe that visual complexity based on features such as decorative color can hurt attention, while well-structured complexity such as arrangements of relevant information enhances attention and the attitude toward the brand (Table 16 ).

4.5.4 Affect

Finally, some research emerges into the potential mediating role of affect . Harrison ( 2013 ) shows in her large-scale online experiment that affective priming can significantly influence judgment accuracy in tasks supported visually and that the graphs themselves can cause a change in affect valence. Similarly, Plass et al. ( 2014 ) demonstrate in their educational research that color and shape in visualizations can evoke positive affect and are associated with better student learning (Table 17 ).

5 Discussion

In this paper, I have presented a systematic and integrative review of the current state of research on the effect of information visualization in the social and information sciences. I structured and summarized the results of my systematic literature review along the type of variable interactions present in experimental research. In order to discuss and synthesize the variety of literature insights, I categorize them into three groups: Descriptions of the positive effects for visualizations within decision-making, elaborations on moderators of this potential, and insights into negative effects of misguided visualization use. Table 18 highlights this categorization of results by application domain.

5.1 Positive Effect 1: Information visualization improves decision accuracy and quality

Research findings overwhelmingly confirm the hypothesis that visualizations enable the user to comprehend information more effectively, subsequently improving performance in judgments and decisions. The reason behind this effect is most commonly attributed to cognitive mechanisms. Suwa and Tversky ( 2002 ) point out that based on cognitive load theory, less working memory is needed when visuals provide external representations of concepts, which one can easily refer back to and thereby need not keep in mind, leading to improved judgments. Allen et al. ( 2014 ) show in their experiment that under externally induced cognitive load, well-designed charts suffer less than cluttered ones. Furthermore, graphs enable a simpler gazing pattern than text, which can be used as an indicator of cognitive effort (Smerecnik et al. 2010 ). Based on the concept of cognitive load reduction, visualizations are effectively used in various application areas including management research (Falschlunger et al. 2014 ) and more specifically managerial decision-making (Yildiz and Boehme 2017 ), next to psychology and information sciences more broadly.

5.2 Positive Effect 2: Information visualization steers attention towards uncertainty

A large share of studies identified points towards the strength of visualizations in enhancing uncertainty and risk features in a data set. Beyond increasing the awareness of uncertainty (Dong and Hayes 2012 ), the question of whether visualizations can also improve the reasoning with probabilistic information is studied extensively. Various studies show that visualizations can reduce typical comprehension issues, resulting in the more accurate use of probabilities from a statistical perspective (Allen et al. 2014 ; Wu et al. 2017 ; Stone et al. 2018 ). Positive effects in risk understanding are evaluated particularly in the contexts of safety, such as food safety (Honda et al. 2015 ) and violence risk (Hilton et al. 2017 ). Studies investigating the cognitive processes more closely provide evidence that simpler charts indeed perform best (Edwards et al. 2012 ) since they can reduce cognitive load (Anderson et al. 2011 ) and ultimately improve the internal processing of probabilistic models (Tak et al. 2015 ). As Quattrone ( 2017 ) points out, ambiguity and uncertainty are inherent in managerial decision-making and should be embraced by information visualization, but research on this insight in management is scarce.

5.3 Positive Effect 3: Information Visualization Speeds Up Cognitive Processing

There is evidence that graphs lead to faster processing, learning, and decision-making (Block 2013 ), as judgment and decision efficiency are measured and operationalized as the response time in various experiments. Utilizing eye-tracking technology, Reani et al. ( 2019 ) point out that different types of graphs result in varying gazing patterns in users and hypothesize a link to the reasoning processes. Based on the principle of saliency, multiple studies show that graphs optimally designed to focus attention on the most relevant information lead to more efficient and thereby faster gazing (Falschlunger et al. 2014 , 2015a ), since more time can be spent focusing on highly relevant information (Vila and Gomez 2016 ). Much of this existing work stems from the area of management reporting, investigating quantitative financial data. Overall, the evidence for visual aids speeding up cognitive processing and decision-making appears robust and applicable to management research.

5.4 Moderator 1: The effects of visualization depend on cognitive fit within the decision context

Cognitive fit is a moderator in the effectiveness of visualizations that has been well validated across psychological, management, and information science. Introducing cognitive fit theory, Vessey ( 1991 ) explains many existing research findings in the graph versus table literature claiming that graphs are not (always) more effective, most notably by DeSanctis ( 1984 ). Cognitive fit theory is validated widely (Vessey and Galletta 1991 ; Carey and White 1991 ; Coll et al. 1994 ; Meyer et al. 1997 ; Meyer 2000 ; Porat et al. 2009 ; Perdana et al. 2019 ). Padilla (2018) recognizes that this well-documented effect arises because a cognitive mismatch between data, task, and approach (format) requires more working memory, which negatively affects cognitive processing effectiveness and efficiency. Though highly reliable, many studies investigate elementary processing tasks with limited external validity for more complex decision-making in practice. Umanath and Vessey ( 1994 ) and others (Tuttle and Kershaw 1998 ; Hirsch et al. 2015 ) extend the original cognitive fit theory and successfully apply it to multi-attribute judgments—though at a potential time-accuracy tradeoff. Finally, the idea of matching task and format complexity can be seen as an extension to cognitive fit theory, where graphs are only helpful when they represent as much data complexity as necessary to complete the respective task, but as little as possible (Pieters et al. 2010 ; Van der Linden et al. 2014 ; Géryk 2017 ).

5.5 Moderator 2: Differences within users can be more relevant than the visualization design

Task complexity in relation to user ability needs to be strictly controlled for as a moderator of positive visualization effects. Early studies including individual differences hypothesize that graph potential may be limited to users with a high level of ability (Subramanian et al. 1992 ). Other studies claim that the positive effects of visualizations may be more significant for (McIntire et al. 2014 ) or even limited to (Mayer and Gallini 1990 ) less-skilled individuals. However, these seemingly conflicting results can be explained by the idea that since graphs are effective by requiring less working memory than other formats, improvements are only visible where working memory capacity is limited and needed elsewhere (Lohse 1997 ).

Furthermore, the majority of studies including user factors emphasize the importance of training and expertise, as opposed to inherent ability. Various studies support the claim that experience significantly enhances the contribution of visuals (Porat et al. 2009 ; Edwards et al. 2012 ; Falschlunger et al. 2015a ; Ognjanovic et al. 2019 ), with some claiming that training constitutes a requirement (Géryk 2017 ; Hilton et al. 2017 ) or that users without training are subject to stronger biases (Raschke and Steinbart 2008 ). Consequently, the training factor needs to be closely monitored particularly for a novel or complex visualization. However, extensive training of users is frequently time-consuming and costly. Therefore, the imperative arises for interactive visualization interfaces to accommodate for varying user needs in demanding decision situations. Interactive data visualization software is shown to improve investment decisions (Perdana et al. 2018 ) and judgments by reducing cognitive load (Ajayi 2014 ), for example with flexible performance management dashboards that reduce information load while hosting a full set of KPIs (Yigitbasioglu and Velcu 2012 ). Contrary to much of the early research on static visualizations, the progress in interactivity studies has been driven by practice and case studies, with calls for science to follow suit (Marchak 1994 ; McInerny et al. 2014 ). Overall, I conclude that a match in ability and training with format complexity and novelty, respectively, is a significant determinant of the effectiveness of visualizations. However, there has been little to no empirical research on the subject in the domain of management.

5.6 Negative Effect 1: Visualizations May Not Always Be Helpful: Risk to Impair Decision Making by Misguiding Attention

Several studies, including in management research, argue that visualizations misguide attention even in the presence of cognitive and user fit. For example, Hutchinson (2010) finds graphs to be as exposed to cognitive biases as tables in data-based managerial decision-making. Similarly, other studies identify graphical representations as equally or less effective than verbal formats in financial reports (Volkov and Laing 2012 ), forecasting (Chan 2001 ), probabilistic comprehension (Parrott et al. 2005 ), evidence evaluation (Sanfey and Hastie 1998 ), and communication (Rose 1966 ). The common denominator in these studies is the suboptimal use of salient visual elements, leading to distraction. For example, overly realistic visualizations encompassing color and higher complexity (DeSanctis 1984 ), may lead to visual clutter that decreases performance (Alhadad 2018 ). As Padilla et al. ( 2018 ) argue, visualizations are powerful because they attract fast cognitive bottom-up processing. However, when this superficial processing is focused on irrelevant elements, decision quality can suffer. A well-studied example of this effect is the addition of superfluous three-dimensional cues to quantitative graphs, which lowers accuracy in using the graph (Zacks et al. 1998 ; Fischer 2000 ).

5.7 Negative Effect 2: Visualizations can increase decision-maker overconfidence

The most documented cognitive bias in my review is overconfidence, which can be aggravated by the use of visualizations (O’Keefe and Pitt 1991 ). Multiple studies demonstrate that graph use can increase decision confidence without enhancing decision quality to the same extent in the context of management and finance (Tang et al. 2014 ; Yildiz and Boehme 2017 ; Wesslen et al. 2019 ). This may result from the perception that visualizations show more information at once (Miettinen 2014 ), thereby seemingly requiring less search for additional information (Phillips et al. 2014 ). In particular, this can be the case when graphs appear to visually simplify a problem and the decision-maker fails to adjust his confidence to the underlying complexity (Sen and Boe 1991 ). There is some research with inconclusive results (Pfaff et al. 2013 ), showing no difference in confidence (Hirsch et al. 2015 ) or even lowered confidence (Dong and Hayes 2012 ; Arshad et al. 2015 ). However, the majority of these studies deal with uncertainty communication, which is inherently tied to a decrease in confidence (Watkins 2000 ). Overall, the evidence demonstrates that unless highlighting uncertainty, visual aids result in higher decision confidence. The case of overconfidence is particularly well established in the area of management controlling and financial reporting but understudied for strategic decisions.

6 Research agenda

In summary, there is ample evidence for the potential of information visualization to improve decision-making in terms of effectiveness and efficiency, yet my review highlights possible limitations and risks where its use is misguided or inappropriate. I argue that several of these are particularly critical for further research since there is little to no application to the domain of strategic management decisions, despite the ubiquity of visualizations to support these in practice. Based on the summary of my insights by application domain in Table 18, I identify five research gaps in the field of strategic management decisions.

First, there is conflicting evidence regarding the effect of information visualization on decision-making under uncertainty, and existing research is mostly limited to information science (Aerts et al. 2003 ). Depending on the context and design, visualization use can increase or reduce risk-taking (Dambacher et al. 2016 ) but has the potential to improve probabilistic reasoning in an objective manner (Allen et al. 2014 ). Given the importance of uncertainty as a defining factor of strategic management decisions (Quattrone 2017 ), the possibility of information visualizations to improve risk understanding in the management context deserves closer evaluation. For example, the framing bias is a well-documented phenomenon in strategic decision-making (Hodgkinson et al. 1999 ), leading to different subjective risk interpretations and subsequent decisions based on the presentation of information. Naturally, the question arises whether information visualization can mitigate this bias and which salient visual features are beneficial. I suggest exploring this question through experiments with strategic management decision vignettes.

Research Gap 1: How can information visualization mitigate the framing bias and improve risk understanding in strategic management decisions?

Second, my review has made clear that the effectiveness of information visualization depends in large parts on user characteristics such as expertise (Hilton et al. 2017 ), numeracy (Honda et al. 2015 ), and graph literacy (Okan et al. 2018b ), yet there exists no transfer of this insight towards individual managerial traits. At the same time, well-established concepts such as the Upper Echelons Theory (Hambrick 2007 ) highlight the relevance of CEO characteristics, both observable and psychological for strategic managerial choices and, subsequently, company performance. While some concepts such as experience may be transferrable from existing visualization research (Falschlunger et al. 2015c ) requiring validation only, others, such as group position or individual values, present opportunities to extend theory substantially. I suggest exploring this area through a dedicated analysis of relevant CEO characteristics and corresponding empirical research with practitioner subjects.

Research Gap 2: How do CEO characteristics influence the effectiveness of information visualization in strategic management decisions?

Third, while the prevalence of visualization use for impression management in financial reporting is well-established (Falschlunger et al. 2015b ), there is a complete lack of transfer of this phenomenon to the realm of strategic management decisions. As Whittington et al. ( 2016 ) highlight, strategy presentations can be seen as an effective tool for CEO impression management. Given the popularity of visualizations in this communication medium – both through quantitative charts and schematic diagrams (Zelazny 2001 ), the question arises to what degree impression management also takes place in this case, for example through the reporting bias (Beattie and Jones 2000 ). I suggest investigating this subject empirically, for example through archival studies.

Research Gap 3: To what extent does CEO impression management occur through visualization use in strategy presentations?

Fourth, while overconfidence in managerial decision-making is a commonly reported issue with significant efforts to develop corrective feedback as a remedy (Chen et al. 2015 ), there is little understanding of the role of information visualization in this matter. My review has demonstrated that visual aids often increase decision confidence as much as they improve the judgment itself (Yildiz and Boehme 2017 ) or even more (Sen and Boe 1991 ), but can also reduce confidence, particularly where uncertainty information is depicted (Dong and Hayes 2012 ). However, the latter effect was only studied for topics unrelated to management. Therefore, there is a complete lack of understanding of the effects of visualizations on managerial overconfidence, and I suggest exploring this research gap empirically with practitioners.

Research Gap 4: How do visual aids influence overconfidence in managerial decision-making?

Finally, a large share of cognitive psychology research discusses the effectiveness of visualization use through the reduction of cognitive load, yet they usually start off with low-load contexts, which is the opposite of high-stress managerial decision-making (Laamanen et al. 2018 ). Allen et al. ( 2014 ) find evidence that the effectiveness of distinct graph types changes with the level of externally induced cognitive load, raising the question to what extent previous insights on helpful visual aids are applicable to managerial decisions in a high-stakes environment filled with distractions and parallel issues requiring attention. Therefore, I suggest studying visualization use in experimental environments with varying levels of cognitive load as the independent variable, ideally with management practitioners and a realistic strategic task setting.

Research Gap 5: How does cognitive load influence the effectiveness of information visualization in strategic management decisions?

7 Conclusion

Information visualization has become ubiquitous in our daily professional and private lives, even more so with the advent of accessible and powerful computer graphics. However, the impact that visualizations have on human cognition and ultimately decisions stills remains unclear to a large extent. While the prevalence of visualization research across a plethora of application domains shows its pertinence, the decentralized approach has led to a scattered and unstructured field of theories and empirical evidence. My literature review thus sought to provide a far-reaching overview of this work and a detailed research agenda. As a result, three contributions arise from my review.

First, I provide an overarching structure to summarize the range of effects and interacting variables that can be found surrounding visualization research. This includes a wide set of dependent variables ranging from decision quality and speed to confidence and attitudes, as well as complex moderating and mediating effects that are crucial to understanding the overall power of visualizations. This precise framework is paramount to a holistic and comprehensive review of the scattered existing literature.

Second, to the best of my knowledge, my systematic literature review is the first on visualizations spanning the whole of social and information sciences simultaneously. While some previous reviews such as the one by Yigitbasioglu and Velcu ( 2012 ) utilize a multidisciplinary approach, they usually define the visualization type investigated more narrowly, for example by focusing on dashboards only. I believe that my integrative overview provides a valid contribution to the ongoing work to synthesize the mixed results in visualization research.

Third, I demonstrate that despite the plethora of evidence at first sight, visualization research is far from complete due to its multitude of moderating variables and at times conflicting results. Building on my systematic review of existing literature, I specify an agenda of potential research directions for future studies to follow in order to advance our understanding of the cognitive implications of visualizations in the context of managerial decision making in particular.

This paper also has direct implications for management practice. As Zhang ( 1998 ) points out, managerial decision-making is particularly well-positioned to profit from good visualizations since it often utilizes unstructured, large sets of information that are computer-centered, dynamic, and need to be interpreted constantly under time pressure. However, the interaction of visualization use with various factors should not be underestimated in the design of computer graphics for decision support. The high validity of the cognitive fit theory and the contingency on user characteristics found in the literature demonstrates that the designer should spend extensive time on clarifying for whom and what the visualization is intended. Furthermore, the potential for overconfidence and automatic processing based on visualized information may result in decision-makers skipping on more elaborate thought, which may be desirable in some, but certainly not all situations.

Availability of data and material

Not applicable.

Code availability

Thanks to the anonymous reviewer for encouraging me to extend my keyword search.

Thanks to the anonymous reviewer for this valuable impulse.

Thanks to the anonymous reviewer for pointing me towards additional, highly relevant articles.

Aerts JC, Clarke KC, Keuper AD (2003) Testing popular visualization techniques for representing model uncertainty. Cartogr Geogr Inf Sci 30:249–261. https://doi.org/10.1559/152304003100011180

Article   Google Scholar  

Ajayi O (2014) Interactive data visualization in accounting contexts: impact on user attitudes, information processing, and decision outcomes. University of Central Florida

Google Scholar  

Alfred KL, Kraemer DJ (2017) Verbal and visual cognition: Individual differences in the lab, in the brain, and in the classroom. Dev Neuropsychol 42:507–520. https://doi.org/10.1080/87565641.2017.1401075

Alhadad SSJ (2018) Visualizing data to support judgement, inference, and decision making in learning analytics: insights from cognitive psychology and visualization science. J Learn Anal 5:60–85. https://doi.org/10.18608/jla.2018.52.5

Allen PM, Edwards JA, Snyder FJ et al (2014) The effect of cognitive load on decision making with graphically displayed uncertainty information. Risk Anal 34:1495–1505. https://doi.org/10.1111/risa.12161

Amer TS (2005) Bias due to visual illusion in the graphical presentation of accounting information. J Inf Syst 19:1–18. https://doi.org/10.2308/jis.2005.19.1.1

Amer TS, Ravindran S (2010) The effect of visual illusions on the graphical display of information. J Inf Syst 24:23–42. https://doi.org/10.2308/jis.2010.24.1.23

Anderson EW, Potter KC, Matzen LE et al (2011) A user study of visualization effectiveness using EEG and cognitive load. Comput Graph Forum 30:791–800. https://doi.org/10.1111/j.1467-8659.2011.01928.x

Andrade EB (2011) Excessive confidence in visually-based estimates. Organ Behav Hum Decis Process 116:252–261. https://doi.org/10.1016/j.obhdp.2011.07.002

Arshad SZ, Zhou J, Bridon C et al (2015) Investigating user confidence for uncertainty presentation in predictive decision making. In: Proceedings of the annual meeting of the Australian special interest group for computer human interaction, pp 352–360

Artacho-Ramírez MA, Diego-Mas JA, Alcaide-Marzal J (2008) Influence of the mode of graphical representation on the perception of product aesthetic and emotional features: an exploratory study. Int J Ind Ergon 38:942–952. https://doi.org/10.1016/j.ergon.2008.02.020

Arunachalam V, Pei BKW, Steinbart PJ (2002) Impression management with graphs: effects on choices. J Inf Syst 16:183–202. https://doi.org/10.2308/jis.2002.16.2.183

Arunachalam D, Kumar N, Kawalek JP (2018) Understanding big data analytics capabilities in supply chain management: Unravelling the issues, challenges and implications for practice. Transp Res Part E Logist Transp Rev 114:416–436. https://doi.org/10.1016/j.tre.2017.04.001

Bajracharya S, Carenini G, Chen K et al (2014) Interactive visualization for group decision analysis. Int J Inf Technol Decis Mak 17:1839–1864. https://doi.org/10.1142/s0219622018500384

Ballard A (2020) Promoting performance information use through data visualization: evidence from an experiment. Public Perform Manag Rev 43:109–128. https://doi.org/10.1080/15309576.2019.1592763

Beattie V, Jones MJ (1992) The use and abuse of graphs in annual reports: theoretical framework and empirical study. Account Bus Res 22:291–303

Beattie VA, Jones MJ (2000) Changing graph use in corporate annual reports: a time-series analysis. Contemp Account Res 17:213–226. https://doi.org/10.1506/aat8-3cgl-3j94-ph4f

Beattie V, Jones MJ (2002a) Measurement distortion of graphs in corporate reports: an experimental study. Account Audit Account J. https://doi.org/10.1108/09513570210440595

Beattie V, Jones MJ (2002b) The impact of graph slope on rate of change judgments in corporate reports. Abacus 38:177–199. https://doi.org/10.1111/1467-6281.00104

Beattie V, Jones M (2008) Corporate reporting using graphs: a review and synthesis. J Account Lit 27:71–110

Binder K, Krauss S, Bruckmaier G (2015) Effects of visualizing statistical information—an empirical study on tree diagrams and 2 × 2 tables. Front Psychol. https://doi.org/10.3389/fpsyg.2015.01186

Block G (2013) Reducing cognitive load using adaptive uncertainty visualization. Nova Southeastern University

Breslow LA, Trafton JG, Ratwani RM (2009) A perceptual process approach to selecting color scales for complex visualizations. J Exp Psychol Appl 15:25–34. https://doi.org/10.1037/a0015085

Brodlie K, Osorio RA, Lopes A (2012) A review of uncertainty in data visualization. In: Expanding the frontiers of visual analytics and visualization. Springer, pp 81–109

Bruckmaier G, Binder K, Krauss S, Kufner H-M (2019) An eye-tracking study of statistical reasoning with tree diagrams and 2 × 2 tables. Front Psychol. https://doi.org/10.3389/fpsyg.2019.00632

Brusilovsky P, Ahn J, Rasmussen E (2010) Teaching Information Retrieval With Web-based Interactive Visualization. J Educ Libr Inf Sci 51:187–200

Butavicius MA, Lee MD (2007) An empirical evaluation of four data visualization techniques for displaying short news text similarities. Int J Hum-Comput Stud 65:931–944. https://doi.org/10.1016/j.ijhcs.2007.07.001

Cardoso RL, de Leite R, O, Aquino ACB de, (2018) The effect of cognitive reflection on the efficacy of impression management. Account Audit Account J 31:1668–1690. https://doi.org/10.1108/aaaj-10-2016-2731

Carey JM, White EM (1991) The effects of graphical versus numerical response on the accuracy of graph-based forecasts. J Manag 17:77. https://doi.org/10.1177/014920639101700106

Cassenti DN, Roy H, Kase SE (2019) Cognitive processing of visually presented data in decision making. Hum Factors 61:78–89. https://doi.org/10.1177/0018720818796009

Chan SY (2001) The use of graphs as decision aids in relation to information overload and managerial decision quality. J Inf Sci 27:417. https://doi.org/10.1177/016555150102700607

Chandar N, Collier D, Miranti P (2012) Graph standardization and management accounting at AT&T during the 1920s. Account Hist 17:35–62. https://doi.org/10.1177/1032373211424889

Chandler P, Sweller J (1991) Cognitive load theory and the format of instruction. Cogn Instr 8:293–332. https://doi.org/10.1207/s1532690xci0804_2

Chen G, Crossland C, Luo S (2015) Making the same mistake all over again: CEO overconfidence and corporate resistance to corrective feedback. Strateg Manag J 36:1513–1535. https://doi.org/10.1002/smj.2291

Chen J, Wang M, Grotzer TA, Dede C (2018) Using a three-dimensional thinking graph to support inquiry learning. J Res Sci Teach 55:1239–1263. https://doi.org/10.1002/tea.21450

Cho CH, Michelon G, Patten DM (2012a) Impression management in sustainability reports: an empirical investigation of the use of graphs. Account Public Interest 12:16–37

Cho CH, Michelon G, Patten DM (2012b) Enhancement and obfuscation through the use of graphs in sustainability reports. Sustain Account Manag Policy J 3:74–88. https://doi.org/10.1108/20408021211223561

Clark WR, Clark LA, Raffo DM, Williams RI (2021) Extending Fisch and Block’s (2018) tips for a systematic review in management and business literature. Manag Rev Q 71:215–231. https://doi.org/10.1007/s11301-020-00184-8

Coll RA, Coll JH, Thakur G (1994) Graphs and tables: a four-factor experiment. Commun ACM 37:77–86. https://doi.org/10.1145/175276.175283

Conati C, Carenini G, Hoque E et al (2014) Evaluating the impact of user characteristics and different layouts on an interactive visualization for decision making. Comput Graph Forum 33:371–380. https://doi.org/10.1111/cgf.12393

Conati C, Maclaren H (2008) Exploring the role of individual differences in information visualization, pp 199–206

Cornelius B, Wagner U, Natter M (2010) Managerial applicability of graphical formats to support positioning decisions. J Für Betriebswirtschaft 60:167–201. https://doi.org/10.1007/s11301-010-0061-y

Correll M, Gleicher M (2014) Error bars considered harmful: exploring alternate encodings for mean and error. IEEE Trans vis Comput Graph 20:2142–2151. https://doi.org/10.1109/tvcg.2014.2346298

Dambacher M, Haffke P, Groß D, Hübner R (2016) Graphs versus numbers: how information format affects risk aversion in gambling. Judgm Decis Mak 11:223–242

Daron JD, Lorenz S, Wolski P et al (2015) Interpreting climate data visualisations to inform adaptation decisions. Clim Risk Manag 10:17–26. https://doi.org/10.1016/j.crm.2015.06.007

Davis W (1986) The origins of image making. Curr Anthropol 27:193–215. https://doi.org/10.1086/203422

Dean JW, Sharfman MP (1996) Does decision process matter? A study of strategic decision-making effectiveness. Acad Manage J 39:368–392. https://doi.org/10.5465/256784

DeSanctis G (1984) Computer graphics as decision aids: directions for research. Decis Sci 15:463–487. https://doi.org/10.1111/j.1540-5915.1984.tb01236.x

Diacon S, Hasseldine J (2007) Framing effects and risk perception: the effect of prior performance presentation format on investment fund choice. J Econ Psychol 28:31–52

Diamond L, Lerch FJ (1992) Fading frames: data presentation and framing effects. Decis Sci 23:1050–1071. https://doi.org/10.1111/j.1540-5915.1992.tb00435.x

Dilla WN, Janvrin DJ (2010) Voluntary disclosure in annual reports: the association between magnitude and direction of change in corporate financial performance and graph use. Account Horiz 24:257–278. https://doi.org/10.2308/acch.2010.24.2.257

Dilla WN, Steinbart PJ (2005) Using information display characteristics to provide decision guidance in a choice task under conditions of strict uncertainty. J Inf Syst 19:29–55. https://doi.org/10.2308/jis.2005.19.2.29

Dong X, Hayes CC (2012) Uncertainty visualizations: helping decision makers become more aware of uncertainty and its implications. J Cogn Eng Decis Mak 6:30–56. https://doi.org/10.1177/1555343411432338

Dori YJ, Belcher J (2005) How does technology-enabled active learning affect undergraduate students’ understanding of electromagnetism concepts? J Learn Sci 14:243–279. https://doi.org/10.1207/s15327809jls1402_3

Dull RB, Tegarden DP (1999) A comparison of three visual representations of complex multidimensional accounting information. J Inf Syst 13:117. https://doi.org/10.2308/jis.1999.13.2.117

Edwards JA, Snyder FJ, Allen PM et al (2012) Decision making for risk management: a comparison of graphical methods for presenting quantitative uncertainty. Risk Anal Int J 32:2055–2070. https://doi.org/10.1111/j.1539-6924.2012.01839.x

Eppler MJ, Aeschimann M (2009) A systematic framework for risk visualization in risk management and communication. Risk Manage 11:67–89. https://doi.org/10.1057/rm.2009.4

Fabrikant SI, Hespanha SR, Hegarty M (2010) Cognitively inspired and perceptually salient graphic displays for efficient spatial inference making. Ann Assoc Am Geogr 100:13–29. https://doi.org/10.1080/00045600903362378

Falschlunger L, Eisl C, Losbichler H, Greil A (eds) (2014) Improving information perception of graphical displays – an experimental study on the display of column graphs. In: Proceedings from the 22th international conference in central europe on computer graphics, visualization and computer vision. Vaclav Skala - Union Agency

Falschlunger L, Eisl C, Losbichler H, Grabmann E (eds) (2015a) Report optimization using visual search strategies - an experimental study with eye tracking technology. In: 6th international conference on information visualization theory and applications

Falschlunger L, Eisl C, Losbichler H, Greil AM (2015b) Impression management in annual reports of the largest European companies. J Appl Account Res 16:383–399. https://doi.org/10.1108/jaar-10-2014-0109

Falschlunger L, Grabmann E et al (eds) (2015c) Deriving a holistic cognitive fit model for an optimal visualization of data for management decisions. Seville, Spain

Fisch C, Block J (2018) Six tips for your (systematic) literature review in business and management research. Manag Rev Q 68:103–106. https://doi.org/10.1007/s11301-018-0142-x

Fischer MH (2000) Do irrelevant depth cues affect the comprehension of bar graphs? Appl Cogn Psychol 14:151–162. https://doi.org/10.1002/(SICI)1099-0720(200003/04)14:2%3c151::AID-ACP629%3e3.0.CO;2-Z

Fischer MH, Dewulf N, Hill RL (2005) Designing bar graphs: orientation matters. Appl Cogn Psychol 19:953–962. https://doi.org/10.1002/acp.1105

Géryk J (2017) Visual analytics of educational time-dependent data using interactive dynamic visualization. Expert Syst Int J Knowl Eng Neural Netw. https://doi.org/10.1111/exsy.12175

Gkiouzepas L, Hogg MK (2011) Articulating a new framework for visual metaphors in advertising: a structural, conceptual, and pragmatic investigation. J Advert 40:103–120. https://doi.org/10.2753/joa0091-3367400107

Gooding DC (2006) Visual cognition: where cognition and culture meet. Philos Sci 73:688–698. https://doi.org/10.1086/518523

Gusenbauer M, Haddaway NR (2020) Which academic search systems are suitable for systematic reviews or meta-analyses? Evaluating retrieval qualities of Google Scholar, PubMed, and 26 other resources. Res Synth Methods 11:181–217. https://doi.org/10.1002/jrsm.1378

Hambrick DC (2007) Upper echelons theory: an update. Academy of Management Briarcliff Manor, NY, p 10510

Harrison L (2013) The role of emotion in visualization. Doctoral thesis, University of North Carolinab

Hegarty M (2013) Cognition, metacognition, and the design of maps. Curr Dir Psychol Sci 22:3–9. https://doi.org/10.1177/0963721412469395

Hellmann A, Yeow C, De Mello L (2017) The influence of textual presentation order and graphical presentation on the judgements of non-professional investors. Account Bus Res 47:455–470. https://doi.org/10.1080/00014788.2016.1271737

Hilton NZ, Ham E, Nunes KL et al (2017) Using graphs to improve violence risk communication. Crim Justice Behav 44:678–694. https://doi.org/10.1177/0093854816668916

Hirsch B, Seubert A, Sohn M (2015) Visualisation of data in management accounting reports. J Appl Account Res. https://doi.org/10.1108/jaar-08-2012-0059

Hodgkinson GP, Bown NJ, Maule AJ et al (1999) Breaking the frame: an analysis of strategic cognition and decision making under uncertainty. Strateg Manag J 20:977–985. https://doi.org/10.1002/(SICI)1097-0266(199910)20:10%3c977::AID-SMJ58%3e3.0.CO;2-X

Honda H, Ogawa M, Murakoshi T et al (2015) Effect of visual aids and individual differences of cognitive traits in judgments on food safety. Food Policy 55:33. https://doi.org/10.1016/j.foodpol.2015.05.010

Huang W, Eades P (2005) How people read graphs. Australian Computer Society Inc, London, pp 51–58

Huestegge L, Pötzsch TH (2018) Integration processes during frequency graph comprehension: performance and eye movements while processing tree maps versus pie charts. Appl Cogn Psychol 32:200–216. https://doi.org/10.1002/acp.3396

Hutchinson JW, Alba JW, Eisenstein EM (2010) Heuristics and biases in data-based decision making: effects of experience, training, and graphical data displays. J Mark Res 47:627–642. https://doi.org/10.1509/jmkr.47.4.627

Jolicœur P, Dell’Acqua R (1999) Attentional and structural constraints on visual encoding. Psychol Res 62:154–164. https://doi.org/10.1007/s004260050048

Jones MJ (2011) The nature, use and impression management of graphs in social and environmental accounting. Account Forum 35:75–89. https://doi.org/10.1016/j.accfor.2011.03.002

Keahey TA (2013) Using visualization to understand big data. IBM Soft Bus Anal Adv Visu

King WC Jr, Dent MM, Miles EW (1991) The persuasive effect of graphics in computer-mediated communication. Comput Hum Behav 7:269–279. https://doi.org/10.1016/0747-5632(91)90015-s

Klockow-McClain KE, McPherson RA, Thomas RP (2020) Cartographic design for improved decision making: trade-offs in uncertainty visualization for Tornado threats. Ann Am Assoc Geogr 110:314–333. https://doi.org/10.1080/24694452.2019.1602467

Korenman LM, Peynircioglu ZF (2007) Individual differences in learning and remembering music: auditory versus visual presentation. J Res Music Educ 55:48–64. https://doi.org/10.1177/002242940705500105

Laamanen T, Maula M, Kajanto M, Kunnas P (2018) The role of cognitive load in effective strategic issue management. Long Range Plann 51:625–639. https://doi.org/10.1016/j.lrp.2017.03.001

Scimago Lab (2021) SJR : scientific journal rankings. In: SJR Sci. J. Rank. https://www.scimagojr.com/journalrank.php . Accessed 11 Jun 2021

Lawrence M, O’Connor M (1993) Scale, variability, and the calibration of judgmental prediction intervals. Organ Behav Hum Decis Process 56:441. https://doi.org/10.1006/obhd.1993.1063

Lefebre E, Fabrizio M, Merbitz C (2008) Accuracy and efficiency of data interpretation: a comparison of data display methods. J Precis Teach Celeration 24:2–20

Lem S, Onghena P, Verschaffel L, Van Dooren W (2013) On the misinterpretation of histograms and box plots. Educ Psychol 33:155–174. https://doi.org/10.1080/01443410.2012.674006

Lohse GL (1997) The role of working memory on graphical information processing. Behav Inf Technol 16:297–308. https://doi.org/10.1080/014492997119707

Lorenz S, Dessai S, Forster PM, Paavola J (2015) Tailoring the visual communication of climate projections for local adaptation practitioners in Germany and the UK. Philos Trans Math Phys Eng Sci 373:1–17. https://doi.org/10.1098/rsta.2014.0457

MacEachren AM (1992) Application of environmental learning theory to spatial knowledge acquisition from maps. Ann Assoc Am Geogr 82:245–274. https://doi.org/10.1111/j.1467-8306.1992.tb01907.x

MacEachren AM, Roth RE, O’Brien J et al (2012) Visual semiotics and uncertainty visualization: an empirical study. IEEE Trans vis Comput Graph 18:2496–2505. https://doi.org/10.1109/tvcg.2012.279

Marchak FM (1994) An overview of scientific visualization techniques applied to experimental psychology. Behav Res Methods Instrum Comput 26:177–180. https://doi.org/10.3758/BF03204613

Mason RO, Mitroff II (1981) Challenging strategic planning assumptions: theory, cases, and techniques. Wiley

Mayer RE, Gallini JK (1990) When is an illustration worth ten thousand words? J Educ Psychol 82:715. https://doi.org/10.1037/0022-0663.82.4.715

Mcbride M, Caldara M (2013) The efficacy of tables versus graphs in disrupting dark networks: an experimental study. Soc Netw 35:406–422. https://doi.org/10.1016/j.socnet.2013.04.008

McInerny GJ, Chen M, Freeman R et al (2014) Information visualisation for science and policy: engaging users and avoiding bias. Trends Ecol Evol 29:148–157. https://doi.org/10.1016/j.tree.2014.01.003

McIntire JP, Havig PR, Geiselman EE (2014) Stereoscopic 3D displays and human performance: a comprehensive review. Displays 35:18–26. https://doi.org/10.1016/j.displa.2013.10.004

Meyer AD (1991) Visual data in organizational research. Organ Sci 2:218–236. https://doi.org/10.1287/orsc.2.2.218

Meyer J (2000) Performance with tables and graphs: effects of training and a visual search model. Ergonomics 43:1840–1865. https://doi.org/10.1080/00140130050174509

Meyer J, Shinar D, Leiser D (1997) Multiple factors that determine performance with tables and graphs. Hum Factors 39:268–286. https://doi.org/10.1518/001872097778543921

Meyer J, Shamo MK, Gopher D (1999) Information structure and the relative efficacy of tables and graphs. Hum Factors 41:570–587. https://doi.org/10.1518/001872099779656707

Miettinen K (2014) Survey of methods to visualize alternatives in multiple criteria decision making problems. Spectr 36:3–37. https://doi.org/10.1007/s00291-012-0297-0

Miniard PW, Bhatla S, Lord KR et al (1991) Picture-based persuasion processes and the moderating role of involvement. J Consum Res 18:92–107. https://doi.org/10.1086/209244

Nadav-Greenberg L, Joslyn SL, Taing MU (2008) The effect of uncertainty visualizations on decision making in weather forecasting. J Cogn Eng Decis Mak 2:24–47. https://doi.org/10.1518/155534308X284354

Nelson PM, Van Norman ER, Christ TJ (2017) Visual analysis among novices: training and trend lines as graphic aids. Contemp Sch Psychol 21:93–102. https://doi.org/10.1007/s40688-016-0107-9

Newman GE, Scholl BJ (2012) Bar graphs depicting averages are perceptually misinterpreted: the within-the-bar bias. Psychon Bull Rev 19:601–607. https://doi.org/10.3758/s13423-012-0247-5

O’Keefe RM, Pitt IL (1991) Interaction with a visual interactive simulation, and the effect of cognitive style. Eur J Oper Res 54:339–348. https://doi.org/10.1016/0377-2217(91)90109-9

Ognjanovic S, Thüring M, Murphy RO, Hölscher C (2019) Display clutter and its effects on visual attention distribution and financial risk judgment. Appl Ergon 80:168–174. https://doi.org/10.1016/j.apergo.2019.05.008

Okan Y, Garcia-Retamero R, Cokely ET, Maldonado A (2018a) Biasing and debiasing health decisions with bar graphs: costs and benefits of graph literacy. Q J Exp Psychol 71:2506–2519. https://doi.org/10.1177/1747021817744546

Okan Y, Stone ER, Bruine W, de Bruin, (2018b) Designing graphs that promote both risk understanding and behavior change. Risk Anal 38:929–946. https://doi.org/10.1111/risa.12895

Padilla LM, Hansen G, Ruginski IT et al (2015) The influence of different graphical displays on nonexpert decision making under uncertainty. J Exp Psychol Appl 21:37–46. https://doi.org/10.1037/xap0000037

Padilla LM, Creem-Regehr SH, Hegarty M, Stefanucci JK (2018) Decision making with visualizations: a cognitive framework across disciplines. Cogn Res Princ Implic. https://doi.org/10.1186/s41235-018-0120-9

Parrott R, Silk K, Dorgan K et al (2005) Risk comprehension and judgments of statistical evidentiary appeals: When a picture is not worth a thousand words. Hum Commun Res 31:423–452. https://doi.org/10.1093/hcr/31.3.423

Peebles D (2008) The effect of emergent features on judgments of quantity in configural and separable displays. J Exp Psychol Appl 14:85–100. https://doi.org/10.1037/1076-898x.14.2.85

Peng C-H, Lurie NH, Slaughter SA (2019) Using technology to persuade: visual representation technologies and consensus seeking in virtual teams. Inf Syst Res 30:948–962. https://doi.org/10.1287/isre.2019.0843

Pennington R, Tuttle B (2009) Managing impressions using distorted graphs of income and earnings per share: the role of memory. Int J Account Inf Syst 10:25–45. https://doi.org/10.1016/j.accinf.2008.10.001

Perdana A, Robb A, Rohde F (2018) Does visualization matter? The role of interactive data visualization to make sense of information. Australas J Inf Syst 22:1–35. https://doi.org/10.3127/ajis.v22i0.1681

Perdana A, Robb A, Rohde F (2019) Interactive data and information visualization: unpacking its characteristics and influencing aspects on decision-making. Pac Asia J Assoc Inf Syst 11:75–104. https://doi.org/10.17705/1pais.11404

Pfaff MS, Klein GL, Drury JL et al (2013) Supporting complex decision making through option awareness. J Cogn Eng Decis Mak 7:155–178. https://doi.org/10.1177/1555343412455799

Phillips B, Prybutok VR, Peak DA (2014) Decision confidence, information usefulness, and information seeking intention in the presence of disconfirming information. Inform Sci Int J Emerg Transdiscipl 17:1–25. https://doi.org/10.28945/1932

Pieters R, Wedel M, Batra R (2010) The stopping power of advertising: measures and effects of visual complexity. J Mark 74:48–60. https://doi.org/10.1509/jmkg.74.5.48

Plass JL, Heidig S, Hayward EO et al (2014) Emotional design in multimedia learning: effects of shape and color on affect and learning. Learn Instrum 29:128–140. https://doi.org/10.1016/j.learninstruc.2013.02.006

Platts K, Tan KH (2004) Strategy visualisation: knowing, understanding, and formulating. Manag Decis 42:667–676. https://doi.org/10.1108/00251740410538505

Podsakoff PM, MacKenzie SB, Bachrach DG, Podsakoff NP (2005) The influence of management journals in the 1980s and 1990s. Strateg Manag J 26:473–488. https://doi.org/10.1002/smj.454

Porat T, Oron-Gilad T, Meyer J (2009) Task-dependent processing of tables and graphs. Behav Inf Technol 28:293–307. https://doi.org/10.1080/01449290701803516

Quattrone P (2017) Embracing ambiguity in management controls and decision-making processes: on how to design data visualisations to prompt wise judgement. Account Bus Res 47:588–612. https://doi.org/10.1080/00014788.2017.1320842

Radley KC, Dart EH, Wright SJ (2018) The effect of data points per x- to y-axis ratio on visual analysts evaluation of single-case graphs. Sch Psychol Q 33:314–322. https://doi.org/10.1037/spq0000243

Raidvee A, Toom M, Averin K, Allik J (2020) Perception of means, sums, and areas. Atten Percept Psychophys. https://doi.org/10.3758/s13414-019-01938-7

Raschke RL, Steinbart PJ (2008) Mitigating the effects of misleading graphs on decisions by educating users about the principles of graph design. J Inf Syst 22:23–52. https://doi.org/10.2308/jis.2008.22.2.23

Reani M, Peek N, Jay C (2019) How different visualizations affect human reasoning about uncertainty: an analysis of visual behaviour. Comput Hum Behav 92:55–64. https://doi.org/10.1016/j.chb.2018.10.033

Rebotier TP, Kirsh DJ, McDonough L (2003) Image-Dependent Interaction of Imagery and Vision. Am J Psychol 116:343–366. https://doi.org/10.2307/1423498

Rose ED (1966) Image, sound, and meaning. J Univ Film Prod Assoc 18:21–23

Sanfey A, Hastie R (1998) Does evidence presentation format affect judgment? An experimental evaluation of displays of data for judgments. Psychol Sci 9:99–103. https://doi.org/10.1111/1467-9280.00018

Sato Y, Stapleton G, Jamnik M, Shams Z (2019) Human inference beyond syllogisms: an approach using external graphical representations. Cogn Process 20:103–115. https://doi.org/10.1007/s10339-018-0877-2

Semmler C, Brewer N (2002) Using a flow-chart to improve comprehension of jury instructions. Psychiatry Psychol Law 9:262–267. https://doi.org/10.1375/13218710260612136

Sen T, Boe WJ (1991) Confidence and accuracy in judgements using computer displayed information. Behav Inf Technol 10:53–64. https://doi.org/10.1080/01449299108924271

Smerecnik CMR, Mesters I, Kessels LTE et al (2010) Understanding the positive effects of graphical risk information on comprehension: Measuring attention directed to written, tabular, and graphical risk information. Risk Anal 30:1387–1398. https://doi.org/10.1111/j.1539-6924.2010.01435.x

Speier C (2006) The influence of information presentation formats on complex task decision-making performance. Int J Hum-Comput Stud 64:1115–1131. https://doi.org/10.1016/j.ijhcs.2006.06.007

Spence I, Kutlesa N, Rose DL (1999) Using color to code quantity in spatial displays. J Exp Psychol Appl 5:393–412. https://doi.org/10.1037/1076-898X.5.4.393

Stone ER (2018) Link to external site this link will open in a new window, Reeder EC, et al. salience versus proportional reasoning: rethinking the mechanism behind graphical display effects. J Behav Decis Mak 31:473–486. https://doi.org/10.1002/bdm.2051

Stone ER, Bruin W, Wilkins AM et al (2017) Designing graphs to communicate risks: understanding how the choice of graphical format influences decision making. Risk Anal 37:612–628. https://doi.org/10.1111/risa.12660

Subramanian GH, Nosek J, Rahunathan SP, Kanitkar SS (1992) A comparison of the decision table and tree. Commun ACM 35:89–94. https://doi.org/10.1145/129617.129621

Sun Y, Li S, Bonini N (2010) Attribute salience in graphical representations affects evaluation. Judgm Decis Mak 5:151–158

Sun Y, Li S, Bonini N, Liu Y (2016) Effect of graph scale on risky choice: evidence from preference and process in decision-making. PLoS ONE. https://doi.org/10.1371/journal.pone.0146914

Suwa M, Tversky B (2002) External representations contribute to the dynamic construction of ideas. Springer, pp 341–343

Symmank C (2019) Extrinsic and intrinsic food product attributes in consumer and sensory research: literature review and quantification of the findings. Manag Rev Q 69:39–74. https://doi.org/10.1007/s11301-018-0146-6

Tak S, Toet A, van Erp J (2015) Public understanding of visual representations of uncertainty in temperature forecasts. J Cogn Eng Decis Mak 9:241–262. https://doi.org/10.1177/1555343415591275

Tang F, Hess TJ, Valacich JS, Sweeney JT (2014) The Effects of visualization and interactivity on calibration in financial decision-making. Behav Res Account 26:25–58. https://doi.org/10.2308/bria-50589

Tintarev N, Masthoff J (2016) Effects of individual differences in working memory on plan presentational choices. Front Psychol 7:1793. https://doi.org/10.3389/fpsyg.2016.01793

Toker D, Conati C (eds) (2017) Leveraging pupil dilation measures for understanding users’ cognitive load during visualization processing, pp 267–270

Tuttle BM, Kershaw R (1998) Information presentation and judgment strategy from a cognitive fit perspective. J Inf Syst 12:1

Umanath NS, Vessey I (1994) Multiattribute data presentation and human judgment: a cognitive fit perspective. Decis Sci 25:795–824. https://doi.org/10.1111/j.1540-5915.1994.tb01870.x

van den Berg R, Cornelissen FW, Roerdink JBTM (2007) Perceptual dependencies in information visualization assessed by complex visual search. ACM Trans Appl Percept. https://doi.org/10.1145/1278760.1278763

Van der Linden SL, Leiserowitz AA, Feinberg GD, Maibach EW (2014) How to communicate the scientific consensus on climate change: plain facts, pie charts or metaphors? Clim Change 126:255–262. https://doi.org/10.1007/s10584-014-1190-4

Van Norman ER, Nelson PM, Shin J-E, Christ TJ (2013) An evaluation of the effects of graphic aids in improving decision accuracy in a continuous treatment design. J Behav Educ 22:283–301. https://doi.org/10.1007/s10864-013-9176-2

Verovsek Š, Juvancic M, Zupancic T (2013) Using visual language to represent interdisciplinary content in urban development. Urbani Izziv 24:144–155. https://doi.org/10.5379/urbani-izziv-en-2013-24-02-006

Vessey I (1991) Cognitive fit: a theory-based analysis of the graphs versus tables literature. Decis Sci 22:219–240. https://doi.org/10.1111/j.1540-5915.1991.tb00344.x

Vessey I, Galletta D (1991) Cognitive fit: An empirical study of information acquisition. Inf Syst Res 2:63–84. https://doi.org/10.1287/isre.2.1.63

Vila J, Gomez Y (2016) Extracting business information from graphs: an eye tracking experiment. J Bus Res 69:1741. https://doi.org/10.1016/j.jbusres.2015.10.048

Volkov A, Laing GK (2012) Assessing the value of graphical presentations in financial reports. Australas Account Bus Finance J 6:85–107

Wang D, Guo D, Zhang H (eds) (2017) Spatial temporal data visualization in emergency management: a view from data-driven decision. Rolando Beach, CA, USA, pp 1–7

Washburne JN (1927) An experimental study of various graphic, tabular, and textual methods of presenting quantitative material. J Educ Psychol 18:361. https://doi.org/10.1037/h0070054

Watkins ET (2000) Improving the analyst and decision-maker’s perspective through uncertainty visualization. Master’s thesis, Air Force Institute of Technology, Wright-Patterson AFB, Ohio

Wesslen R, Santhanam S, Karduni A et al (2019) Investigating effects of visual anchors on decision-making about misinformation. Comput Graph Forum 38:161–171. https://doi.org/10.1111/cgf.13679

Whittington R, Yakis-Douglas B, Ahn K (2016) Cheap talk? Strategy presentations as a form of chief executive officer impression management. Strateg Manag J 37:2413–2424. https://doi.org/10.1002/smj.2482

Wu CM, Meder B, Filimon F, Nelson JD (2017) Asking better questions: How presentation formats influence information search. J Exp Psychol Learn Mem Cogn 43:1274–1297. https://doi.org/10.1037/xlm0000374

Xu Y (2005) The effect of graphic disclosures on users’ perceptions: an experiment. J Account Finance Res 13:39–50

Yigitbasioglu OM, Velcu O (2012) A review of dashboards in performance management: Implications for design and research. Int J Account Inf Syst 13:41–59. https://doi.org/10.1016/j.accinf.2011.08.002

Yildiz E, Boehme R (eds) (2017) Effects of information security risk visualization on managerial decision making. Internet Society, Paris, France

Yoon SA (2011) Using social network graphs as visualization tools to influence peer selection decision-making strategies to access information about complex socioscientific issues. J Learn Sci 20:549–588. https://doi.org/10.1080/10508406.2011.563655

Zabukovec A, Jaklič J (2015) The impact of information visualisation on the quality of information in business decision-making. Int J Technol Hum Interact IJTHI 11:61–79. https://doi.org/10.4018/ijthi.2015040104

Zacks J, Levy E, Tversky B, Schiano DJ (1998) Reading bar graphs: effects of extraneous depth cues and graphical context. J Exp Psychol Appl 4:119–138. https://doi.org/10.1037/1076-898X.4.2.119

Zelazny G (2001) Say it with charts: the executive’s guide to visual communication. McGraw-Hill Education

Zhang P (1998) An image construction method for visualizing managerial data. Decis Support Syst 23:371. https://doi.org/10.1016/s0167-9236(98)00050-5

Download references

Open Access funding enabled and organized by Projekt DEAL.

Author information

Authors and affiliations.

Chair of Strategic and International Management, Philipps-University Marburg, Universitätsstr. 24, 35037, Marburg, Germany

Karin Eberhard

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Karin Eberhard .

Ethics declarations

Conflict of interest.

The author declares that there is no conflict of interest.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Eberhard, K. The effects of visualization on judgment and decision-making: a systematic literature review. Manag Rev Q 73 , 167–214 (2023). https://doi.org/10.1007/s11301-021-00235-8

Download citation

Received : 26 October 2020

Accepted : 11 August 2021

Published : 25 August 2021

Issue Date : February 2023

DOI : https://doi.org/10.1007/s11301-021-00235-8

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Information visualization
• Strategic decision-making
• Decision quality
• Cognitive load
• User characteristics
• Task characteristics

JEL Classification

• Find a journal
• Publish with us
• Track your research

• Original article
• Open access
• Published: 17 September 2020

A new measure of group decision-making efficiency

• Cheng-Ju Hsieh 1 ,
• Mario Fifić 2 &
• Cheng-Ta Yang   ORCID: orcid.org/0000-0002-6561-763X 3 , 4  

Cognitive Research: Principles and Implications volume  5 , Article number:  45 ( 2020 ) Cite this article

16k Accesses

7 Citations

Metrics details

It has widely been accepted that aggregating group-level decisions is superior to individual decisions. As compared to individuals, groups tend to show a decision advantage in their response accuracy. However, there has been a lack of research exploring whether group decisions are more efficient than individual decisions with a faster information-processing speed. To investigate the relationship between accuracy and response time (RT) in group decision-making, we applied systems’ factorial technology, developed by Townsend and Nozawa ( Journal of Mathematical Psychology 39 , 321–359, 1995) and regarded as a theory-driven methodology, to study the information-processing properties. More specifically, we measured the workload capacity C AND ( t ), which only considers the correct responses, and the assessment function of capacity A AND ( t ), which considers the speed-accuracy trade-off, to make a strong inference about the system-level processing efficiency. A two-interval, forced-choice oddball detection task, where participants had to detect which interval contains an odd target, was conducted in Experiment 1. Then, in Experiment 2, a yes/no Gabor detection task was adopted, where participants had to detect the presence of a Gabor patch. Our results replicated previous findings using the accuracy-based measure: Group detection sensitivity was better than the detection sensitivity of the best individual, especially when the two individuals had similar detection sensitivities. On the other hand, both workload capacity measures,  C AND ( t ) and  A AND ( t ), showed evidence of supercapacity processing, thus suggesting a collective benefit. The ordered relationship between accuracy-based and RT-based collective benefit was limited to the A AND ( t ) of the correct and fast responses, which may help uncover the processing mechanism behind collective benefits. Our results suggested that  A AND ( t ), which combines both accuracy and RT into inferences, can be regarded as a novel and diagnostic tool for studying the group decision-making process.

Significance

Previous studies have shown the so-called collective benefit. That is, performance is more accurate when participants work as a group in which they can communicate with each other verbally or non-verbally, and with an exchange of decision evidence or internal estimate of confidence. However, it is still unclear whether group decisions are more efficient than individual decisions since a tradeoff may exist between speed and accuracy. In other words, increasing the number of group members may increase the group’s response accuracy, but at the same time would slow down the processing speed. The aim of the study was to learn about the relationship between accuracy and response-time (RT) measures in group decision-making. To sum up, our results replicated the previous findings that showed the collective benefit for accuracy: the group’s detection sensitivity was higher than the best individual’s detection sensitivity only when group members’ detection sensitivities were similar. The measures of processing speed, the workload capacity measures, revealed that group decision-making was of supercapacity processing. In addition, our results suggested that the assessment function of workload capacity, which combines both accuracy and RT into inferences, can be regarded as a novel and diagnostic tool to study the group decision-making process. The current study is not only a replication of the previous studies, but also highlights the importance of combined accuracy and RT measures in the inference of the group decision-making process.

Introduction

An old saying goes, “Two heads are better than one.” Combining group members’ opinions to make a coherent decision is usually regarded as a better means of decision-making than having an individual make a decision alone (Clemen, 1989 ). Many important real-world decisions are collective decisions, such as juries rendering verdicts or a team of radiologists reading X-rays. The situation in which group decisions are considered to be superior to individual decisions is termed “wisdom of crowds” Footnote 1 (Surowiecki, 2004 ).

In the literature, the primary focus has been on learning about the mechanism that underlies group decisions. Previous studies have investigated the properties of cooperation between two or more participants in perceptual decision tasks (Bahrami, Olsen, Latham, Roepstorff, Rees, & Frith, 2010 ; Bahrami, Olsen, Bang, Roepstorff, Rees, & Frith, 2012a ; Sorkin & Dai, 1994 ; Sorkin, Hays, & West, 2001 ; Sorkin, West, & Robinson, 1998 ) and found that a group exhibits a decision advantage—a so-called “collective benefit”—over an individual decision-maker. There are several proposed possible explanations for the existence of the collective benefit.

First, it is possible that the collective benefit results from a reduced workload, as group members strategically split information among themselves. The fact that each member can focus attention on a subset of information (which means the group does not have to focus its attention on the entirety of information) could lead to an increase in collective processing efficiency. However, group decision accuracy may be limited by individual ability because the group must rely on the capabilities of each member. This possibility has been challenged and ruled out by Barr and Gold ( 2014 ). Barr and Gold ( 2014 ) manipulated the group size (one to four members) and the quantity of information (partial or full) that each member received. Their results showed that groups viewing the entirety of information significantly outperformed groups whose members viewed limited portions of information and suggested that strategically splitting information does not necessarily lead to a collective benefit.

Second, the collective benefit could be due to the statistical facilitation effect (Green & Swets, 1966 ; Lorge & Solomon, 1955 ; Sorkin & Dai, 1994 ; Sorkin et al., 2001 ; Swets, Shipley, McKey, & Green, 1959 ). In stochastic modeling, adding more independent random variables to a parallel processing system can lead to a faster and more accurate task completion. The statistical facilitation that is achieved through the redundancy gain, has been used by engineers to decrease failure rate of their machines. Analogously, one can demonstrate a similar effect in human group decision-making, by increasing the number of independent decision-makers who work in parallel. The overall group’s achievement will be better than that of any individual member working alone. Interestingly, the collective benefit, that is due to the statistical facilitation and the redundancy gain, is merely an outcome of the statistical improvement – that is, the group benefit is not achieved by the group members’ interaction. Such a statistical facilitation effect is conditioned on the use of the so-called first-termination rule, Footnote 2 which means that the system would wait for the fastest and correctly responding unit to complete and would then use it to make the final decision while ignoring the unfinished or incorrectly responding units.

Third, the collective benefit could be a result of the integration of evidence collected by each group member via social interaction (Bahrami et al., 2010 ; Lorenz, Rauhut, Schweitzer, & Helbing, 2011 ). This explanation is different from the first two in that it assumes the collective sum of knowledge occurs not only as a simple sum of individual knowledge, but as novel knowledge created through a series of social interactions. As a result, the performance of a group is better than that of the best observer or exceeds the expectations of individual members working in isolation (Collins & Guetzkow, 1964; Davis, 1969 , 1973 ). Consistent with the coactive model Footnote 3 (Houpt & Townsend, 2011 ; Schwarz, 1989 , 1994 ; Townsend & Nozawa, 1995 ), the collective benefit may have occurred because the individual contribution is weighted and integrated into a single information channel following a “weighted-and-sum” principle of information integration.

Recently, an increasing number of studies has challenged the idea that group decisions would always outperform individual decisions. For example, Fific and Gigerenzer ( 2014 ) suggested that adding more decision-makers does not necessarily enhance the group performance. The best individual may match the collective decision accuracy or even outperform the group, especially when free riders exist. In addition, a single expert, under certain conditions, can outperform the group (e.g., Gordon, 1924 ; Graham, 1996 ; Winkler & Poses, 1993 ). The debate over the potential negative effect of collaboration on decision-making has intensified work that explores the conditions under which the collective benefit/cost may arise.

A potential factor that may influence the collective effect was performance similarity between group members. In a two-interval forced-choice oddball detection task in which participants had to decide which interval contained an odd target, Bahrami et al. ( 2010 ) investigated whether participants can utilize their partner’s confidence rating to improve group decision sensitivity. Footnote 4 The results showed that only in a consistent group, in which the two group members had similar detection sensitivity, was the group decision superior to individual decisions. Specifically, the authors used S min and S max to represent the detection sensitivity of the worse and better individual, respectively, and only when  S min / S max  ≥ 0.4 did the group show the collective benefit, i.e., S dyad / S max  > 1 (S dyad denotes group detection sensitivity). By contrast, in an inconsistent group (i. e. , S min / S max  < 0.4), the group decision was worse than the decision made by its better group member. Bahrami et al. ( 2010 ) further suggested that the results supported the weighted confidence-sharing model (WCS), which assumes that individuals can take advantage of the confidence information, i.e., an internal estimate of the probability of being correct; the final decision is made based on the weighting function of the group members’ confidence. The WCS model can be considered a variant of the coactive models.

Another important factor in understanding the collective is the tradeoff between accuracy and speed in group decision-making (see Heitz, 2014 for a review). The collective effect can be measured by both response accuracy and RT. However, when used in the same task, the two measures can have an inverse relationship. That is, increasing the number of group members may increase the group’s response accuracy, but, at the same time, could create longer RTs; for example, as the group size increases, group members require more time to communicate with each other to reach a consensus. Thus, it is reasonable to speculate that collaboration can increase response accuracy but slow down the decision RT. The above-mentioned studies focused mainly on the effect of collaboration on accuracy measures (e.g., Bahrami et al., 2010 , 2012a , b ) while neglecting its effect on the measure of processing speed.

Two other studies utilized RT measures (e.g., RT distribution analysis) to assess the collective effect (e.g., Brennan & Enns, 2015 ; Yamani, Neider, Kramer, & McCarley, 2017 ). Brennan and Enns ( 2015 ) tested the violation of the race-model inequality to infer the collective benefit. In general, the race model assumes that two processing units are racing to reach a certain decision criterion (Miller, 1982 ). The two units work independently, and the faster unit, which reaches the decision criteria first, will determine the response outcome. In our domain of interest, the units are defined as the individual decision-makers. The race-model inequality assumes that two group members work independently and in parallel (simultaneously). In a nutshell, a violation of the race-model inequality would suggest that the two decision-makers did not work independently of each other and that at some point, prior to making a final decision, they interacted with each other. In terms of modeling processing systems, this situation is defined by coactive processing. Using a visual enumeration task in which participants were required to count the number of targets (0/1/2) presented against the distractors, Brennan and Enns ( 2015 ) demonstrated that the observed RT data violated the race-model inequality, thereby supporting the notion that the two decision-makers did not work independently and collaboration would facilitate decision RTs.

Using another RT measure, i.e., workload capacity, Yamani et al. ( 2017 ) examined how collaboration affects individual processing efficiency in terms of the information-processing speed. Workload capacity is a measure of the change in processing efficiency (speed) at the individual subject level when the system’s workload (i.e., the number of decision-makers) increases, as proposed by the framework of Systems Factorial Technology (SFT, Little, Altieri, Fific, & Yang, 2017 ; Townsend & Nozawa, 1995 ). According to SFT, increasing the number of processing units (i.e., the system’s workload) can have three different effects on the processing speed of an individual processing unit. In the case of limited-capacity processing, the speed of processing per processing unit slows down when more units operate at the same time. In the case of unlimited-capacity processing, the speed of processing per processing unit remains unchanged when more processing units are added. In the case of supercapacity processing, the speed of processing per processing unit speeds up by the addition of more decision units. This could be the result of a facilitatory interaction between decision units. It is notable that an unlimited-capacity system implies that the efficiency of an individual unit remains unchanged, whereas the system overall can be performing better compared with the individual unit, due to stochastic considerations. In the context of group decision-making, the limited capacity indicates some form of inhibition between individual decision-makers when they work as a group. In the case of unlimited capacity, the addition of more group members does not affect individual efficiency. In the case of supercapacity, efficiency improves as a result of the group members’ facilitatory interaction. Yamani et al. ( 2017 ) study adopted the shared-gaze technique which allows participants to see where their partner is looking in order to study whether collaboration can benefit scan in teams. Results showed supercapacity processing when both group members were required to find and respond to a target; by contrast, limited capacity was found when the faster searcher found and responded to the target. The supercapacity results suggested that, by holding fixation on the target, the faster searcher can cue their partner to the target location, which, in turn, boosts the processing speed of the slower searcher. The limited-capacity results suggested that shared gaze offered no benefits but slowed down the processing for the faster searcher.

To summarize, collective benefit/cost can be measured by either response accuracy or RTs. These two measures play complementary roles in understanding the mechanism underlying group decisions. However, to our knowledge, there is no prior study that combines accuracy and RT measures to infer the dynamic process of group decision-making. This raises several related questions. Are the inferences from the two measures consistent enough to draw similar conclusions? Is it possible to use a single performance index to quantify the collective effect by considering the two measures simultaneously? In the present study, we integrated, within one study, the two approaches by applying SFT (Little et al., 2017 ; Townsend & Nozawa, 1995 ). Before we go into the details about the present study, we first briefly introduce the theory and methodology of Systems Factorial Technology (SFT).

Systems Factorial Technology

SFT (Little et al., 2017 ; Townsend & Nozawa, 1995 ) is a useful tool for analyzing and diagnosing the dynamic decision-making process. A wide range of fields in cognitive research have utilized SFT, such as visual search (Zehetleitner, Krummenacher, & Müller, 2009 ), memory search (Townsend & Fific, 2004 ; Van Zandt & Townsend, 1993 ), face perception (Ingvalson & Wenger, 2005 ; Yang, Altieri, & Little, 2018 ), classification (Fific, Nosofsky, & Townsend, 2008 ), change detection (Yang, 2011 ; Yang, Chang, & Wu, 2013 ; Yang, Hsu, Huang, & Yeh, 2011 ), cued detection (Yang, Little, & Hsu, 2014 ; Yang, Wang, Chang, Yu, & Little, 2019 ), word processing (Houpt, Sussman, Townsend, & Newman, 2015 ; Houpt, Townsend, & Donkin, 2014 ), audiovisual processing (Altieri & Yang, 2016 ; Yang et al., 2018 ; Yang, Yu, & Chang, 2016 ), and group decision-making (Yamani et al., 2017 ). According to SFT, two important information-processing properties of group decision-making can be uncovered, such as a group’s organization during task participation (i.e., how two individuals work together to achieve a group decision) and workload capacity (i.e., individual decision efficiency varies as a function of the number of decision-makers).

In this paper, we used the workload capacity measure to quantify group decision-making efficacy. The workload capacity measures can be used to indicate the amount and type of the potential collective benefit. Here, we introduced two types of capacity measures. First is the AND capacity Footnote 5 ( C AND ( t )), a standard measure of workload capacity, developed by Townsend and colleagues (e.g., Townsend & Nozawa, 1995 ), which considers only the RT data of correct responses. The AND capacity is analyzed by comparing the group processing efficiency to a baseline predicted from the UCIP model (i.e., unlimited-capacity, independent, parallel model), which assumes that all group members work independently and in parallel. The workload capacity is formalized as a ratio of the cumulative reverse hazard functions, K ( t ) = ln  F ( t ) where F ( t ) = P (RT ≤  t ) (Chechile, 2003 , 2011 ; Townsend & Eidels, 2011 ; Townsend & Wenger, 2004 ), and is expressed as:

for t  > 0, where K 1 , K 2 , and K 12 represent the cumulative reverse hazard function of the two non-collaborative conditions, in which participants perform the task independently, and the collaborative condition, in which participants work together with social interaction (here, the non-verbal communication), respectively. The interpretation of C AND ( t ) > 1 implies that group performance is better than the prediction from the UCIP model—that is, the system engages in supercapacity processing. When C AND ( t ) = 1, it suggests an unlimited-capacity processing system, implying that individual decision performance is unchanged when the number of group members increases. When C AND ( t ) < 1, it suggests a limited-capacity processing system, implying that social interaction may, in fact, even slow down individual decision time.

Second, we introduced a new measure, assessment function of workload capacity A AND ( t ), to analyze group decision efficiency. To our knowledge, A AND ( t ) has not been used to study the group decision-making process. Similar to C AND ( t ), A AND ( t ) has the advantage of inferring the dynamic processing efficiency as a function of RT (Donkin et al., 2014 ; Townsend & Altieri, 2012 ). Better than C AND ( t ), A AND ( t ) combines both accuracy and RT data into the analysis, such that we can analyze the decision efficiency of four response conditions: (a) correct and fast, (b) correct and slow, (c) incorrect and fast, and (d) incorrect and slow. The inferences of A AND ( t ) were similar to C AND ( t ); the details of the data analysis and inferences will be introduced in the “Data analysis” section. Please see Table  1 for the SFT-related theoretical glossary. More details can also be found in Townsend and Altieri ( 2012 ).

The present study

In the present study, we conducted two psychophysics experiments by collecting both accuracy and RT data to test whether collaboration through the exchange of confidence information can promote group decision efficiency. In the first experiment, we extended the study by Bahrami et al. ( 2010 ) by adopting the two-interval forced-choice oddball detection task. In the second experiment, we used a yes/no Gabor detection task. Both accuracy-based and time-based measures were computed to infer the collective effect. First, the accuracy-based collective effect was computed by comparing the dyad’s sensitivity with the maximum individual sensitivity (i.e., S dyad / S max ). Of particular interest was testing whether this effect would change as a function of relative detection sensitivity between the two group members (i.e., S min / S max ). Second, the time-based collective effect was inferred from two workload capacity measures ( C AND ( t )) and A AND ( t ))) as introduced in the previous section.

Our first goal of the study is to replicate the effect of sensitivity similarity on joint decisions. We expect that only when group members have similar detection sensitivities, the collective benefit would be observed; that is, group detection sensitivity will be higher than the detection sensitivity of the best observer. Our second goal is to evaluate how RT and accuracy measures are consistent enough to draw similar conclusions about the collective effect. We expect to observe significant correlations between the time-based and accuracy-based measures of collective effect. Our last goal is to establish the assessment function of workload capacity as a standard measure of group decision efficiency. Considering the speed-accuracy trade-off effect, A AND ( t ), which combines both accuracy and RT into the analysis, can be regarded as a better index for quantifying the group decision advantage.

Experiment 1

In Experiment 1, a two-interval forced-choice oddball detection task was adopted. The relative detection sensitivity between the group and the best observer was computed to infer the accuracy-based collective effect. Additionally, the workload capacity of group decision-making was assessed as a time-based measure of the collective effect. If both accuracy-based and time-based measures can reflect the collective benefit, we should expect that the accuracy-based collective effect (i.e., S dyad / S max ) is greater than 1 and that time-based measures reveal supercapacity results.

Participants

Fourteen (eight male and six female; age: 21.3 ± 2.05 years) undergraduate students at National Cheng Kung University volunteered to participate in this experiment and were randomly divided into seven pairs. All of the participants were right-handed. Before the experiment, participants signed an informed consent form. The ethics approval for the study was obtained from the Human Research Ethics Committee of National Cheng Kung University, and the experiment was conducted in accordance with the approved guidelines and regulations. The participants were either compensated in the amount of NTD 140 per hour or received class credit for their participation.

A desktop computer with a 3.20 G-Hz Intel Core i7–8700 Processor, Intel UHD Graphics 630, and 8 GB of RAM controlled the display and recorded the manual responses. Stimuli were presented on a 19-in. CRT monitor with a refresh rate of 75 Hz and a display resolution of 1024 x 768 pixels. The viewing distance was kept at 60 cm and a chin-rest was used to prevent any head movements. The experiment was programmed using Psychtoolbox ( http://psychtoolbox.org/ ) from MATLAB (Mathworks Inc., Natick, MA, USA).

Design, stimuli, and procedure

Figure  1 shows the flowchart of the task design and Fig.  2 shows an illustration of the trial procedure for each condition. Footnote 6 All the participants participated in three cooperation conditions: the individual condition, the non-collaborative condition, and the collaborative condition. Participants first performed the individual condition. Then they were randomly paired to form a dyad. Footnote 7 As a dyad, each participant faced their screen and was positioned next to their partner at a distance of 55 cm. They could only see their partner’s responding hand. Therefore, we believe that other forms of communication (e.g., body language) are not available. The order of the collaborative and non-collaborative conditions was counterbalanced across dyads (see Fig.  1 ).

A flowchart of the task design. a the non-collaborative condition first. b the collaborative condition first

An illustration of the trial procedure in Experiment 1

The response mode was slightly different across the three cooperation conditions. In the individual condition, participants were required to complete the task alone and to provide a response by clicking the left or right button of the mouse to indicate which interval contained the odd target. After they had made their decision, they were asked to rate their confidence in their judgment using a Likert-type scale from 1 (“very doubtful”) to 5 (“absolutely sure”). The confidence information was used for the collaborative condition in the next stage.

In the non-collaborative condition, two participants working as a dyad performed the task together but without any communication. One participant delivered a response with a mouse click while the other participant responded with a keyboard press. During each trial, only one of the participants was required to respond and the person who was required to respond was dependent on the color of the question mark, with the color green indicating that the participant with the keyboard should respond and the color red indicating that the participant with the mouse should respond. The color (green/red) was randomly selected with equal probability. See Fig. 2 for an example, because the question mark is colored in red, the participant with the mouse should respond and the participant with the keyboard does not need to respond. After a response was made, a feedback display was shown to indicate the correctness of the present decision (middle) and the correctness of the participants’ individual decision about the same trial tested in the individual condition (top and bottom colored in red and green, respectively).

In the collaborative condition, the procedure was the same as that in the non-collaborative condition except that each participant’s confidence information was presented on a horizontal line to indicate the participants’ confidence in the judgment regarding the same trial that was tested in the individual condition. Participants could only communicate by exchanging their confidence information and no other forms of communication (e.g., verbal communication) were allowable. Specifically, the confidence information was represented by two colored marks (i.e., red and green marks representing the confidence rating made by the two participants, respectively). The farthest-left side represented the fact that the participants were very confident that the target was presented at Interval 1, while the farthest-right side represented the fact that the participants were very confident that the target was presented at Interval 2. This display of the participants’ confidence allowed them to communicate with each other and non-verbally exchange their confidence ratings.

Each trial started with a fixation cross displayed for a random duration of between 500 and 1000 ms. Afterward, two consecutive stimulus displays were presented for 85 ms with a 1000-ms blank interval in between the two displays. Each display contained six Gabor patches, which were placed in an imaginary circle with a radius of 232.7 pixels (8.0°). All the Gabor patches were placed equidistant from each other. Each Gabor patch was vertically oriented (standard deviation of the Gaussian envelope: 0.45°; spatial frequency: 1.049 cycle/°). In one of the two displays, there were an odd target and five distractors; in the other display, there were six identical distractors. The contrast of the distractor was 10%. The contrast of the target was 1.5, 3.5, 7.0, or 15.0% higher than the distractor (i.e., the contrast level of the target was 11.5, 13.5, 17.0, or 25.0%) and was randomly selected from one of the four contrast levels. The background was colored in gray with a luminance of 36.34 cd/m 2 . The participant’s task was to make a two-interval forced-choice (2IFC) to indicate which interval contained the odd target, as accurately and quickly as possible after the question mark was presented. The question mark was presented until the participants delivered a response or 2 s had elapsed.

Each cooperation condition involved two sessions in order to collect enough data points. In each session, participants first performed a practice block of 64 trials and then 10 blocks of formal trials. Each block consisted of 2 (target was presented at Interval 1 or 2) × 4 (contrast levels of the target) × 8 (trials per combination).

Data analysis

The practice trials were excluded from the analysis. Correct RTs within a range of the 2.5% quartile and the 97.5% quartile were extracted for further analysis to exclude the outliers. To compute the collective effect, we compared the data of the collaborative condition (i.e., two participants working together with an exchange of the partner’s confidence) to the data of the non-collaborative condition (i.e., two participants working independently without any communication). In doing that, we were able to conclude that the collective benefit stems from the non-verbal communication rather than from the social facilitation effect. It is notable that, in the following, the term “non-collaborative-individual” represents the individual performance extracted from the non-collaborative condition rather than from the individual condition. The reason why we did not directly compare the data of the collaborative condition to the data of the individual condition is that a large processing difference existed between the two conditions. That is, in the collaborative and non-collaborative conditions, participants were required to first process the color of the question mark such that they would know who was going to respond in the present trial, whereas there was no need to process the color of the question mark in the individual condition. If the processing was less efficient in the collaborative condition than it was in the individual condition, we did not know whether the results were due to collective cost or whether making a group decision would require an additional process.

First, we conducted a mixed-design analysis of variance (ANOVA) to test the differences in accuracy and mean correct RTs, with the social condition (better observer, worse observer, collaboration) serving as a between-subject factor and with contrast level (11.5, 13.5, 17.0, or 25.0%) serving as a within-subject factor. It is notable that the data of the worse observer and better observer was extracted from the data of the non-collaborative condition and that the collaboration data was extracted from the data of the collaborative condition. We did not incorporate the data of the individual condition into analysis, as mentioned above; however, the data would provide necessary information for the collaborative condition.

The participants’ perceptual sensitivity was derived from response accuracy. To quantify perceptual sensitivity, the maximum slope of the psychometric function of contrast sensitivity was estimated for individuals and dyads, respectively (Bahrami et al., 2010 , 2012a , b ). We fitted the data with a cumulative Gaussian function to estimate the maximum slope of the psychometric function. A larger slope indicates higher sensitivity. The cumulative Gaussian function with two parameters, i.e., bias ( b ) and variance ( σ 2 ), was fitted by a procedure of maximum likelihood estimation via the MATLAB Palamedes toolbox ( http://www.palamedestoolbox.org/ ) (Mathworks Inc., Natick, MA, USA). The psychometric curve, denoted as P ( ∆c ) where ∆c is the contrast difference between the target and distractor, can be expressed as:

where H ( z ) is the cumulative Gaussian function:

Here, P ( ∆c ) corresponds to the probability of responding that the odd target was presented during the second interval. By definition, the variance σ 2 is related to the maximum slope of the psychometric curve, denoted as s , which can be expressed as:

We defined “collective effect” as the ratio of the dyad’s slope ( S dyad ) to that of the better observer (i.e., the individual with a larger slope, S max ). A collective effect larger than 1 indicates that the dyad managed to gain an advantage over its better individual, suggesting collective benefit. Values below 1 indicate that collaboration was counterproductive and that the dyad did worse than its better individual—namely, collective cost. The effect of relative detection sensitivity on the collective benefit can be revealed by plotting the collective effect against the relative sensitivity between the worse and the better observer ( S min / S max ).

Moreover, we adopted SFT to calculate the workload capacity of group decision-making (Little et al., 2017 ; Townsend & Nozawa, 1995 ). We used the SFT R package (Houpt, Blaha, McIntire, Havig, & Townsend, 2014 ) to compute workload capacity C AND ( t ) and the assessment function of workload capacity A AND ( t ) (please see the introduction for more details). Notably, A AND ( t ) incorporates both RT and accuracy data into the analysis. The probabilities of responses can be classified into four categories: (a) the probability that a correct response is made by time t (correct and fast decision), (b) the probability that an incorrect response is made by time t (incorrect and fast decision), (c) the probability that a correct response will be made but it has not happened by time t (correct and slow decision), and (d) the probability that an incorrect response will be made but it has not happened by time t (incorrect and slow decision).

The influence of additional decision-makers was measured for each of the four response types by comparing the collaborative performance with the predicted UCIP baseline derived from the non-collaborative condition. This interpretation is more nuanced than the standard capacity coefficient. In contrast to the standard capacity coefficient C AND ( t ), when interpreting the A AND ( t ) function, one must consider the type of response being made. The following examples may help readers understand how we make inferences using A AND ( t ). The interpretation of A AND ( t ) for correct and fast responses bears the closest resemblance to the standard capacity coefficient. When A AND ( t ) = 1 for correct and fast responses, the implication is that the observed responses made before time t are as probable as expected by the UCIP baseline model. A correct and fast A AND ( t ) > 1 means that participants make more correct responses by time t than expected and, thus, exhibit a form of supercapacity. Similarly, correct and fast A AND ( t ) < 1 implies that fewer correct responses are made by time t than expected by the UCIP model (i.e., capacity is limited). The interpretation differs for the other types of responses. For example, for the incorrect and slow responses,  A AND ( t ) > 1 indicates that more incorrect responses are made after time t than is expected by the UCIP model, which implies a form of limited capacity.

Finally, a quantitative comparison was made to observe the relationship between the accuracy-based measure and time-based measure. We first employed functional principal components analysis (fPCA) with varimax rotation to decompose C AND ( t ) and A AND ( t ) into several principal components (Burns, Houpt, Townsend, & Endres, 2013 ). The SFT R package was adopted (Houpt, Blaha, et al., 2014 ). fPCA is a structural extension of standard PCA (Ramsay & Silverman, 2005 ) and can be used to describe the entire functions using a small number of scalar values (i.e., the loading of the principal component) (Burns et al., 2013 ). The approach enabled us to describe which part of the function-level property is crucial for distinguishing collective effect as a function of RT. Then, we tested the relationship between the factor scores and the accuracy-based collective effect measured by sensitivity such that the relationship between the accuracy-based measure and the component functions can be uncovered.

In order to let the readers follow our SFT analysis steps to replicate our results (e.g., results of C AND ( t ), A AND ( t ), and fPCA analysis), please see the Supplementary material (S. 1 ). We upload our script and data into Github. Files can be downloaded at https://github.com/hanekaze/A-new-measure-of-group-decision-making-efficiency .

Table  2 presents the mean correct RTs and accuracy for all the combinations of the social condition and contrast level.

For RT, the results showed a significant main effect of contrast level [ F (3, 54) = 63.46, p  < 0.001, \( {\eta}_p^2 \) = 0.78]. Post hoc comparison showed that the mean RT of the contrast level of 25.0% was the fastest and that the mean RT of the contrast level of 17.0% was faster than that of the contrast level of 13.5% and that of the contrast level of 11.5% ( ps  < .01 for all comparisons). The difference between the contrast levels of 13.5% and 11.5% was not significant. There was a main effect of social condition [ F (2, 18) = 4.005, p  = 0.04, \( {\eta}_p^2 \) = 0.31]. Post hoc comparison showed that the mean RT of the better observer was faster than that of the worse observer ( ps  < .05). However, there were no significant differences between the performance of the collaboration and the better observer or worse observer. The two-way interaction did not reach the significance level.

For accuracy, the results showed that a significant main effect of contrast level [ F (3, 54) = 560.53, p  < 0.001, \( {\eta}_p^2 \) = 0.97]. Post hoc comparison showed that accuracy increased as the contrast level increased and that the differences between every two contrast levels were all significant ( ps  < .01 for all comparisons). Moreover, there was a significant main effect of social condition [ F (2, 18) = 3.936, p  = 0.04, \( {\eta}_p^2 \) = 0.30]. Post hoc comparison showed that the accuracy of the worse observer was lower than that of the better observer ( ps  < .05). However, there were no observable, significant differences between collaboration and the better observer or worse observer. The two-way interaction did not reach the significance level.

Detection sensitivity

Figure  3 plots the relationship between the relative sensitivity between the worse observer and the better observer and the accuracy-based collective effect (i.e., S dyad / S max ). Although we observed a slight trend of positive correlation, as found in the Bahrami et al. ( 2010 ) study, this positive correlation did not reach the significance level ( R 2 = 0.16, slope = 0.58, p  = 0.37). The non-significant result may have occurred because of the restriction of range—namely, in the present experiment, the relative detection sensitivities of all dyads were above 0.6.

Plot of the accuracy-based collective effect ( Sdyad / Smax ) as a function of relative detection sensitivity ( Smin / Smax ). The red line represents the regression line

Capacity coefficient Footnote 8

Figure  4 shows the plot of the capacity coefficient function for each dyad. The capacity functions are plotted according to the level of the accuracy-based collective effect (i.e., S dyad / S max ) to reveal the relationship between the capacity level and the accuracy-based collective effect. The brightness level represents the level of the accuracy-based collective effect. Our visual inspection indicated that all dyads were of supercapacity with all C AND ( t ) greater than 1 for all times t . However, we did not find a robust relationship between the capacity values and the level of the accuracy-based collective effect.

Plot of the capacity coefficient function for each dyad in Experiment 1. The capacity functions were plotted by the level of the accuracy-based collective effect represented by its brightness level

Assessment function

Figure  5 shows the A AND ( t ) for all four response types; the functions were plotted according to the level of the accuracy-based collective effect. The assessment functions for each response type can be summarized as follows:

For the correct and fast responses, values of A AND ( t ) were consistently greater than 1, suggesting that correct collaborative responses were faster and more frequent than expected—that is, supercapacity processing.

For the correct and slow responses, values of A AND ( t ) were greater than 1 at the faster RTs and were then less than 1 at the slower RTs. This indicates that correct and slow responses made after time t were more probable than expected at faster RTs but less probable than expected at slower RTs.

For the incorrect and fast responses, the results showed that most dyads made more incorrect responses by time t at the faster RTs; however, incorrect and fast responses were less probable than expected at the slower RTs.

For the incorrect and slow responses, values of A AND ( t ) were less than 1 for all times t , indicating that fewer incorrect and slow responses were observed than the expectation from the UCIP model.

Plots of the assessment function of workload capacity in Experiment 1. The functions were colored by the level of accuracy-based collective effect. a correct and fast, b correct and slow, c incorrect and fast, and d incorrect and slow

To sum up, the results of A AND ( t ) indicate that the collaborative performance was processed more efficiently than the predicted baseline model, suggesting supercapacity processing. However, similar to the results of C AND ( t ), we did not find a strong relationship between A AND ( t ) and the collective effect indicated by relative detection sensitivity. Thus, the current results did not provide evidence in support of the relationship between time-based measures and accuracy-based measures.

fPCA and quantitative comparison

We employed fPCA with the varimax rotation to decompose C AND ( t ) and A AND ( t ) into several component functions and plotted the factor score of each component against the accuracy-based collective effect to investigate the relationship between the component function and the accuracy-based collective effect. However, we failed to find a robust relationship between the component functions and the accuracy-based collective effect—that is, all the correlations did not reach the significant level ( ps  > .05). Due to the non-significance, we do not present the results in the main text. Please refer to the Supplementary material ( S.2 ) for detailed descriptions of the fPCA results.

In Experiment 1, we conducted a two-interval forced-choice oddball detection task as employed in the study by Bahrami et al. ( 2010 , 2012a , b ) but with several modifications on the test procedure. We measured both accuracy and RT, from which we estimated the accuracy-based and time-based collective effects, respectively. Our accuracy results were consistent with the Bahrami et al. ( 2010 ) findings. That is, when two individuals had similar detection sensitivity ( S min / S max > 0.8), we observed a collective benefit with S dyad / S max larger than 1. Note that the sensitivity measure did not reveal a collective benefit in all dyads. In four of seven dyads, S dyad / S max was larger than 1, while in the other three, the effect was less than or equal to 1.

In addition, the collective benefit was inferred from the RT data in three ways. First, at the mean RT level, we found no significant difference between the mean RT of the collaboration condition and that of the worse or better observer, suggesting a lack of collective benefits at the mean RT level. Second, the observed C AND ( t ) values which were greater than 1 for all times t , suggested supercapacity for all dyads—namely, collective benefits were consistently observed for all pairs. Third, when RT and accuracy were combined, the results of A AND ( t ) again supported supercapacity processing. In particular, the results of A AND ( t ) with respect to correct and fast responses showed that dyads made correct and faster responses more frequently than expected, suggesting supercapacity processing. On the other hand, the correct and slower responses were more probable than expected at the faster RTs, also implying supercapacity processing. For the two types of incorrect responses, the results of A AND ( t ) suggested that dyads were more efficient since fewer incorrect responses were made.

It is interesting to observe that, despite the various applied measures that demonstrated the collective effect, the time-based and accuracy-based measures did not show a robust correlation between the two. Here, we came up with two likely explanations for the non-significant correlational results. First, we consider a likely case of the range of restriction. Participants in the present study had a similar detection sensitivity that restricted the full range of possible sensitivities in the general population, which could lead to non-significant sample correlations. To ameliorate a possible restricted range issue in Experiment 2, we introduced a stimulus noise; an extra background white noise was added on top of the visual display of one of the participants, thus heightening the relative differences in detection sensitivity between the two participants. Second, we suspect that there is another reason why the procedure of Experiment 1 was not sensitive enough to measure the time-based collective benefit. In the current setting, participants’ RT was recorded from the onset of a question mark. However, the participants may have already made their decisions while the two intervals were presented; therefore, the RT may reflect only the motor execution time rather than the information accumulation and decision time. To ameliorate this issue and correctly measure the decision time in Experiment 2, we replaced the 2IFC oddball detection task with a yes/no Gabor detection task. With a better estimation of the decision time, we expect to find a robust relationship between the time-based and accuracy-based measures.

Experiment 2

In Experiment 2, a yes/no Gabor detection task was conducted to enable a better estimation of the information accumulation time and decision time. In addition, we manipulated the transparency of the noise mask that was superimposed on the target stimulus, to manipulate the detection difficulty and thereby estimate the psychometric function. The experiment could be separated into Experiments 2a and 2b depending on whether an additional background noise was introduced to one of the participant’s displays. In Experiment 2a, dyads were tested without any additional background noise such that participants might have had similar detection sensitivity. In Experiment 2b, one of the participant’s displays was covered by additional background noise such that participants might have had different detection sensitivities. This manipulation could solve the problem of the restriction of range raised by Experiment 1 and allowed us to observe the effect of the relative detection sensitivity on collaborative performance. Following the Bahrami et al. ( 2010 ) study, we expected that relative detection sensitivity would reduce the collective benefit, which could be observed by both RT- and accuracy-based measures.

Twenty-six (18 male and eight female; age: 21.2 ± 2.48 years) and 20 (15 male and five female; age: 22.4 ± 1.55 years) undergraduate students at National Cheng Kung University volunteered to participate in Experiments 2a and 2b, respectively. They were randomly assigned and paired into 13 and 10 dyads, respectively. All the participants were right-handed and signed a written informed consent form prior to the experiment. Upon completion of the experiment, each participant received either a total of NTD 140 per hour or class bonus course credits.

The procedure was similar to that used in Experiment 1, though with several modifications. Figure  6 shows an illustration of the trial procedure. The task was to detect the presence of a target as accurately and quickly as possible. The stimulus was presented until a participant delivered a response or 5 s had elapsed. Unlike in Experiment 1, RT had been recorded since the test stimulus was presented. Another difference between Experiments 1 and 2 was that the cue, which informed the participants as to who was going to respond, was the color of the frame of the test stimulus.

An illustration of the experimental procedure in Experiment 2

The size of the test stimulus was 256 (horizontal) × 256 (vertical) pixels and the target stimulus was a vertically oriented Gabor patch (standard deviation of the Gaussian envelope: 0.45°; spatial frequency: 1.049 cycles/°; contrast: 10%). In half of the trials, the target was presented and the stimulus was superimposed by a white noise mask with a different degree of transparency to manipulate the detection difficulty. The degree of transparency was adjusted by the value of the alpha channel, i.e., a color component that represents the degree of transparency (or opacity) of a color, with a value ranging from 0 (high transparency) to 1 (low transparency). Hence, if the alpha value was high, it would become difficult to detect the target. Specifically, in Experiment 2a, both participants received the same level of noise, with the alpha value being either 0.81, 0.85, 0.865, 0.885, or 0.9. In Experiment 2b, the alpha value was 0.69, 0.77, 0.83, 0.87, or 0.9. The alpha values were slightly different from those used in Experiment 2a because this adjustment can create a better estimation of the slope of the psychometric function. The critical difference between Experiments 2a and 2b was that extra background noise was introduced into one of the participant’s displays with an alpha value of 0.5. This manipulation would result in relative detection sensitivity—namely, the participant with the extra background noise would have more difficulty detecting the target than would the participant without any extra background noise. Figure  7 shows an example of how we manipulated the transparency of the noise mask and the extra background noise in Experiment 2.

An illustration of stimuli with different levels of noise in Experiment 2

Similar to Experiment 1, participants were required to participate in three cooperation conditions. Each condition was repetitively tested in two sessions in order to collect enough data points. In each session, participants first performed a practice block of 60 trials and then 10 blocks of formal trials. Each block consisted of 2 (presence or absence of the target) × 5 (difficulty levels) × 6 (trials per combination).

The data analysis procedure was similar to that used in Experiment 1, though with several modifications. For example, for the ANOVA and detection sensitivity analyses, the within-subject factor was the difficulty level, which was manipulated by the alpha value. For the psychometric function analysis, we estimated the probability of responses that the target was presented, denoted as P (− c ), where c denotes the alpha value.

Table  3 presents the mean correct RTs and accuracy for all the combinations of the social conditions and difficulty levels in Experiments 2a and 2b.

Experiment 2a

For RT, the results showed a main effect of social condition [ F (2, 36) = 4.628, p  = 0.02, \( {\eta}_p^2 \) = 0.20]. Post hoc comparison showed that the mean RT of a better observer was faster than that of a worse observer ( ps  < .05); however, there was no observable significant difference between collaboration and the worse or better observer. Moreover, the results showed a significant main effect of difficulty level [ F (4, 144) = 154.23, p  < 0.001, \( {\eta}_p^2 \) = 0.81]. Post hoc comparison showed that the mean RT was slower as the difficulty level increased and that the differences between every two difficulty levels were all significant ( ps  < .01 for all comparisons). The two-way interaction did not reach the significance level.

In terms of accuracy, the results showed a significant main effect of the social condition [ F (2, 36) = 27.59, p  < 0.001, \( {\eta}_p^2 \) = 0.61]. Post hoc comparison showed that the accuracy of collaboration was higher than that of the better observer ( ps  < .05) and that of the worse observer ( ps  < .01). Moreover, the accuracy of the better observer was higher than that of the worse observer ( ps  < .01). The results showed a significant main effect of difficulty level [ F (4, 144) = 345.49, p  < 0.001, \( {\eta}_p^2 \) = 0.91]. Post hoc comparison showed that accuracy declined as the difficulty level increased and that the differences between every two difficulty levels were all significant ( ps  < .01 for all comparisons). The interaction effect was also significant [ F (8, 144) = 4.635, p  < 0.001, \( {\eta}_p^2 \) = 0.20]. Post hoc comparison showed that the accuracy of the worse observer was lower than that of the better observer and collaboration when the alpha value was 0.85, 0.865, 0.885, 0.9 ( ps  < .01) but the differences were not significant when the alpha value was 0.81 (i.e., the easiest condition). Moreover, the results showed that the accuracy of collaboration was significantly higher than that of the better observer when the alpha values were 0.885 ( ps  < .01) and 0.85, 0.865, and 0.9 ( ps  < .05), suggesting a collective benefit; however, the difference was not significant when the alpha value was 0.81.

Experiment 2b

For RT, the results showed a main effect of difficulty level [ F (4, 108) = 34.780, p  < 0.001, \( {\eta}_p^2 \) = 0.56]. Post hoc comparison showed that the mean RTs were slower when the alpha values were 0.83 and 0.87 than when the alpha values were 0.69 and 0.77 ( ps  < .01). Moreover, the mean RT was slower when the alpha value was 0.9 than when the alpha values were 0.69, 0.77, and 0.83 ( ps  < .01 for all comparisons). No other effects reached the significance level.

In terms of accuracy, the results showed that a main effect of difficulty level [ F (4, 108) = 121.59, p  < 0.001, \( {\eta}_p^2 \) = 0.82]. Post hoc comparison showed that accuracy declined as the difficulty level increased and that the differences between every two difficulty levels were significant ( ps  < .01) except for the difference between the two easiest difficulty levels (i.e., the alpha values were 0.69 and 0.77). Moreover, there was a significant main effect of social condition [ F (2, 27) = 9.324, p  < 0.001, \( {\eta}_p^2 \) = 0.41]. Post hoc comparison showed that the accuracy of the worse observer was lower than that of the better observer and collaboration ( ps  < .01); however, the difference between collaboration and the better observer was not significant. The interaction effect was significant [ F (8, 108) = 2.763, p  = 0.008, \( {\eta}_p^2 \) = 0.17]. Post hoc comparison showed that the accuracy of the worse observer was lower than that of the better observer only when the alpha value was 0.77 or 0.83 ( ps  < .01). Moreover, the accuracy of the worse observer was lower than that of collaboration when the alpha value was 0.77 ( ps  < .05) and 0.83 and 0.87 ( ps  < .01). However, there were no observable differences between collaboration and the better observer.

Figure 8 plots the relationship between the relative detection sensitivity between individuals and the accuracy-based collective effect (i.e., S dyad / S max ). The results were similar to what Bahrami et al. ( 2010 ) observed and the correlation was significantly positive ( R 2 = 0.62, slope = 1.34, p  < 0.001). Fitted by linear regression, the results suggested that the cutoff for having a collective benefit ( S dyad / S max > 1) was about 0.57.

Plot of the accuracy-based collective effect ( Sdyad / Smax ) as a function of relative detection sensitivity ( Smin/Smax ). The red line represents the regression line

Capacity coefficient

Figure 9 shows the plot of the capacity coefficient function for each dyad. The capacity functions were plotted according to the level of the accuracy-based collective effect to reveal the relationship between the capacity level and the accuracy-based collective effect. The brightness level represents the level of the accuracy-based collective effect. Our visual inspection revealed that most dyads were of supercapacity for all time t and that few of them were of limited capacity at the slower RTs. However, similar to Experiment 1, we did not find a robust relationship between the capacity functions ( C AND ( t )) and the level of the accuracy-based collective effect ( S dyad / S max ).

Plot of the capacity coefficient function for each dyad in Experiment 2. The capacity functions were plotted by the level of the accuracy-based collective effect represented by the brightness level

Figure  10 shows the A AND ( t ) for all four response types. The functions were plotted according to the level of the accuracy-based collective effect. The assessment functions for each response type can be summarized as follows:

For the correct and fast response, values of A AND ( t ) were consistently greater than 1, suggesting that correct group responses were faster and more frequent than expected (i.e., supercapacity processing). Importantly, our visual inspection revealed that A AND ( t ) systematically increased as a function of the accuracy-based collective effect.

For the correct and slow responses,  A AND  ( t ) values were above 1 at the faster RTs and reached an asymptote at the value of 1. This suggests that correct and slow responses, made after time t , were more probable than expected at the faster RTs.

For the incorrect and fast responses, the results showed that most dyads delivered incorrect responses by time t more frequently than expected at the faster RTs; however, incorrect and fast responses were less probable than expected at the slower RTs. Moreover, the results suggested that A AND ( t ) values decreased upon an increase in the accuracy-based collective effect.

For the incorrect and slow responses,  A AND  ( t ) for most dyads was consistently less than 1, indicating that fewer incorrect and slow responses were observed than the expectation from the UCIP model. Similar to the results of the incorrect and fast responses, A AND ( t ) values decreased upon an increase in the accuracy-based collective effect.

Plots of the assessment function of workload capacity in Experiment 2. The functions were plotted by the level of accuracy-based collective effect. a correct and fast, b correct and slow, c incorrect and fast, and d incorrect and slow

To sum up, the pattern of A AND ( t ) was similar to that of Experiment 1, indicating that the group decision-making was of supercapacity. We found that the dyads with a higher accuracy-based collective effect tended to have larger A AND ( t ) values for correct and fast responses and smaller A AND ( t ) values for incorrect responses. Thus, the results converged to suggest the existence of a positive correlation between time-based and accuracy-based measures.

When we correlated the factor score of each component with the accuracy-based collective effect (please refer to the Supplementary material ( S.3 ) for detailed results), we found a significant correlation for only the first component of A AND ( t ) of the correct and fast responses. In Fig.  11 a, the left panel shows the capacity function of the first component and the mean capacity function. The right panel shows the contrast function (i.e., the mean capacity function subtracted from the first principal component function). The first principal component function accounts for 46.0% of the variance and indicates a general increase in the capacity values at the faster RTs. In other words, the first component captures the profile of supercapacity processing. Figure  11 b shows the scatter plot of the loading of the first component and the accuracy-based collective effect; its correlation reached the significance level ( R 2 = 0.19, slope = 0.65, p  < 0.05).

Results of the first principle component of A(t) for correct and fast responses. a Mean capacity function and the first principal component function (left panel) and the contrast function (right panel). b Plot of the time-based measure (the factor score of the first component) against the accuracy-based measure ( Sdyad/Smax ). The red line represents the regression line

In Experiment 2, we conducted a yes/no Gabor detection task. Both accuracy and RT were collected to estimate a potential collective benefit. The sensitivity results were consistent with the Bahrami et al. ( 2010 ) findings. That is, when the two participants had similar detection sensitivity (i.e., S min / S max > 0.57), the collective benefit was observed with S dyad / S max greater than 1. By contrast, when two individuals’ sensitivities were dissimilar ( S min / S max ≤ 0.57), the collective cost was observed.

Similar to Experiment 1, we can infer the collective benefit from the RT data in three ways. First, at the mean RT level, the results were similar to those of Experiment 1, suggesting that there was no collective benefit. Second, the results of C AND ( t ) showed supercapacity with capacity values greater than 1 for all times t for most dyads; however, for some dyads, the capacity went from supercapacity to limited capacity as a function of RT. This implied that collective benefit can be found in most dyads but that only a few dyads had collective benefits at the faster RTs. Third, when one combined both RT and accuracy, the results of A AND ( t ) again supported supercapacity processing and showed evidence of the ordered relationship between accuracy-based collective effect and A AND ( t ). In particular, for correct and fast responses, the results of A AND ( t ) showed that dyads made correct and faster responses more frequently than expected, thereby suggesting supercapacity processing. In addition, we observed that the dyads with a higher accuracy-based collective effect showed larger values of A AND ( t ) for the correct and fast responses. On the other hand, for the correct and slow responses, the values of A AND ( t ) were larger than 1 at the faster RTs, also implying supercapacity processing. For the two types of incorrect responses, the results of A AND ( t ) suggested that dyads were of supercapacity because there were fewer incorrect responses than the expectation from the UCIP model.

When we correlated the time-based and accuracy-based measures, we found a significant positive correlation between accuracy-based collective effect and the first principal component of A AND ( t ) of the correct and fast responses. The first principle component indicates an increase in A AND ( t ) for correct and fast responses at the faster RTs, implying that dyads with higher factor scores had a higher level of supercapacity. Hence, the correlation shows that dyads with a higher collective effect in the accuracy-based measure tended to have a larger supercapacity in the time-based measure.

General discussion

Summary of the present findings.

In the present study, we examined the collective effect in perceptual decision-making tasks, such as a two-interval forced-choice oddball detection task (Experiment l) and a yes/no Gabor detection task (Experiment 2). We followed the Bahrami et al. ( 2010 ) study to estimate the accuracy-based collective effect through a comparison of the collaborative detection sensitivity to the sensitivity of the better individual. We utilized SFT to analyze the decision efficiency by comparing the performance of the collaborative condition to the capacity baseline, which assumes that group members work independently in a non-collaborative fashion. We investigated the relationship between the accuracy-based and timed-based measures of the collective effect and proposed that the assessment function of workload capacity A AND ( t ) can be regarded as a novel and diagnostic measure for quantifying group decision-making efficiency.

To address our first goal of replicating the effect of sensitivity similarity on joint decisions, we first conducted the accuracy analyses. The results of Experiment 1 showed that four of the dyads had S dyad / S max  ≥ 1, suggesting collective benefit. The other three dyads had S dyad / S max  < 1, suggesting collective cost. There was a slight trend of a positive correlation between the accuracy-based collective effect and the relative detection sensitivity between observers; however, it did not reach the significance level. The inspection of the measured relative differences in the subjects’ detection sensitivities implied the possibility of a restricted range. The random subject selection had formed dyads of subjects whose detection sensitivities were, overall, more similar than dissimilar. The restriction of the range is a reasonable explanation for the failure to fully replicate the Bahrami et al. ( 2010 ) findings. To amend the issue of the restriction of range, Experiment 2 increased the relative differences in detection sensitivity between observers by introducing an extra background noise to one of the participants. In Experiment 2, we found that 15 pairs of participants had S dyad / S max  ≥ 1, suggesting a collective benefit, while the other eight pairs had S dyad / S max  < 1, suggesting a collective cost. It is noted that the cutoff point for observing the collective benefit is when the relative detection sensitivity ( S min / S max ) equals 0.57. As expected, the effect of inducing more individual variability in detection sensitivities led to the replication of the main findings of Bahrami et al.: a significant correlation between the accuracy-based collective effect and the relative detection sensitivity was observed. This more strongly supported findings of Bahrami et al. that the collective benefit was only observable when the group members had similar levels of detection sensitivity.

We further analyzed RT to address the second goal of testing whether RT and accuracy measures are consistent to draw similar conclusions about the collective effect. We used both C AND ( t ) and A AND ( t ) to quantify the RT-based collective effect. The results of Experiments 1 and 2 were consistent in suggesting that all the pairs had C AND ( t ) and A AND ( t ) of the correct and fast responses larger than 1, implying that they were all of supercapacity in group decision-making—that is, collaborative decision-making is always more efficient than non-collaborative decision-making, although collaboration can have both a benefit and a cost in terms of detection sensitivity (i.e., individual differences in group detection sensitivity). Though we did not observe a robust relationship between C AND ( t ) and the accuracy-based collective effect measured by detection sensitivity (i.e., S dyad / S max ), we found an ordered relationship between the accuracy-based collective effect ( S dyad / S max ) and A AND ( t ) (see Fig. 10 ). This suggests that both the time-based and accuracy-based measures can capture the collective benefit.

Our last goal is to establish the assessment function of workload capacity as a standard measure of group decision efficiency. It is important to note that while most of the A AND ( t ) analyses and applications are based on the position that A AND ( t ) provides a point estimate of capacity for a single dyad, A AND ( t ) is a function of response time and, as seen in Figs.  5 and 10 , its shape changes differently for correct/incorrect and fast/slow response times. One of the recent studies employed an advanced idea to analyze the shape of the A AND ( t ) function by using the fPCA analysis (Burns et al., 2013 ; Houpt, Blaha, et al., 2014 ). We also decided to utilize the fPCA analysis on A AND ( t ) to further investigate whether different component functions would provide additional insights into the accuracy measures as well as into the nature of the A AND ( t ) function. Further analysis of the data of Experiment 2 via fPCA showed that it is the first component of A AND ( t ), i.e., an increase of A AND ( t ) values at the faster RTs, which positively correlated with S dyad / S max . That is, the accuracy-based collective benefit may imply a decision-processing advantage (i.e., supercapacity processing and more efficient processing) especially for the faster RTs. The combined fPCA and A AND ( t ) analyses, as a novel and diagnostic tool, offer a potentially promising direction for the study of group decision efficiency, as well as for diagnosing individual differences in terms of how efficiently and accurately group members work together to make a final decision.

To sum up, the current study was motivated primarily by a desire to provide more evidence through which to learn about the underlying processing mechanism of group decision-making. Our study extended the Bahrami et al. ( 2010 ) psychophysical method by simultaneously recording the accuracy and RT when participants made group decisions. Following a careful examination of the information accumulation time and decision time in the yes/no Gabor detection task in Experiment 2, several major findings and novel contributions can be noted. First, we replicated Bahrami et al. ( 2010 ) main findings, which showed that collaboration benefits detection sensitivity only when group members have similar detection sensitivity. By contrast, collaboration hinders detection sensitivity when individuals’ detection sensitivities are dissimilar. Second, following the SFT approach, both results for group decision efficiency, C AND ( t ) and A AND ( t ), showed that collaborative decision-making is always processed more efficiently than is non-collaborative decision-making, with evidence of supercapacity processing for all pairs throughout Experiments 1 and 2. These results implied that participants were able to utilize their partners’ confidence to boost and improve their decision-making. Third, we found no strong correlation between the accuracy-based collective effect and  C AND ( t ), which considers only the processing efficiency of the correct responses. Alternatively, we found a robust relationship between the accuracy-based collective effect and A AND ( t ), which considers both RT and accuracy at the same time. It is notable that the first component of the A AND ( t ) of the correct and fast responses, which captures the signature of an increase in capacity at the faster RTs, is significantly correlated with the accuracy-based collective effect. Based on the converging evidence, we may suggest that the supercapacity processing of the correct and fast trials is closely related to the source of the accuracy-based collective effect, which could imply the sharing of similar mental mechanisms.

Processing mechanism underlying group decision-making

In the context of group decision-making, supercapacity processing suggests more efficient processing for collaborative decisions in which participants can exchange their confidence than for non-collaborative decisions in which participants work independently without any communication. The supercapacity processing indicates that group performance is better than the UCIP predictions, which assumes that the group decision-making system is an unlimited-capacity, independent, and parallel processing system. The evidence of supercapacity can be used to rule out the statistical facilitation account. This means that the collective benefit is not likely to be an artifact of an increase in the number of individual units.

The results of supercapacity processing may suggest two possible mechanisms. First, it implies that, in a group, all members’ processing efforts, or their individually collected evidence, are accumulated and combined through a series of social interactions. This model for group interaction is equivalent to the cognitive processing model known as coactive processing, which is used to characterize a combination of processing outcomes of several mental structures (Houpt & Townsend, 2011 ; Schwarz, 1989 , 1994 ; Townsend & Nozawa, 1995 ). Similarly, as in the coactive processing framework, a collective benefit could arise as the result of a “weighted-and-sum” principle of information integration of the individual contributions. Following this analogy, a “coactive” group member can weigh and integrate their confidence based on their perceptual observation with their partner’s confidence by considering the partner’s credibility based on the previous experience acquired from the trial-by-trial feedback. The details of such a process of combining opinions are beyond the scope of the current study; however, it provides a direction for investigation in future research. Second, supercapacity processing may imply information-sharing between parallel channels (Eidels, Houpt, Altieri, Pei, & Townsend, 2011 ; Miller, 1982 ; Mordkoff & Yantis, 1991 ; Townsend & Wenger, 2004 ). We were able to observe several types of social interactions while participants worked together on making decisions. In the case of non-verbal communication, in which a participant could see their partner showing confidence in one of the responses, this form of interaction boosted the participant’s accumulation of decision information toward that response. While the current evidence is still insufficient for making strong conclusions about various forms of information exchange between the partners in the tasks, and about whether the present results are due to the coactivation or facilitatory interaction between group members, future studies are encouraged to further explore the difference between the two possibilities.

It is generally accepted that efficient group decision-making may have occurred based on the following considerations: (1) What is the form of social interaction? (i.e., verbal or non-verbal), (2) Is the partner trustworthy or reliable? (i.e., the presence or lack of feedback), and (3) What is the decision input from the partner? (i.e., is the partner’s confidence in the current trial high or low?). First, previous studies have compared the effect of collaboration between verbal communication and non-verbal communication on group detection sensitivity (e.g., Bahrami et al., 2012a ). The results showed that both forms of communication can enhance group detection sensitivity through the exchange of decision evidence and decision confidence, respectively. However, in the current context of perceptual decision-making tasks, it is difficult to demonstrate the verbal communication effect on the group decision RT because verbal communication usually takes longer. According to our data, for most trials, it takes less than 2 s to complete a perceptual decision. It seems that such a short time interval is not sufficient for members to communicate with each other verbally. As a result, in our study, we considered only the non-verbal form of collaboration (i.e., internal confidence estimate). Future works are encouraged to modify the tasks or increase the tasks’ difficulty, which could create sufficient time for more verbal communication between group members and, as such, allow this to be explored in more detail. Second, it has been shown that participants can utilize a partner’s confidence to increase group detection sensitivity only when feedback is provided (Bahrami et al., 2010 ; Sorkin et al., 2001 ). The reason why the response feedback is important in observing the collective benefit is that a participant can learn whether their partner is trustworthy or reliable. If a partner is consistently making a correct decision, the participant would assign a higher weight to their partner’s confidence. By contrast, if the partner is consistently making an incorrect decision, the participant would assign a lower weight to their confidence. The feedback would help participants adjust the weighting process while integrating the partner’s confidence into the practice of making a group decision. However, to our knowledge, collaborative decision-making efficiency has not been tested without a presentation of the response feedback; therefore, we are still unclear as to whether presenting feedback would, indeed, enhance group decision efficiency as compared to the no-feedback condition. Future studies are encouraged to manipulate the presence/absence of feedback to test its effect on group decision efficiency. Third, if a partner is very evidently confident in their decision, participants may simply follow their partner’s opinions to make a decision or give very high weight to the partner’s confidence. Thus, the decision-making process would become very efficient. On the other hand, if the participant received a low-confidence input, they would need to make a decision by integrating their decision evidence with their partner’s decision confidence, which might slow down the processing for the faster decision-maker. As a result, the efficient group decision may have occurred because the partner’s high-confidence input boosted the decision efficiency. Although we have not yet systematically examined how high/low confidence affects group decision efficiency (due to a limited number of trials), we are open to this possibility and leave it for future study. It is also notable that possibilities (2) and (3) may interactively affect group decision efficiency. That is, when a second partner is very reliable and consistent in making a correct decision, the first partner may simply follow the second partner’s decision to offer a response regardless of whether the confidence is high or low. On the other hand, if the partner is not reliable, the first partner may still need to rely on their perception and integrate it with the second partner’s decision confidence to make a final decision, even in the high-confidence trial. More research with sophisticated analyses would be appropriate to investigate the interaction effects.

Capacity coefficient and assessment function of workload capacity

It is intriguing to note that, in our study, the accuracy-based measure showed both the benefit and cost of collaboration, whereas the time-based measures (i.e., C AND ( t ) and A AND ( t )) showed only the benefit of collaboration. A reasonable question to ask is: why are the two measures not tightly connected? If one must choose, which measure should be regarded as being the better one in terms of diagnosing the collective benefit? In the introduction, we mentioned a serious challenge to data interpretation if either measure is considered in isolation: that there may be a time-accuracy tradeoff in making a group decision. That is, a group may make a very fast decision by sacrificing the decision accuracy, or it may make an accurate group decision by slowing down its information accumulation process because social interaction takes time. Therefore, the examination of only one type of measure cannot clearly reveal whether collaboration, indeed, presents an advantage. In a surprising turnaround, we could not find strong evidence for the time-accuracy tradeoff; when we considered both measures at the same time, the correlation between the accuracy-based and time-based collective effects was not strong enough (i.e., there was a correlation only between the accuracy-based collective effect and  A AND ( t )). The absence of evidence of a strong correlation may imply that the two measures are related to different aspects of collective effects. Two possibilities can be considered. First, the baselines for computation and inference of the accuracy-based and time-based collective effects are different. The accuracy-based collective effect is quantified by dividing group detection sensitivity by the sensitivity of the better observer (i.e.,  S dyad / S max ). On the other hand, the time-based collective effect is quantified by comparing group decision efficiency to the UCIP baseline, which is generated from the two individuals’ decision efficiencies when they work independently. Therefore, we would argue that, in the previous condition, the collective benefit is defined only when the performance exceeds the better individual, while, in the latter condition, the collective benefit is defined by the group outperforming the integration of the two individuals’ decision efficiencies. The difference in baselines might explain why both benefits and costs are found in the former condition instead of in the latter condition. The second possibility concerns the cooperation rule for group decisions. For the accuracy-based collective effect, the cooperation rule is not considered. Therefore, it is unclear whether a group decision that outperforms the individual decisions occurred because the group decision follows a “winner-takes-all” rule by making a choice based on the opinion of the more confident member or because the group exhaustively processed the two members’ opinions and made a decision by integrating decision evidence. For the time-based measure, we assessed the AND capacity by assuming that group decisions always take place after all the decision evidence is exhaustively processed. That is, both members’ confidence (i.e., the participant’s own and that of their partner) is always considered when making the final decision. In comparison, Brennan and Enns ( 2015 ) used the test of the violation of the race-model inequality; the underlying assumption of the test of the race-model inequality is an OR cooperation rule. However, we believe that the AND rule is a better way to characterize how group members make a coherent decision in the present context of perceptual decision-making. The different assumptions in the decision rules may also explain why we could not find a robust relationship between the accuracy-based and time-based collective effects.

To conclude, we argue that A AND ( t ) could serve as a novel and diagnostic measure of group decision efficiency. A AND ( t ) retains the property of accuracy but can also infer the processing efficiency under both correct and incorrect responses. In addition, A AND ( t ) can provide information about the dynamic changes in processing efficiency as a function of RT. Also, A AND ( t ) can select a specific cooperation rule for the analysis of, and inferences about, group decision efficiency. Therefore, we strongly recommend that researchers test the collective effect by using the A AND ( t ) measure.

Availability of data and materials

The datasets and codes of the current study are available at https://github.com/hanekaze/A-new-measure-of-group-decision-making-efficiency .

The “wisdom of crowds” does not occur only in a context of working as a group. It can be used to refer to the averaging effect while working in isolation.

It is commonly assumed that the first-termination rule is adopted in an OR task, when multiple units are analyzed simultaneously, and in which a decision is determined when either decision unit reaches the decision criterion. However, it should be noted that group decision-making does not necessarily follow the first-termination rule; other forms of decision rules (e.g., majority rule, consensus) may be adopted for group decisions.

The coactive models assume the information from multiple decision units is summed into a single unit, and that the sum of information drives the total activation which is then compared to a threshold for decision-making (e.g., Schwarz, 1989 , 1994 ; Townsend & Nozawa, 1995 ; Houpt & Townsend, 2011 ).

The detection sensitivity is defined as the maximum slope of the psychometric function. With the steeper the slope, the higher the detection sensitivity.

In the present study, we used the AND capacity rather than the OR capacity to quantify the group decision efficiency. The AND capacity assesses information-processing efficiency when an exhaustive stopping rule is adopted (i.e., the maximum-time rule) whereas the OR capacity assess the efficiency when a self-terminating rule is adopted (i.e., the minimum-time rule). The reason is that we believe that A group decision is made by reaching the consensus through non-verbal communication and the independent-race model following a winner-takes-all rule is not likely to occur in the current context.

It is noted that the current testing procedure is different from that used by Bahrami et al. ( 2010 ). In their study, a joint decision was required only when the participants made inconsistent individual decisions. By contrast, in the current setting, all the participants were required to perform the individual condition, non-collaborative condition, and collaborative condition, tested in separate blocks.

All the participants did not know each other; thus, the friendship cannot account for the group performance.

We conducted a one-way repeated measures analysis of variance (ANOVA) to test the effect of contrast level on the collective effect but the effect did not reach the significance level. Hence, we aggregated all the contrast levels to estimate the capacity coefficient function and the assessment function.

Altieri, N., & Yang, C.-T. (2016). Parallel linear dynamic models can mimic the McGurk effect in clinical populations. Journal of Computational Neuroscience , 41 (2), 143–155. https://doi.org/10.1007/s10827-016-0610-z .

Bahrami, B., Olsen, K., Bang, D., Roepstorff, A., Rees, G., & Frith, C. (2012a). Together, slowly but surely: The role of social interaction and feedback on the build-up of benefit in collective decision-making. Journal of Experimental Psychology. Human Perception and Performance , 38 (1), 3–8. https://doi.org/10.1037/a0025708 .

Bahrami, B., Olsen, K., Bang, D., Roepstorff, A., Rees, G., & Frith, C. (2012b). What failure in collective decision-making tells us about metacognition. Philosophical Transactions of the Royal Society, B: Biological Sciences , 367 (1594), 1350–1365. https://doi.org/10.1098/rstb.2011.0420 .

Bahrami, B., Olsen, K., Latham, P. E., Roepstorff, A., Rees, G., & Frith, C. D. (2010). Optimally interacting minds. Science , 329 (5995), 1081–1085. https://doi.org/10.1126/science.1185718 .

Barr, S., & Gold, J. M. (2014). Redundant visual information enhances group decisions. Journal of Experimental Psychology: Human Perception and Performance, 40(6), 2124–2130. https://doi.org/10.1037/a0038224 .

Brennan, A. A., & Enns, J. T. (2015). When two heads are better than one: Interactive versus independent benefits of collaborative cognition. Psychonomic Bulletin & Review , 22 (4), 1076–1082. https://doi.org/10.3758/s13423-014-0765-4 .

Burns, D. M., Houpt, J. W., Townsend, J. T., & Endres, M. J. (2013). Functional principal components analysis of workload capacity functions. Behavior Research Methods , 45 (4), 1048–1057. https://doi.org/10.3758/s13428-013-0333-2 .

Chechile, R. A. (2003). Mathematical tools for hazard function analysis. Journal of Mathematical Psychology , 47 (5–6), 478–494. https://doi.org/10.1016/S0022-2496(03)00063-4 .

Chechile, R. A. (2011). Properties of reverse hazard functions. Journal of Mathematical Psychology , 55 (3), 203–222. https://doi.org/10.1016/j.jmp.2011.03.001 .

Clemen, R. T. (1989). Combining forecasts: a review and annotated bibliography. International Journal of Forecasting , 5 (4), 559–583. https://doi.org/10.1016/0169-2070(89)90012-5 .

Davis, J. H. (1969). Group performance . Reading: Addison-Wesley.

Google Scholar  

Davis, J. H. (1973). Group decision and social interaction: a theory of social decision schemes. Psychological Review , 80 (2), 97–125. https://doi.org/10.1037/h0033951 .

Donkin, C., Little, D. R., & Houpt, J. W. (2014). Assessing the speed-accuracy Trade-off effect on the capacity of information processing. Journal of Experimental Psychology: Human Perception and Performance , 40 (3), 1183–1202. https://doi.org/10.1037/a0035947 .

Eidels, A., Houpt, J. W., Altieri, N., Pei, L., & Townsend, J. T. (2011). Nice guys finish fast and bad guys finish last: facilitatory vs. inhibitory interaction in parallel systems. Journal of Mathematical Psychology , 55 (2), 176–190. https://doi.org/10.1016/j.jmp.2010.11.003 .

Fific, M., & Gigerenzer, G. (2014). Are two interviewers better than one? Journal of Business Research , 67 (8), 1771–1779. https://doi.org/10.1016/j.jbusres.2014.03.003 .

Fific, M., Nosofsky, R. M., & Townsend, J. T. (2008). Information-processing architectures in multidimensional classification: a validation test of the systems factorial technology. Journal of Experimental Psychology. Human Perception and Performance , 34 (2), 356–375. https://doi.org/10.1037/0096-1523.34.2.356 .

Gordon, K. (1924). Group judgments in the field of lifted weights. Journal of Experimental Psychology. Human Perception and Performance , 7 , 398–400. https://doi.org/10.1037/h0074666 .

Graham, J. R. (1996). Is a group of economists better than one? Than none? Journal of Business , 69 (2), 193–232. https://doi.org/10.1086/209688 .

Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics. New York: Wiley.

Heitz, R. P. (2014). The speed-accuracy tradeoff: methodology, and behavior. Frontiers in Neuroscience , 8 , ARTN 150. https://doi.org/10.3389/fnins.2014.00150 .

Houpt, J. W., Blaha, L. M., McIntire, J. P., Havig, P. R., & Townsend, J. T. (2014). Systems factorial technology with R. Behavior Research Methods , 46 (2), 307–330. https://doi.org/10.3758/s13428-013-0377-3 .

Houpt, J. W., Sussman, B. L., Townsend, J. T., & Newman, S. D. (2015). Dyslexia and configural perception of character sequences. Frontiers in Psychology, 6, Article 482. https://doi.org/10.3389/fpsyg.2015.00482 .

Houpt, J. W., & Townsend, J. T. (2011). An extension of SIC predictions to the Wiener coactive model. Journal of Mathematical Psychology , 55 (3), 267–270. https://doi.org/10.1016/j.jmp.2011.02.002 .

Houpt, J. W., Townsend, J. T., & Donkin, C. (2014). A new perspective on visual word processing efficiency. Acta Psychologica , 145 (0), 118–127. https://doi.org/10.1016/j.actpsy.2013.10.013 .

Ingvalson, E. M., & Wenger, M. J. (2005). A strong test of the dual-mode hypothesis. Perception & Psychophysics , 67 , 14–35.

Article   Google Scholar  

Little, D. R., Altieri, N., Fific, M., & Yang, C.-T. (2017) Systems factorial technology: A theory driven methodology for the identification of perceptual and cognitive mechanisms. New York: Academic Press.

Lorenz, J., Rauhut, H., Schweitzer, F., & Helbing, D. (2011). How social influence can undermine the wisdom of crowd effect. Proceedings of the National Academy of Sciences , 108 (22), 9020–9025. https://doi.org/10.1073/pnas.1008636108 .

Lorge, I., & Solomon, H. (1955). Two models of group behavior in the solution of eureka-type problems. Psychometrika , 20 (2), 139–148. https://doi.org/10.1007/bf02288986 .

Miller, J. (1982). Divided attention: evidence for coactivation with redundant signals. Cognitive Psychology , 14 (2), 247–279. https://doi.org/10.1016/0010-0285%2882%2990010-X .

Mordkoff, J. T., & Yantis, S. (1991). An interactive race model of divided attention. Journal of Experimental Psychology. Human Perception and Performance , 17 (2), 520–538. https://doi.org/10.1037/0096-1523.17.2.520 .

Ramsay, J. O., & Silverman, B. W. (2005). Functional data analysis (2nd ed.). New York: Springer.

Schwarz, W. (1989). A new model to explain the redundant-signals effect. Attention, Perception, & Psychophysics , 46 (5), 498–500. https://doi.org/10.3758/bf03210867 .

Schwarz, W. (1994). Diffusion, superposition, and the redundant-targets effect. Journal of Mathematical Psychology , 38 (4), 504–520. https://doi.org/10.1006/jmps.1994.1036 .

Sorkin, R. D., & Dai, H. (1994). Signal detection analysis of the ideal group. Organizational Behavior and Human Decision Processes , 60 (1), 1–13. https://doi.org/10.1006/obhd.1994.1072 .

Sorkin, R. D., Hays, C. J., & West, R. (2001). Signal-detection analysis of group decision making. Psychological Review , 108 (1), 183–203. https://doi.org/10.1037/0033-295X.108.1.183 .

Sorkin, R. D., West, R., & Robinson, D. E. (1998). Group performance depends on the majority rule. Psychological Science , 9 (6), 456–463. https://doi.org/10.1111/1467-9280.00085 .

Surowiecki, J. (2004). The wisdom of crowds: why the many are smarter than the few and how collective wisdom shapes business, economies, societies, and nations (1st ed.). New York: Doubleday.

Swets, J. A., Shipley, E. F., McKey, M. J., & Green, D. M. (1959). Multiple observations of signals in noise. The Journal of the Acoustical Society of America , 31 (4), 514–521. https://doi.org/10.1121/1.1907745 .

Townsend, J. T., & Altieri, N. (2012). An accuracy-response time capacity assessment function that measures performance against standard parallel predictions. Psychological Review , 119 (3), 500–516. https://doi.org/10.1037/a0028448 .

Townsend, J. T., & Eidels, A. (2011). Workload capacity spaces: a unified methodology for response time measures of efficiency as workload is varied. Psychonomic Bulletin & Review , 18 (4), 659–681. https://doi.org/10.3758/s13423-011-0106-9 .

Townsend, J. T., & Fific, M. (2004). Parallel versus serial processing and individual differences in high-speed search in human memory. Perception & Psychophysics , 66 (6), 953–962.

Townsend, J. T., & Nozawa, G. (1995). Spatio-temporal properties of elementary perception: an investigation of parallel, serial, and coactive theories. Journal of Mathematical Psychology , 39 (4), 321–359.

Townsend, J. T., & Wenger, M. J. (2004). A theory of interactive parallel processing: New capacity measures and predictions for a response time inequality series. Psychological Review , 111 (4), 1003–1035. https://doi.org/10.1037/0033-295x.111.4.1003 .

Van Zandt, T., & Townsend, J. T. (1993). Self-terminating versus exhaustive processes in rapid visual and memory search: An evaluative review. Perception & Psychophysics , 53 (5), 563–580.

Winkler, R. L., & Poses, R. M. (1993). Evaluating and combining physicians probabilities of survival in an intensive-care unit. Management Science , 39 (12), 1526–1543. https://doi.org/10.1287/mnsc.39.12.1526 .

Yamani, Y., Neider, M. B., Kramer, A. F., & McCarley, J. S. (2017). Characterizing the efficiency of collaborative visual search with systems factorial technology. Archives of Scientific Psychology, 5(1), 1-9. https://doi.org/10.1037/arc0000030 .

Yang, C.-T. (2011). Relative saliency in change signals affects perceptual comparison and decision processes in change detection. Journal of Experimental Psychology. Human Perception and Performance , 37 (6), 1708–1728.

Yang, C.-T., Altieri, N., & Little, D. R. (2018). An examination of parallel versus coactive processing accounts of redundant-target audiovisual signal processing. Journal of Mathematical Psychology , 82 , 138–158.

Yang, C.-T., Chang, T.-Y., & Wu, C.-J. (2013). Relative change probability affects the decision process of detecting multiple feature changes. Journal of Experimental Psychology. Human Perception and Performance , 39 (5), 1365–1385.

Yang, C.-T., Hsu, Y.-F., Huang, H.-Y., & Yeh, Y.-Y. (2011). Relative salience affects the process of detecting changes in orientation and luminance. Acta Psychologica , 138 (3), 377–389. https://doi.org/10.1016/j.actpsy.2011.09.003 .

Yang, C.-T., Little, D. R., & Hsu, C.-C. (2014). The influence of cueing on attentional focus in perceptual decision making. Attention, Perception & Psychophysics , 76 (8), 2256–2275.

Yang, C.-T., Wang, C.-H., Chang, T.-Y., Yu, J.-C., & Little, D. R. (2019). Cue-driven changes in detection strategies reflect trade-offs in strategic efficiency. Computational Brain & Behavior , 2 (2), 109–127. https://doi.org/10.1007/s42113-019-00027-0 .

Yang, C.-T., Yu, J.-C., & Chang, W.-S. (2016). Using systems factorial technology to investigate cognitive processing in redundant visual-auditory signals. Chinese Journal of Psychology , 58 (2), 1–20.

Zehetleitner, M., Krummenacher, J., & Müller, H. J. (2009). The detection of feature singletons defined in two dimensions is based on salience summation, rather than on serial exhaustive or interactive race architectures. Attention, Perception, & Psychophysics , 71 (8), 1739–1759. https://doi.org/10.3758/app.71.8.1739 .

Download references

Acknowledgements

We thank Leah Stampfer for her assistance with English editing.

The third author (CTY) is supported by the Ministry of Science and Technology, Taiwan (MOST 107–2410-H-006 -055 -MY2, MOST 105–2410-H-006 -020 -MY2, and MOST 108–2321-B-006 -022 -MY2).

Author information

Authors and affiliations.

Department of Chemical Engineering, National Cheng Kung University, Tainan, Taiwan

Cheng-Ju Hsieh

Department of Psychology, Grand Valley State University, Allendale, MI, USA

Mario Fifić

Department of Psychology, National Cheng Kung University, Tainan, Taiwan

Cheng-Ta Yang

Institute of Allied Health Sciences, National Cheng Kung University, No. 1, University Road, Tainan, Taiwan, 701

You can also search for this author in PubMed   Google Scholar

Contributions

All authors have been involved in the design of the experiments and drafting of the manuscript and have read and approved the final version.

Corresponding author

Correspondence to Cheng-Ta Yang .

Ethics declarations

Ethics approval and consent to participate.

The ethics approval for the study was obtained from the Human Research Ethics Committee of National Cheng Kung University, and the experiment was conducted in accordance with the approved guidelines and regulations. All participants provided informed consent to take part in the experiments.

Consent for publication

All participants gave permission for the publication of their data.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Additional file 1., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Hsieh, CJ., Fifić, M. & Yang, CT. A new measure of group decision-making efficiency. Cogn. Research 5 , 45 (2020). https://doi.org/10.1186/s41235-020-00244-3

Download citation

Received : 29 November 2019

Accepted : 06 August 2020

Published : 17 September 2020

DOI : https://doi.org/10.1186/s41235-020-00244-3

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Systems factorial technology
• Workload capacity
• Group decision-making

About Stanford GSB

• The Leadership
• Dean’s Updates
• School News & History
• Commencement
• Business, Government & Society
• Centers & Institutes
• Center for Entrepreneurial Studies
• Center for Social Innovation
• Stanford Seed

About the Experience

• Learning at Stanford GSB
• Experiential Learning
• Guest Speakers
• Entrepreneurship
• Social Innovation
• Communication
• Life at Stanford GSB
• Collaborative Environment
• Activities & Organizations
• Student Services
• Housing Options
• International Students

Full-Time Degree Programs

• Why Stanford MBA
• Academic Experience
• Financial Aid
• Why Stanford MSx
• Research Fellows Program
• See All Programs

Non-Degree & Certificate Programs

• Executive Education
• Stanford Executive Program
• Programs for Organizations
• The Difference
• Online Programs
• Stanford LEAD
• Seed Transformation Program
• Aspire Program
• Seed Spark Program
• Faculty Profiles
• Academic Areas
• Awards & Honors
• Conferences

Faculty Research

• Publications
• Working Papers
• Case Studies

Research Hub

• Research Labs & Initiatives
• Business Library
• Data, Analytics & Research Computing
• Behavioral Lab

Research Labs

• Cities, Housing & Society Lab
• Golub Capital Social Impact Lab

Research Initiatives

• Corporate Governance Research Initiative
• Corporations and Society Initiative
• Policy and Innovation Initiative
• Rapid Decarbonization Initiative
• Stanford Latino Entrepreneurship Initiative
• Value Chain Innovation Initiative
• Venture Capital Initiative
• Career & Success
• Climate & Sustainability
• Corporate Governance
• Culture & Society
• Finance & Investing
• Government & Politics
• Leadership & Management
• Markets & Trade
• Operations & Logistics
• Opportunity & Access
• Organizational Behavior
• Political Economy
• Social Impact
• Technology & AI
• Opinion & Analysis
• Email Newsletter

Welcome, Alumni

• Communities
• Digital Communities & Tools
• Regional Chapters
• Women’s Programs
• Identity Chapters
• Find Your Reunion
• Career Resources
• Job Search Resources
• Career & Life Transitions
• Programs & Services
• Career Video Library
• Alumni Education
• Research Resources
• Volunteering
• Alumni News
• Class Notes
• Alumni Voices
• Contact Alumni Relations
• Upcoming Events

Admission Events & Information Sessions

• MBA Program
• MSx Program
• PhD Program
• Alumni Events
• All Other Events
• Operations, Information & Technology
• Classical Liberalism
• The Eddie Lunch
• Accounting Summer Camp
• Videos, Code & Data
• California Econometrics Conference
• California Quantitative Marketing PhD Conference
• California School Conference
• China India Insights Conference
• Homo economicus, Evolving
• Political Economics (2023–24)
• Scaling Geologic Storage of CO2 (2023–24)
• A Resilient Pacific: Building Connections, Envisioning Solutions
• Adaptation and Innovation
• Changing Climate
• Civil Society
• Climate Impact Summit
• Climate Science
• Corporate Carbon Disclosures
• Earth’s Seafloor
• Environmental Justice
• Operations and Information Technology
• Organizations
• Sustainability Reporting and Control
• Taking the Pulse of the Planet
• Urban Infrastructure
• Watershed Restoration
• Junior Faculty Workshop on Financial Regulation and Banking
• Ken Singleton Celebration
• Quantitative Marketing PhD Alumni Conference
• Presentations
• Theory and Inference in Accounting Research
• Stanford Closer Look Series
• Quick Guides
• Core Concepts
• Journal Articles
• Glossary of Terms
• Faculty & Staff
• Researchers & Students
• Research Approach
• Charitable Giving
• Financial Health
• Government Services
• Workers & Careers
• Short Course
• Adaptive & Iterative Experimentation
• Incentive Design
• Social Sciences & Behavioral Nudges
• Bandit Experiment Application
• Conferences & Events
• Get Involved
• Reading Materials
• Teaching & Curriculum
• Energy Entrepreneurship
• Faculty & Affiliates
• SOLE Report
• Responsible Supply Chains
• Current Study Usage
• Pre-Registration Information
• Participate in a Study

Defining and Improving Decision Making Effectiveness

Evaluation is seen as key in any systematic attempt to improve decision making effectiveness in unstructured tasks. Decision Support Systems (DSS) are systems designed to extend a manager’s ability to deal with the complexity of the decision situation. It is argued that outcome-oriented evaluation (an evaluation that appears to dominate current organizational practices) limits the development of DDS. Process-based measures of decision making effectiveness are developed and analyzed using the concept of an evaluation space as an organizing framework. Both in theory and in practice, improving decision making effectiveness in unstructured tasks in a systematic fashion is found to be an elusive goal. Seeking a better understanding of decision making—by identifying relevant simplifying procedures, rules and beliefs; by defining an explicit reference group of effective decision makers; by developing a model of the decision task—should instead be an important operational objective in any effort to improve effectiveness.

• See the Current DEI Report
• Supporting Data
• Research & Insights
• Share Your Thoughts
• Search Fund Primer
• Affiliated Faculty
• Faculty Advisors
• Louis W. Foster Resource Center
• Defining Social Innovation
• Impact Compass
• Global Health Innovation Insights
• Faculty Affiliates
• Student Awards & Certificates
• Changemakers
• Dean Jonathan Levin
• Dean Garth Saloner
• Dean Robert Joss
• Dean Michael Spence
• Dean Robert Jaedicke
• Dean Rene McPherson
• Dean Arjay Miller
• Dean Ernest Arbuckle
• Dean Jacob Hugh Jackson
• Dean Willard Hotchkiss
• Faculty in Memoriam
• Stanford GSB Firsts
• Certificate & Award Recipients
• Dean’s Remarks
• Keynote Address
• Teaching Approach
• Analysis and Measurement of Impact
• The Corporate Entrepreneur: Startup in a Grown-Up Enterprise
• Data-Driven Impact
• Designing Experiments for Impact
• Digital Business Transformation
• The Founder’s Right Hand
• Marketing for Measurable Change
• Product Management
• Public Policy Lab: Financial Challenges Facing US Cities
• Public Policy Lab: Homelessness in California
• Lab Features
• Curricular Integration
• View From The Top
• Formation of New Ventures
• Managing Growing Enterprises
• Startup Garage
• Explore Beyond the Classroom
• Stanford Venture Studio
• Summer Program
• Workshops & Events
• The Five Lenses of Entrepreneurship
• Leadership Labs
• Executive Challenge
• Arbuckle Leadership Fellows Program
• Selection Process
• Training Schedule
• Time Commitment
• Learning Expectations
• Post-Training Opportunities
• Who Should Apply
• Introductory T-Groups
• Leadership for Society Program
• Certificate
• 2023 Awardees
• 2022 Awardees
• 2021 Awardees
• 2020 Awardees
• 2019 Awardees
• 2018 Awardees
• Social Management Immersion Fund
• Stanford Impact Founder Fellowships and Prizes
• Stanford Impact Leader Prizes
• Social Entrepreneurship
• Stanford GSB Impact Fund
• Economic Development
• Energy & Environment
• Stanford GSB Residences
• Environmental Leadership
• Stanford GSB Artwork
• A Closer Look
• California & the Bay Area
• Voices of Stanford GSB
• Business & Beneficial Technology
• Business & Sustainability
• Business & Free Markets
• Business, Government, and Society Forum
• Second Year
• Global Experiences
• JD/MBA Joint Degree
• MA Education/MBA Joint Degree
• MD/MBA Dual Degree
• MPP/MBA Joint Degree
• MS Computer Science/MBA Joint Degree
• MS Electrical Engineering/MBA Joint Degree
• MS Environment and Resources (E-IPER)/MBA Joint Degree
• Academic Calendar
• Clubs & Activities
• LGBTQ+ Students
• Military Veterans
• Minorities & People of Color
• Partners & Families
• Students with Disabilities
• Student Support
• Residential Life
• Student Voices
• MBA Alumni Voices
• A Week in the Life
• Career Support
• Employment Outcomes
• Cost of Attendance
• Knight-Hennessy Scholars Program
• Yellow Ribbon Program
• BOLD Fellows Fund
• Application Process
• Loan Forgiveness
• Contact the Financial Aid Office
• Evaluation Criteria
• GMAT & GRE
• English Language Proficiency
• Personal Information, Activities & Awards
• Professional Experience
• Letters of Recommendation
• Optional Short Answer Questions
• Application Fee
• Reapplication
• Deferred Enrollment
• Joint & Dual Degrees
• Entering Class Profile
• Event Schedule
• Ambassadors
• New & Noteworthy
• Ask a Question
• See Why Stanford MSx
• Is MSx Right for You?
• MSx Stories
• Leadership Development
• Career Advancement
• Career Change
• How You Will Learn
• Admission Events
• Personal Information
• Information for Recommenders
• GMAT, GRE & EA
• English Proficiency Tests
• After You’re Admitted
• Daycare, Schools & Camps
• U.S. Citizens and Permanent Residents
• Requirements
• Requirements: Behavioral
• Requirements: Quantitative
• Requirements: Macro
• Requirements: Micro
• Annual Evaluations
• Field Examination
• Research Activities
• Research Papers
• Dissertation
• Oral Examination
• Current Students
• Education & CV
• International Applicants
• Statement of Purpose
• Reapplicants
• Application Fee Waiver
• Deadline & Decisions
• Job Market Candidates
• Academic Placements
• Stay in Touch
• Faculty Mentors
• Current Fellows
• Standard Track
• Fellowship & Benefits
• Group Enrollment
• Program Formats
• Developing a Program
• Diversity & Inclusion
• Strategic Transformation
• Program Experience
• Contact Client Services
• Campus Experience
• Live Online Experience
• Silicon Valley & Bay Area
• Digital Credentials
• Faculty Spotlights
• Participant Spotlights
• Eligibility
• International Participants
• Stanford Ignite
• Frequently Asked Questions
• Founding Donors
• Location Information
• Participant Profile
• Network Membership
• Program Impact
• Collaborators
• Entrepreneur Profiles
• Company Spotlights
• Seed Transformation Network
• Responsibilities
• Current Coaches
• How to Apply
• Meet the Consultants
• Meet the Interns
• Intern Profiles
• Collaborate
• Research Library
• News & Insights
• Program Contacts
• Databases & Datasets
• Research Guides
• Consultations
• Research Workshops
• Career Research
• Research Data Services
• Course Reserves
• Course Research Guides
• Material Loan Periods
• Fines & Other Charges
• Document Delivery
• Interlibrary Loan
• Equipment Checkout
• Print & Scan
• MBA & MSx Students
• PhD Students
• Other Stanford Students
• Faculty Assistants
• Research Assistants
• Stanford GSB Alumni
• Telling Our Story
• Staff Directory
• Site Registration
• Alumni Directory
• Alumni Email
• Privacy Settings & My Profile
• Success Stories
• The Story of Circles
• Support Women’s Circles
• Stanford Women on Boards Initiative
• Alumnae Spotlights
• Insights & Research
• Industry & Professional
• Entrepreneurial Commitment Group
• Recent Alumni
• Half-Century Club
• Fall Reunions
• Spring Reunions
• MBA 25th Reunion
• Half-Century Club Reunion
• Faculty Lectures
• Ernest C. Arbuckle Award
• Alison Elliott Exceptional Achievement Award
• ENCORE Award
• Excellence in Leadership Award
• John W. Gardner Volunteer Leadership Award
• Robert K. Jaedicke Faculty Award
• Jack McDonald Military Service Appreciation Award
• Jerry I. Porras Latino Leadership Award
• Tapestry Award
• Student & Alumni Events
• Executive Recruiters
• Interviewing
• Land the Perfect Job with LinkedIn
• Negotiating
• Elevator Pitch
• Email Best Practices
• Resumes & Cover Letters
• Self-Assessment
• Whitney Birdwell Ball
• Margaret Brooks
• Bryn Panee Burkhart
• Margaret Chan
• Ricki Frankel
• Peter Gandolfo
• Cindy W. Greig
• Natalie Guillen
• Carly Janson
• Sloan Klein
• Sherri Appel Lassila
• Stuart Meyer
• Tanisha Parrish
• Virginia Roberson
• Philippe Taieb
• Michael Takagawa
• Terra Winston
• Johanna Wise
• Debbie Wolter
• Rebecca Zucker
• Complimentary Coaching
• Changing Careers
• Work-Life Integration
• Career Breaks
• Flexible Work
• Encore Careers
• D&B Hoovers
• Data Axle (ReferenceUSA)
• EBSCO Business Source
• Firsthand (Vault)
• Global Newsstream
• Market Share Reporter
• ProQuest One Business
• Student Clubs
• Entrepreneurial Students
• Stanford GSB Trust
• Alumni Community
• How to Volunteer
• Springboard Sessions
• Consulting Projects
• 2020 – 2029
• 2010 – 2019
• 2000 – 2009
• 1990 – 1999
• 1980 – 1989
• 1970 – 1979
• 1960 – 1969
• 1950 – 1959
• 1940 – 1949
• Service Areas
• ACT History
• ACT Awards Celebration
• ACT Governance Structure
• Building Leadership for ACT
• Individual Leadership Positions
• Leadership Role Overview
• Purpose of the ACT Management Board
• Contact ACT
• Business & Nonprofit Communities
• Reunion Volunteers
• Ways to Give
• Fiscal Year Report
• Business School Fund Leadership Council
• Planned Giving Options
• Planned Giving Benefits
• Planned Gifts and Reunions
• Legacy Partners
• Giving News & Stories
• Giving Deadlines
• Development Staff
• Submit Class Notes
• Class Secretaries
• Board of Directors
• Health Care
• Sustainability
• Class Takeaways
• All Else Equal: Making Better Decisions
• If/Then: Business, Leadership, Society
• Grit & Growth
• Think Fast, Talk Smart
• Spring 2022
• Spring 2021
• Autumn 2020
• Summer 2020
• Winter 2020
• In the Media
• For Journalists
• DCI Fellows
• Other Auditors
• Academic Calendar & Deadlines
• Course Materials
• Entrepreneurial Resources
• Campus Drive Grove
• Campus Drive Lawn
• CEMEX Auditorium
• King Community Court
• Seawell Family Boardroom
• Stanford GSB Bowl
• Stanford Investors Common
• Town Square
• Vidalakis Courtyard
• Vidalakis Dining Hall
• Catering Services
• Policies & Guidelines
• Reservations
• Contact Faculty Recruiting
• Lecturer Positions
• Postdoctoral Positions
• Accommodations
• CMC-Managed Interviews
• Recruiter-Managed Interviews
• Virtual Interviews
• Campus & Virtual
• Search for Candidates
• Think Globally
• Recruiting Calendar
• Recruiting Policies
• Full-Time Employment
• Summer Employment
• Entrepreneurial Summer Program
• Global Management Immersion Experience
• Social-Purpose Summer Internships
• Process Overview
• Project Types
• Client Eligibility Criteria
• Client Screening
• ACT Leadership
• Social Innovation & Nonprofit Management Resources
• Develop Your Organization’s Talent
• Centers & Initiatives
• Student Fellowships

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 08 April 2024

A case study on the relationship between risk assessment of scientific research projects and related factors under the Naive Bayesian algorithm

• Xuying Dong 1 &
• Wanlin Qiu 1  

Scientific Reports volume  14 , Article number:  8244 ( 2024 ) Cite this article

Metrics details

• Computer science
• Mathematics and computing

This paper delves into the nuanced dynamics influencing the outcomes of risk assessment (RA) in scientific research projects (SRPs), employing the Naive Bayes algorithm. The methodology involves the selection of diverse SRPs cases, gathering data encompassing project scale, budget investment, team experience, and other pertinent factors. The paper advances the application of the Naive Bayes algorithm by introducing enhancements, specifically integrating the Tree-augmented Naive Bayes (TANB) model. This augmentation serves to estimate risk probabilities for different research projects, shedding light on the intricate interplay and contributions of various factors to the RA process. The findings underscore the efficacy of the TANB algorithm, demonstrating commendable accuracy (average accuracy 89.2%) in RA for SRPs. Notably, budget investment (regression coefficient: 0.68, P < 0.05) and team experience (regression coefficient: 0.51, P < 0.05) emerge as significant determinants obviously influencing RA outcomes. Conversely, the impact of project size (regression coefficient: 0.31, P < 0.05) is relatively modest. This paper furnishes a concrete reference framework for project managers, facilitating informed decision-making in SRPs. By comprehensively analyzing the influence of various factors on RA, the paper not only contributes empirical insights to project decision-making but also elucidates the intricate relationships between different factors. The research advocates for heightened attention to budget investment and team experience when formulating risk management strategies. This strategic focus is posited to enhance the precision of RAs and the scientific foundation of decision-making processes.

Similar content being viewed by others

Prediction of SMEs’ R&D performances by machine learning for project selection

Hyoung Sun Yoo, Ye Lim Jung & Seung-Pyo Jun

Bayesian Analysis Reporting Guidelines

John K. Kruschke

An intelligent methodology for the use of multi-criteria decision analysis in impact assessment: the case of real-world offshore construction

Mariza Tsakalerou, Damianos Efthymiadis & Almat Abilez

Introduction

Scientific research projects (SRPs) stand as pivotal drivers of technological advancement and societal progress in the contemporary landscape 1 , 2 , 3 . The dynamism of SRP success hinges on a multitude of internal and external factors 4 . Central to effective project management, Risk assessment (RA) in SRPs plays a critical role in identifying and quantifying potential risks. This process not only aids project managers in formulating strategic decision-making approaches but also enhances the overall success rate and benefits of projects. In a recent contribution, Salahuddin 5 provides essential numerical techniques indispensable for conducting RAs in SRPs. Building on this foundation, Awais and Salahuddin 6 delve into the assessment of risk factors within SRPs, notably introducing the consideration of activation energy through an exploration of the radioactive magnetohydrodynamic model. Further expanding the scope, Awais and Salahuddin 7 undertake a study on the natural convection of coupled stress fluids. However, RA of SRPs confronts a myriad of challenges, underscoring the critical need for novel methodologies 8 . Primarily, the intricate nature of SRPs renders precise RA exceptionally complex and challenging. The project’s multifaceted dimensions, encompassing technology, resources, and personnel, are intricately interwoven, posing a formidable challenge for traditional assessment methods to comprehensively capture all potential risks 9 . Furthermore, the intricate and diverse interdependencies among various project factors contribute to the complexity of these relationships, thereby limiting the efficacy of conventional methods 10 , 11 , 12 . Traditional approaches often focus solely on the individual impact of diverse factors, overlooking the nuanced relationships that exist between them—an inherent limitation in the realm of RA for SRPs 13 , 14 , 15 .

The pursuit of a methodology capable of effectively assessing project risks while elucidating the intricate interplay of different factors has emerged as a focal point in SRPs management 16 , 17 , 18 . This approach necessitates a holistic consideration of multiple factors, their quantification in contributing to project risks, and the revelation of their correlations. Such an approach enables project managers to more precisely predict and respond to risks. Marx-Stoelting et al. 19 , current approaches for the assessment of environmental and human health risks due to exposure to chemical substances have served their purpose reasonably well. Additionally, Awais et al. 20 highlights the significance of enthalpy changes in SRPs risk considerations, while Awais et al. 21 delve into the comprehensive exploration of risk factors in Eyring-Powell fluid flow in magnetohydrodynamics, particularly addressing viscous dissipation and activation energy effects. The Naive Bayesian algorithm, recognized for its prowess in probability and statistics, has yielded substantial results in information retrieval and data mining in recent years 22 . Leveraging its advantages in classification and probability estimation, the algorithm presents a novel approach for RA of SRPs 23 . Integrating probability analysis into RA enables a more precise estimation of project risks by utilizing existing project data and harnessing the capabilities of the Naive Bayesian algorithms. This method facilitates a quantitative, statistical analysis of various factors, effectively navigating the intricate relationships between them, thereby enhancing the comprehensiveness and accuracy of RA for SRPs.

This paper seeks to employ the Naive Bayesian algorithm to estimate the probability of risks by carefully selecting distinct research project cases and analyzing multidimensional data, encompassing project scale, budget investment, and team experience. Concurrently, Multiple Linear Regression (MLR) analysis is applied to quantify the influence of these factors on the assessment results. The paper places particular emphasis on exploring the intricate interrelationships between different factors, aiming to provide a more specific and accurate reference framework for decision-making in SRPs management.

This paper introduces several innovations and contributions to the field of RA for SRPs:

Comprehensive Consideration of Key Factors: Unlike traditional research that focuses on a single factor, this paper comprehensively considers multiple key factors, such as project size, budget investment, and team experience. This holistic analysis enhances the realism and thoroughness of RA for SRPs.

Introduction of Tree-Enhanced Naive Bayes Model: The naive Bayes algorithm is introduced and further improved through the proposal of a tree-enhanced naive Bayes model. This algorithm exhibits unique advantages in handling uncertainty and complexity, thereby enhancing its applicability and accuracy in the RA of scientific and technological projects.

Empirical Validation: The effectiveness of the proposed method is not only discussed theoretically but also validated through empirical cases. The analysis of actual cases provides practical support and verification, enhancing the credibility of the research results.

Application of MLR Analysis: The paper employs MLR analysis to delve into the impact of various factors on RA. This quantitative analysis method adds specificity and operability to the research, offering a practical decision-making basis for scientific and technological project management.

Discovery of New Connections and Interactions: The paper uncovers novel connections and interactions, such as the compensatory role of team experience for budget-related risks and the impact of the interaction between project size and budget investment on RA results. These insights provide new perspectives for decision-making in technology projects, contributing significantly to the field of RA for SRPs in terms of both importance and practical value.

The paper is structured as follows: “ Introduction ” briefly outlines the significance of RA for SRPs. Existing challenges within current research are addressed, and the paper’s core objectives are elucidated. A distinct emphasis is placed on the innovative aspects of this research compared to similar studies. The organizational structure of the paper is succinctly introduced, providing a brief overview of each section’s content. “ Literature review ” provides a comprehensive review of relevant theories and methodologies in the domain of RA for SRPs. The current research landscape is systematically examined, highlighting the existing status and potential gaps. Shortcomings in previous research are analyzed, laying the groundwork for the paper’s motivation and unique contributions. “ Research methodology ” delves into the detailed methodologies employed in the paper, encompassing data collection, screening criteria, preprocessing steps, and more. The tree-enhanced naive Bayes model is introduced, elucidating specific steps and the purpose behind MLR analysis. “ Results and discussion ” unfolds the results and discussions based on selected empirical cases. The representativeness and diversity of these cases are expounded upon. An in-depth analysis of each factor’s impact and interaction in the context of RA is presented, offering valuable insights. “ Discussion ” succinctly summarizes the entire research endeavor. Potential directions for further research and suggestions for improvement are proposed, providing a thoughtful conclusion to the paper.

Literature review

A review of ra for srps.

In recent years, the advancement of SRPs management has led to the evolution of various RA methods tailored for SRPs. The escalating complexity of these projects poses a challenge for traditional methods, often falling short in comprehensively considering the intricate interplay among multiple factors and yielding incomplete assessment outcomes. Scholars, recognizing the pivotal role of factors such as project scale, budget investment, and team experience in influencing project risks, have endeavored to explore these dynamics from diverse perspectives. Siyal et al. 24 pioneered the development and testing of a model geared towards detecting SRPs risks. Chen et al. 25 underscored the significance of visual management in SRPs risk management, emphasizing its importance in understanding and mitigating project risks. Zhao et al. 26 introduced a classic approach based on cumulative prospect theory, offering an optional method to elucidate researchers’ psychological behaviors. Their study demonstrated the enhanced rationality achieved by utilizing the entropy weight method to derive attribute weight information under Pythagorean fuzzy sets. This approach was then applied to RA for SRPs, showcasing a model grounded in the proposed methodology. Suresh and Dillibabu 27 proposed an innovative hybrid fuzzy-based machine learning mechanism tailored for RA in software projects. This hybrid scheme facilitated the identification and ranking of major software project risks, thereby supporting decision-making throughout the software project lifecycle. Akhavan et al. 28 introduced a Bayesian network modeling framework adept at capturing project risks by calculating the uncertainty of project net present value. This model provided an effective means for analyzing risk scenarios and their impact on project success, particularly applicable in evaluating risks for innovative projects that had undergone feasibility studies.

A review of factors affecting SRPs

Within the realm of SRPs management, the assessment and proficient management of project risks stand as imperative components. Consequently, a range of studies has been conducted to explore diverse methods and models aimed at enhancing the comprehension and decision support associated with project risks. Guan et al. 29 introduced a new risk interdependence network model based on Monte Carlo simulation to support decision-makers in more effectively assessing project risks and planning risk management actions. They integrated interpretive structural modeling methods into the model to develop a hierarchical project risk interdependence network based on identified risks and their causal relationships. Vujović et al. 30 provided a new method for research in project management through careful analysis of risk management in SRPs. To confirm the hypothesis, the study focused on educational organizations and outlined specific project management solutions in business systems, thereby improving the business and achieving positive business outcomes. Muñoz-La Rivera et al. 31 described and classified the 100 identified factors based on the dimensions and aspects of the project, assessed their impact, and determined whether they were shaping or directly affecting the occurrence of research project accidents. These factors and their descriptions and classifications made significant contributions to improving the security creation of the system and generating training and awareness materials, fostering the development of a robust security culture within organizations. Nguyen et al. concentrated on the pivotal risk factors inherent in design-build projects within the construction industry. Effective identification and management of these factors enhanced project success and foster confidence among owners and contractors in adopting the design-build approach 32 . Their study offers valuable insights into RA in project management and the adoption of new contract forms. Nguyen and Le delineated risk factors influencing the quality of 20 civil engineering projects during the construction phase 33 . The top five risks identified encompass poor raw material quality, insufficient worker skills, deficient design documents and drawings, geographical challenges at construction sites, and inadequate capabilities of main contractors and subcontractors. Meanwhile, Nguyen and Phu Pham concentrated on office building projects in Ho Chi Minh City, Vietnam, to pinpoint key risk factors during the construction phase 34 . These factors were classified into five groups based on their likelihood and impact: financial, management, schedule, construction, and environmental. Findings revealed that critical factors affecting office building projects encompassed both natural elements (e.g., prolonged rainfall, storms, and climate impacts) and human factors (e.g., unstable soil, safety behavior, owner-initiated design changes), with schedule-related risks exerting the most significant influence during the construction phase of Ho Chi Minh City’s office building projects. This provides construction and project management practitioners with fresh insights into risk management, aiding in the comprehensive identification, mitigation, and management of risk factors in office building projects.

While existing research has made notable strides in RA for SRPs, certain limitations persist. These studies limitations in quantifying the degree of influence of various factors and analyzing their interrelationships, thereby falling short of offering specific and actionable recommendations. Traditional methods, due to their inherent limitations, struggle to precisely quantify risk degrees and often overlook the intricate interplay among multiple factors. Consequently, there is an urgent need for a comprehensive method capable of quantifying the impact of diverse factors and revealing their correlations. In response to this exigency, this paper introduces the TANB model. The unique advantages of this algorithm in the RA of scientific and technological projects have been fully realized. Tailored to address the characteristics of uncertainty and complexity, the model represents a significant leap forward in enhancing applicability and accuracy. In comparison with traditional methods, the TANB model exhibits greater flexibility and a heightened ability to capture dependencies between features, thereby elevating the overall performance of RA. This innovative method emerges as a more potent and reliable tool in the realm of scientific and technological project management, furnishing decision-makers with more comprehensive and accurate support for RA.

Research methodology

This paper centers on the latest iteration of ISO 31000, delving into the project risk management process and scrutinizing the RA for SRPs and their intricate interplay with associated factors. ISO 31000, an international risk management standard, endeavors to furnish businesses, organizations, and individuals with a standardized set of risk management principles and guidelines, defining best practices and establishing a common framework. The paper unfolds in distinct phases aligned with ISO 31000:

Risk Identification: Employing data collection and preparation, a spectrum of factors related to project size, budget investment, team member experience, project duration, and technical difficulty were identified.

RA: Utilizing the Naive Bayes algorithm, the paper conducts RA for SRPs, estimating the probability distribution of various factors influencing RA results.

Risk Response: The application of the Naive Bayes model is positioned as a means to respond to risks, facilitating the formulation of apt risk response strategies based on calculated probabilities.

Monitoring and Control: Through meticulous data collection, model training, and verification, the paper illustrates the steps involved in monitoring and controlling both data and models. Regular monitoring of identified risks and responses allows for adjustments when necessary.

Communication and Reporting: Maintaining effective communication throughout the project lifecycle ensures that stakeholders comprehend the status of project risks. Transparent reporting on discussions and outcomes contributes to an informed project environment.

Data collection and preparation

In this paper, a meticulous approach is undertaken to select representative research project cases, adhering to stringent screening criteria. Additionally, a thorough review of existing literature is conducted and tailored to the practical requirements of SRPs management. According to Nguyen et al., these factors play a pivotal role in influencing the RA outcomes of SRPs 35 . Furthermore, research by He et al. underscored the significant impact of team members’ experience on project success 36 . Therefore, in alignment with our research objectives and supported by the literature, this paper identifies variables such as project scale, budget investment, team member experience, project duration, and technical difficulty as the focal themes. To ensure the universality and scientific rigor of our findings, the paper adheres to stringent selection criteria during the project case selection process. After preliminary screening of SRPs completed in the past 5 years, considering factors such as project diversity, implementation scales, and achieved outcomes, five representative projects spanning diverse fields, including engineering, medicine, and information technology, are ultimately selected. These project cases are chosen based on their capacity to represent various scales and types of SRPs, each possessing a typical risk management process, thereby offering robust and comprehensive data support for our study. The subsequent phase involves detailed data collection on each chosen project, encompassing diverse dimensions such as project scale, budget investment, team member experience, project cycle, and technical difficulty. The collected data undergo meticulous preprocessing to ensure data quality and reliability. The preprocessing steps comprised data cleaning, addressing missing values, handling outliers, culminating in the creation of a self-constructed dataset. The dataset encompasses over 500 SRPs across diverse disciplines and fields, ensuring statistically significant and universal outcomes. Particular emphasis is placed on ensuring dataset diversity, incorporating projects of varying scales, budgets, and team experience levels. This comprehensive coverage ensures the representativeness and credibility of the study on RA in SRPs. New influencing factors are introduced to expand the research scope, including project management quality (such as time management and communication efficiency), historical success rate, industry dynamics, and market demand. Detailed definitions and quantifications are provided for each new variable to facilitate comprehensive data processing and analysis. For project management quality, consideration is given to time management accuracy and communication frequency and quality among team members. Historical success rate is determined by reviewing past project records and outcomes. Industry dynamics are assessed by consulting the latest scientific literature and patent information. Market demand is gauged through market research and user demand surveys. The introduction of these variables enriches the understanding of RA in SRPs and opens up avenues for further research exploration.

At the same time, the collected data are integrated and coded in order to apply Naive Bayes algorithm and MLR analysis. For cases involving qualitative data, this paper uses appropriate coding methods to convert it into quantitative data for processing in the model. For example, for the qualitative feature of team member experience, numerical values are used to represent different experience levels, such as 0 representing beginners, 0 representing intermediate, and 2 representing advanced. The following is a specific sample data set example (Table 1 ). It shows the processed structured data, and the values in the table represent the specific characteristics of each project.

Establishment of naive Bayesian model

The Naive Bayesian algorithm, a probabilistic and statistical classification method renowned for its effectiveness in analyzing and predicting multi-dimensional data, is employed in this paper to conduct the RA for SRPs. The application of the Naive Bayesian algorithm to RA for SRPs aims to discern the influence of various factors on the outcomes of RA. The Naive Bayesian algorithm, depicted in Fig.  1 , operates on the principles of Bayesian theorem, utilizing posterior probability calculations for classification tasks. The fundamental concept of this algorithm hinges on the assumption of independence among different features, embodying the “naivety” hypothesis. In the context of RA for SRPs, the Naive Bayesian algorithm is instrumental in estimating the probability distribution of diverse factors affecting the RA results, thereby enhancing the precision of risk estimates. In the Naive Bayesian model, the initial step involves the computation of posterior probabilities for each factor, considering the given RA result conditions. Subsequently, the category with the highest posterior probability is selected as the predictive outcome.

Naive Bayesian algorithm process.

In Fig.  1 , the data collection process encompasses vital project details such as project scale, budget investment, team member experience, project cycle, technical difficulty, and RA results. This meticulous collection ensures the integrity and precision of the dataset. Subsequently, the gathered data undergoes integration and encoding to convert qualitative data into quantitative form, facilitating model processing and analysis. Tailored to specific requirements, relevant features are chosen for model construction, accompanied by essential preprocessing steps like standardization and normalization. The dataset is then partitioned into training and testing sets, with the model trained on the former and its performance verified on the latter. Leveraging the training data, a Naive Bayesian model is developed to estimate probability distribution parameters for various features across distinct categories. Ultimately, the trained model is employed to predict new project features, yielding RA results.

Naive Bayesian models, in this context, are deployed to forecast diverse project risk levels. Let X symbolize the feature vector, encompassing project scale, budget investment, team member experience, project cycle, and technical difficulty. The objective is to predict the project’s risk level, denoted as Y. Y assumes discrete values representing distinct risk levels. Applying the Bayesian theorem, the posterior probability P(Y|X) is computed, signifying the probability distribution of projects falling into different risk levels given the feature vector X. The fundamental equation governing the Naive Bayesian model is expressed as:

In Eq. ( 1 ), P(Y|X) represents the posterior probability, denoting the likelihood of the project belonging to a specific risk level. P(X|Y) signifies the class conditional probability, portraying the likelihood of the feature vector X occurring under known risk level conditions. P(Y) is the prior probability, reflecting the antecedent likelihood of the project pertaining to a particular risk level. P(X) acts as the evidence factor, encapsulating the likelihood of the feature vector X occurring.

The Naive Bayes, serving as the most elementary Bayesian network classifier, operates under the assumption of attribute independence given the class label c , as expressed in Eq. ( 2 ):

The classification decision formula for Naive Bayes is articulated in Eq. ( 3 ):

The Naive Bayes model, rooted in the assumption of conditional independence among attributes, often encounters deviations from reality. To address this limitation, the Tree-Augmented Naive Bayes (TANB) model extends the independence assumption by incorporating a first-order dependency maximum-weight spanning tree. TANB introduces a tree structure that more comprehensively models relationships between features, easing the constraints of the independence assumption and concurrently mitigating issues associated with multicollinearity. This extension bolsters its efficacy in handling intricate real-world data scenarios. TANB employs conditional mutual information \(I(X_{i} ;X_{j} |C)\) to gauge the dependency between attributes \(X_{j}\) and \(X_{i}\) , thereby constructing the maximum weighted spanning tree. In TANB, any attribute variable \(X_{i}\) is permitted to have at most one other attribute variable as its parent node, expressed as \(Pa\left( {X_{i} } \right) \le 2\) . The joint probability \(P_{con} \left( {x,c} \right)\) undergoes transformation using Eq. ( 4 ):

In Eq. ( 4 ), \(x_{r}\) refers to the root node, which can be expressed as Eq. ( 5 ):

TANB classification decision equation is presented below:

In the RA of SRPs, normal distribution parameters, such as mean (μ) and standard deviation (σ), are estimated for each characteristic dimension (project scale, budget investment, team member experience, project cycle, and technical difficulty). This estimation allows the calculation of posterior probabilities for projects belonging to different risk levels under given feature vector conditions. For each feature dimension \({X}_{i}\) , the mean \({mu}_{i,j}\) and standard deviation \({{\text{sigma}}}_{i,j}\) under each risk level are computed, where i represents the feature dimension, and j denotes the risk level. Parameter estimation employs the maximum likelihood method, and the specific calculations are as follows.

In Eqs. ( 7 ) and ( 8 ), \({N}_{j}\) represents the number of projects belonging to risk level j . \({x}_{i,k}\) denotes the value of the k -th item in the feature dimension i . Finally, under a given feature vector, the posterior probability of a project with risk level j is calculated as Eq. ( 9 ).

In Eq. ( 9 ), d represents the number of feature dimensions, and Z is the normalization factor. \(P(Y=j)\) represents the prior probability of category j . \(P({X}_{i}\mid Y=j)\) represents the normal distribution probability density function of feature dimension i under category j . The risk level of a project can be predicted by calculating the posterior probabilities of different risk levels to achieve RA for SRPs.

This paper integrates the probability estimation of the Naive Bayes model with actual project risk response strategies, enabling a more flexible and targeted response to various risk scenarios. Such integration offers decision support to project managers, enhancing their ability to address potential challenges effectively and ultimately improving the overall success rate of the project. This underscores the notion that risk management is not solely about problem prevention but stands as a pivotal factor contributing to project success.

MLR analysis

MLR analysis is used to validate the hypothesis to deeply explore the impact of various factors on RA of SRPs. Based on the previous research status, the following research hypotheses are proposed.

Hypothesis 1: There is a positive relationship among project scale, budget investment, and team member experience and RA results. As the project scale, budget investment, and team member experience increase, the RA results also increase.

Hypothesis 2: There is a negative relationship between the project cycle and the RA results. Projects with shorter cycles may have higher RA results.

Hypothesis 3: There is a complex relationship between technical difficulty and RA results, which may be positive, negative, or bidirectional in some cases. Based on these hypotheses, an MLR model is established to analyze the impact of factors, such as project scale, budget investment, team member experience, project cycle, and technical difficulty, on RA results. The form of an MLR model is as follows.

In Eq. ( 10 ), Y represents the RA result (dependent variable). \({X}_{1}\) to \({X}_{5}\) represent factors, such as project scale, budget investment, team member experience, project cycle, and technical difficulty (independent variables). \({\beta }_{0}\) to \({\beta }_{5}\) are the regression coefficients, which represent the impact of various factors on the RA results. \(\epsilon\) represents a random error term. The model structure is shown in Fig.  2 .

Schematic diagram of an MLR model.

In Fig.  2 , the MLR model is employed to scrutinize the influence of various independent variables on the outcomes of RA. In this specific context, the independent variables encompass project size, budget investment, team member experience, project cycle, and technical difficulty, all presumed to impact the project’s RA results. Each independent variable is denoted as a node in the model, with arrows depicting the relationships between these factors. In an MLR model, the arrow direction signifies causality, illustrating the influence of an independent variable on the dependent variable.

When conducting MLR analysis, it is necessary to estimate the parameter \(\upbeta\) in the regression model. These parameters determine the relationship between the independent and dependent variables. Here, the Ordinary Least Squares (OLS) method is applied to estimate these parameters. The OLS method is a commonly used parameter estimation method aimed at finding parameter values that minimize the sum of squared residuals between model predictions and actual observations. The steps are as follows. Firstly, based on the general form of an MLR model, it is assumed that there is a linear relationship between the independent and dependent variables. It can be represented by a linear equation, which includes regression coefficients β and the independent variable X. For each observation value, the difference between its predicted and actual values is calculated, which is called the residual. Residual \({e}_{i}\) can be expressed as:

In Eq. ( 11 ), \({Y}_{i}\) is the actual observation value, and \({\widehat{Y}}_{i}\) is the value predicted by the model. The goal of the OLS method is to adjust the regression coefficients \(\upbeta\) to minimize the sum of squared residuals of all observations. This can be achieved by solving an optimization problem, and the objective function is the sum of squared residuals.

Then, the estimated value of the regression coefficient \(\upbeta\) that minimizes the sum of squared residuals can be obtained by taking the derivative of the objective function and making the derivative zero. The estimated values of the parameters can be obtained by solving this system of equations. The final estimated regression coefficient can be expressed as:

In Eq. ( 13 ), X represents the independent variable matrix. Y represents the dependent variable vector. \(({X}^{T}X{)}^{-1}\) is the inverse of a matrix, and \(\widehat{\beta }\) is a parameter estimation vector.

Specifically, solving for the estimated value of regression coefficient \(\upbeta\) requires matrix operation and statistical analysis. Based on the collected project data, substitute it into the model and calculate the residual. Then, the steps of the OLS method are used to obtain parameter estimates. These parameter estimates are used to establish an MLR model to predict RA results and further analyze the influence of different factors.

The degree of influence of different factors on the RA results can be determined by analyzing the value of the regression coefficient β. A positive \(\upbeta\) value indicates that the factor has a positive impact on the RA results, while a negative \(\upbeta\) value indicates that the factor has a negative impact on the RA results. Additionally, hypothesis testing can determine whether each factor is significant in the RA results.

The TANB model proposed in this paper extends the traditional naive Bayes model by incorporating conditional dependencies between attributes to enhance the representation of feature interactions. While the traditional naive Bayes model assumes feature independence, real-world scenarios often involve interdependencies among features. To address this, the TANB model is introduced. The TANB model introduces a tree structure atop the naive Bayes model to more accurately model feature relationships, overcoming the limitation of assuming feature independence. Specifically, the TANB model constructs a maximum weight spanning tree to uncover conditional dependencies between features, thereby enabling the model to better capture feature interactions.

Assessment indicators

To comprehensively assess the efficacy of the proposed TANB model in the RA for SRPs, a self-constructed dataset serves as the data source for this experimental evaluation, as outlined in Table 1 . The dataset is segregated into training (80%) and test sets (20%). These indicators cover the accuracy, precision, recall rate, F1 value, and Area Under the Curve (AUC) Receiver Operating Characteristic (ROC) of the model. The following are the definitions and equations for each assessment indicator. Accuracy is the proportion of correctly predicted samples to the total number of samples. Precision is the proportion of Predicted Positive (PP) samples to actual positive samples. The recall rate is the proportion of correctly PP samples among the actual positive samples. The F1 value is the harmonic average of precision and recall, considering the precision and comprehensiveness of the model. The area under the ROC curve measures the classification performance of the model, and a larger AUC value indicates better model performance. The ROC curve suggests the relationship between True Positive Rate and False Positive Rate under different thresholds. The AUC value can be obtained by accumulating the area of each small rectangle under the ROC curve. The confusion matrix is used to display the prediction property of the model in different categories, including True Positive (TP), True Negative (TN), False Positive (FP), and False Negative (FN).

The performance of TANB in RA for SRPs can be comprehensively assessed to understand the advantages, disadvantages, and applicability of the model more comprehensively by calculating the above assessment indicators.

Results and discussion

Accuracy analysis of naive bayesian algorithm.

On the dataset of this paper, Fig.  3 reveals the performance of TANB algorithm under different assessment indicators.

Performance assessment of TANB algorithm on different projects.

From Fig.  3 , the TANB algorithm performs well in various projects, ranging from 0.87 to 0.911 in accuracy. This means that the overall accuracy of the model in predicting project risks is quite high. The precision also maintains a high level in various projects, ranging from 0.881 to 0.923, indicating that the model performs well in classifying high-risk categories. The recall rate ranges from 0.872 to 0.908, indicating that the model can effectively capture high-risk samples. Meanwhile, the AUC values in each project are relatively high, ranging from 0.905 to 0.931, which once again emphasizes the effectiveness of the model in risk prediction. From multiple assessment indicators, such as accuracy, precision, recall, F1 value, and AUC, the TANB algorithm has shown good risk prediction performance in representative projects. The performance assessment results of the TANB algorithm under different feature dimensions are plotted in Figs.  4 , 5 , 6 and 7 .

Prediction accuracy of TANB algorithm on different budget investments.

Prediction accuracy of TANB algorithm on different team experiences.

Prediction accuracy of TANB algorithm at different risk levels.

Prediction accuracy of TANB algorithm on different project scales.

From Figs.  4 , 5 , 6 and 7 , as the level of budget investment increases, the accuracy of most projects also shows an increasing trend. Especially in cases of high budget investment, the accuracy of the project is generally high. This may mean that a higher budget investment helps to reduce project risks, thereby improving the prediction accuracy of the TANB algorithm. It can be observed that team experience also affects the accuracy of the model. Projects with high team experience exhibit higher accuracy in TANB algorithms. This may indicate that experienced teams can better cope with project risks to improve the performance of the model. When budget investment and team experience are low, accuracy is relatively low. This may imply that budget investment and team experience can complement each other to affect the model performance.

There are certain differences in the accuracy of projects under different risk levels. Generally speaking, the accuracy of high-risk and medium-risk projects is relatively high, while the accuracy of low-risk projects is relatively low. This may be because high-risk and medium-risk projects require more accurate predictions, resulting in higher accuracy. Similarly, project scale also affects the performance of the model. Large-scale and medium-scale projects exhibit high accuracy in TANB algorithms, while small-scale projects have relatively low accuracy. This may be because the risks of large-scale and medium-scale projects are easier to identify and predict to promote the performance of the model. In high-risk and large-scale projects, accuracy is relatively high. This may indicate that the impact of project scale is more significant in specific risk scenarios.

Figure  8 further compares the performance of the TANB algorithm proposed here with other similar algorithms.

Performance comparison of different algorithms in RA of SRPs.

As depicted in Fig.  8 , the TANB algorithm attains an accuracy and precision of 0.912 and 0.920, respectively, surpassing other algorithms. It excels in recall rate and F1 value, registering 0.905 and 0.915, respectively, outperforming alternative algorithms. These findings underscore the proficiency of the TANB algorithm in comprehensively identifying high-risk projects while sustaining high classification accuracy. Moreover, the algorithm achieves an AUC of 0.930, indicative of its exceptional predictive prowess in sample classification. Thus, the TANB algorithm exhibits notable potential for application, particularly in scenarios demanding the recognition and comprehensiveness requisite for high-risk project identification. The evaluation results of the TANB model in predicting project risk levels are presented in Table 2 :

Table 2 demonstrates that the TANB model surpasses the traditional Naive Bayes model across multiple evaluation metrics, including accuracy, precision, and recall. This signifies that, by accounting for feature interdependence, the TANB model can more precisely forecast project risk levels. Furthermore, leveraging the model’s predictive outcomes, project managers can devise tailored risk mitigation strategies corresponding to various risk scenarios. For example, in high-risk projects, more assertive measures can be implemented to address risks, while in low-risk projects, risks can be managed more cautiously. This targeted risk management approach contributes to enhancing project success rates, thereby ensuring the seamless advancement of SRPs.

The exceptional performance of the TANB model in specific scenarios derives from its distinctive characteristics and capabilities. Firstly, compared to traditional Naive Bayes models, the TANB model can better capture the dependencies between attributes. In project RA, project features often exhibit complex interactions. The TANB model introduces first-order dependencies between attributes, allowing features to influence each other, thereby more accurately reflecting real-world situations and improving risk prediction precision. Secondly, the TANB model demonstrates strong adaptability and generalization ability in handling multidimensional data. SRPs typically involve data from multiple dimensions, such as project scale, budget investment, and team experience. The TANB model effectively processes these multidimensional data, extracts key information, and achieves accurate RA for projects. Furthermore, the paper explores the potential of using hybrid models or ensemble learning methods to further enhance model performance. By combining other machine learning algorithms, such as random forests and support vector regressors with sigmoid kernel, through ensemble learning, the shortcomings of individual models in specific scenarios can be overcome, thus improving the accuracy and robustness of RA. For example, in the study, we compared the performance of the TANB model with other algorithms in RA, as shown in Table 3 .

Table 3 illustrates that the TANB model surpasses other models in terms of accuracy, precision, recall, F1 value, and AUC value, further confirming its superiority and practicality in RA. Therefore, the TANB model holds significant application potential in SRPs, offering effective decision support for project managers to better evaluate and manage project risks, thereby enhancing the likelihood of project success.

Analysis of the degree of influence of different factors

Table 4 analyzes the degree of influence and interaction of different factors.

In Table 4 , the regression analysis results reveal that budget investment and team experience exert a significantly positive impact on RA outcomes. This suggests that increasing budget allocation and assembling a team with extensive experience can enhance project RA outcomes. Specifically, the regression coefficient for budget investment is 0.68, and for team experience, it is 0.51, both demonstrating significant positive effects (P < 0.05). The P-values are all significantly less than 0.05, indicating a significant impact. The impact of project scale is relatively small, at 0.31, and its P-value is also much less than 0.05. The degree of interaction influence is as follows. The impact of interaction terms is also significant, especially the interaction between budget investment and team experience and the interaction between budget investment and project scale. The P value of the interaction between budget investment and project scale is 0.002, and the P value of the interaction between team experience and project scale is 0.003. The P value of the interaction among budget investment, team experience, and project scale is 0.005. So, there are complex relationships and interactions among different factors, and budget investment and team experience significantly affect the RA results. However, the budget investment and project scale slightly affect the RA results. Project managers should comprehensively consider the interactive effects of different factors when making decisions to more accurately assess the risks of SRPs.

The interaction between team experience and budget investment

The results of the interaction between team experience and budget investment are demonstrated in Table 5 .

From Table 5 , the degree of interaction impact can be obtained. Budget investment and team experience, along with the interaction between project scale and technical difficulty, are critical factors in risk mitigation. Particularly in scenarios characterized by large project scales and high technical difficulties, adequate budget allocation and a skilled team can substantially reduce project risks. As depicted in Table 5 , under conditions of high team experience and sufficient budget investment, the average RA outcome is 0.895 with a standard deviation of 0.012, significantly lower than assessment outcomes under other conditions. This highlights the synergistic effects of budget investment and team experience in effectively mitigating risks in complex project scenarios. The interaction between team experience and budget investment has a significant impact on RA results. Under high team experience, the impact of different budget investment levels on RA results is not significant, but under medium and low team experience, the impact of different budget investment levels on RA results is significantly different. The joint impact of team experience and budget investment is as follows. Under high team experience, the impact of budget investment is relatively small, possibly because high-level team experience can compensate for the risks brought by insufficient budget to some extent. Under medium and low team experience, the impact of budget investment is more significant, possibly because the lack of team experience makes budget investment play a more important role in RA. Therefore, team experience and budget investment interact in RA of SRPs. They need to be comprehensively considered in project decision-making. High team experience can compensate for the risks brought by insufficient budget to some extent, but in the case of low team experience, the impact of budget investment on RA is more significant. An exhaustive consideration of these factors and their interplay is imperative for effectively assessing the risks inherent in SRPs. Merely focusing on budget allocation or team expertise may not yield a thorough risk evaluation. Project managers must scrutinize the project’s scale, technical complexity, and team proficiency, integrating these aspects with budget allocation and team experience. This holistic approach fosters a more precise RA and facilitates the development of tailored risk management strategies, thereby augmenting the project’s likelihood of success. In conclusion, acknowledging the synergy between budget allocation and team expertise, in conjunction with other pertinent factors, is pivotal in the RA of SRPs. Project managers should adopt a comprehensive outlook to ensure sound decision-making and successful project execution.

Risk mitigation strategies

To enhance the discourse on project risk management in this paper, a dedicated section on risk mitigation strategies has been included. Leveraging the insights gleaned from the predictive model regarding identified risk factors and their corresponding risk levels, targeted risk mitigation measures are proposed.

Primarily, given the significant influence of budget investment and team experience on project RA outcomes, project managers are advised to prioritize these factors and devise pertinent risk management strategies.

For risks stemming from budget constraints, the adoption of flexible budget allocation strategies is advocated. This may involve optimizing project expenditures, establishing financial reserves, or seeking additional funding avenues.

In addressing risks attributed to inadequate team experience, measures such as enhanced training initiatives, engagement of seasoned project advisors, or collaboration with experienced teams can be employed to mitigate the shortfall in expertise.

Furthermore, recognizing the impact of project scale, duration, and technical complexity on RA outcomes, project managers are advised to holistically consider these factors during project planning. This entails adjusting project scale as necessary, establishing realistic project timelines, and conducting thorough assessments of technical challenges prior to project commencement.

These risk mitigation strategies aim to equip project managers with a comprehensive toolkit for effectively identifying, assessing, and mitigating risks inherent in SRPs.

This paper delves into the efficacy of the TANB algorithm in project risk prediction. The findings indicate that the algorithm demonstrates commendable performance across diverse projects, boasting high precision, recall rates, and AUC values, thereby outperforming analogous algorithms. This aligns with the perspectives espoused by Asadullah et al. 37 . Particular emphasis was placed on assessing the impact of variables such as budget investment levels, team experience, and project size on algorithmic performance. Notably, heightened budget investment and extensive team experience positively influenced the results, with project size exerting a comparatively minor impact. Regression analysis elucidates the magnitude and interplay of these factors, underscoring the predominant influence of budget investment and team experience on RA outcomes, whereas project size assumes a relatively marginal role. This underscores the imperative for decision-makers in projects to meticulously consider the interrelationships between these factors for a more precise assessment of project risks, echoing the sentiments expressed by Testorelli et al. 38 .

In sum, this paper furnishes a holistic comprehension of the Naive Bayes algorithm’s application in project risk prediction, offering robust guidance for practical project management. The paper’s tangible applications are chiefly concentrated in the realm of RA and management for SRPs. Such insights empower managers in SRPs to navigate risks with scientific acumen, thereby enhancing project success rates and performance. The paper advocates several strategic measures for SRPs management: prioritizing resource adjustments and team training to elevate the professional skill set of team members in coping with the impact of team experience on risks; implementing project scale management strategies to mitigate potential risks by detailed project stage division and stringent project planning; addressing technical difficulty as a pivotal risk factor through assessment and solution development strategies; incorporating project cycle adjustment and flexibility management to accommodate fluctuations and mitigate associated risks; and ensuring the integration of data quality management strategies to bolster data reliability and enhance model accuracy. These targeted risk responses aim to improve the likelihood of project success and ensure the seamless realization of project objectives.

Achievements

In this paper, the application of Naive Bayesian algorithm in RA of SRPs is deeply explored, and the influence of various factors on RA results and their relationship is comprehensively investigated. The research results fully prove the good accuracy and applicability of Naive Bayesian algorithm in RA of science and technology projects. Through probability estimation, the risk level of the project can be estimated more accurately, which provides a new decision support tool for the project manager. It is found that budget input and team experience are the most significant factors affecting the RA results, and their regression coefficients are 0.68 and 0.51 respectively. However, the influence of project scale on the RA results is relatively small, and its regression coefficient is 0.31. Especially in the case of low team experience, the budget input has a more significant impact on the RA results. However, it should also be admitted that there are some limitations in the paper. First, the case data used is limited and the sample size is relatively small, which may affect the generalization ability of the research results. Second, the factors concerned may not be comprehensive, and other factors that may affect RA, such as market changes and policies and regulations, are not considered.

The paper makes several key contributions. Firstly, it applies the Naive Bayes algorithm to assess the risks associated with SRPs, proposing the TANB and validating its effectiveness empirically. The introduction of the TANB model broadens the application scope of the Naive Bayes algorithm in scientific research risk management, offering novel methodologies for project RA. Secondly, the study delves into the impact of various factors on RA for SRPs through MLR analysis, highlighting the significance of budget investment and team experience. The results underscore the positive influence of budget investment and team experience on RA outcomes, offering valuable insights for project decision-making. Additionally, the paper examines the interaction between team experience and budget investment, revealing a nuanced relationship between the two in RA. This finding underscores the importance of comprehensively considering factors such as team experience and budget investment in project decision-making to achieve more accurate RA. In summary, the paper provides crucial theoretical foundations and empirical analyses for SRPs risk management by investigating RA and its influencing factors in depth. The research findings offer valuable guidance for project decision-making and risk management, bolstering efforts to enhance the success rate and efficiency of SRPs.

This paper distinguishes itself from existing research by conducting an in-depth analysis of the intricate interactions among various factors, offering more nuanced and specific RA outcomes. The primary objective extends beyond problem exploration, aiming to broaden the scope of scientific evaluation and research practice through the application of statistical language. This research goal endows the paper with considerable significance in the realm of science and technology project management. In comparison to traditional methods, this paper scrutinizes project risk with greater granularity, furnishing project managers with more actionable suggestions. The empirical analysis validates the effectiveness of the proposed method, introducing a fresh perspective for decision-making in science and technology projects. Future research endeavors will involve expanding the sample size and accumulating a more extensive dataset of SRPs to enhance the stability and generalizability of results. Furthermore, additional factors such as market demand and technological changes will be incorporated to comprehensively analyze elements influencing the risks of SRPs. Through these endeavors, the aim is to provide more precise and comprehensive decision support to the field of science and technology project management, propelling both research and practice in this domain to new heights.

Limitations and prospects

This paper, while employing advanced methodologies like TANB models, acknowledges inherent limitations that warrant consideration. Firstly, like any model, TANB has its constraints, and predictions in specific scenarios may be subject to these limitations. Subsequent research endeavors should explore alternative advanced machine learning and statistical models to enhance the precision and applicability of RA. Secondly, the focus of this paper predominantly centers on the RA for SRPs. Given the unique characteristics and risk factors prevalent in projects across diverse fields and industries, the generalizability of the paper results may be limited. Future research can broaden the scope of applicability by validating the model across various fields and industries. The robustness and generalizability of the model can be further ascertained through the incorporation of extensive real project data in subsequent research. Furthermore, future studies can delve into additional data preprocessing and feature engineering methods to optimize model performance. In practical applications, the integration of research outcomes into SRPs management systems could provide more intuitive and practical support for project decision-making. These avenues represent valuable directions for refining and expanding the contributions of this research in subsequent studies.

Data availability

All data generated or analysed during this study are included in this published article [and its Supplementary Information files].

Moshtaghian, F., Golabchi, M. & Noorzai, E. A framework to dynamic identification of project risks. Smart and sustain. Built. Environ. 9 (4), 375–393 (2020).

Google Scholar  

Nunes, M. & Abreu, A. Managing open innovation project risks based on a social network analysis perspective. Sustainability 12 (8), 3132 (2020).

Article   Google Scholar  

Elkhatib, M. et al. Agile project management and project risks improvements: Pros and cons. Mod. Econ. 13 (9), 1157–1176 (2022).

Fridgeirsson, T. V. et al. The VUCAlity of projects: A new approach to assess a project risk in a complex world. Sustainability 13 (7), 3808 (2021).

Salahuddin, T. Numerical Techniques in MATLAB: Fundamental to Advanced Concepts (CRC Press, 2023).

Book   Google Scholar  

Awais, M. & Salahuddin, T. Radiative magnetohydrodynamic cross fluid thermophysical model passing on parabola surface with activation energy. Ain Shams Eng. J. 15 (1), 102282 (2024).

Awais, M. & Salahuddin, T. Natural convection with variable fluid properties of couple stress fluid with Cattaneo-Christov model and enthalpy process. Heliyon 9 (8), e18546 (2023).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Guan, L., Abbasi, A. & Ryan, M. J. Analyzing green building project risk interdependencies using Interpretive Structural Modeling. J. Clean. Prod. 256 , 120372 (2020).

Gaudenzi, B. & Qazi, A. Assessing project risks from a supply chain quality management (SCQM) perspective. Int. J. Qual. Reliab. Manag. 38 (4), 908–931 (2021).

Lee, K. T., Park, S. J. & Kim, J. H. Comparative analysis of managers’ perception in overseas construction project risks and cost overrun in actual cases: A perspective of the Republic of Korea. J. Asian Archit. Build. Eng. 22 (4), 2291–2308 (2023).

Garai-Fodor, M., Szemere, T. P. & Csiszárik-Kocsir, Á. Investor segments by perceived project risk and their characteristics based on primary research results. Risks 10 (8), 159 (2022).

Senova, A., Tobisova, A. & Rozenberg, R. New approaches to project risk assessment utilizing the Monte Carlo method. Sustainability 15 (2), 1006 (2023).

Tiwari, P. & Suresha, B. Moderating role of project innovativeness on project flexibility, project risk, project performance, and business success in financial services. Glob. J. Flex. Syst. Manag. 22 (3), 179–196 (2021).

de Araújo, F., Lima, P., Marcelino-Sadaba, S. & Verbano, C. Successful implementation of project risk management in small and medium enterprises: A cross-case analysis. Int. J. Manag. Proj. Bus. 14 (4), 1023–1045 (2021).

Obondi, K. The utilization of project risk monitoring and control practices and their relationship with project success in construction projects. J. Proj. Manag. 7 (1), 35–52 (2022).

Atasoy, G. et al. Empowering risk communication: Use of visualizations to describe project risks. J. Constr. Eng. Manage. 148 (5), 04022015 (2022).

Dandage, R. V., Rane, S. B. & Mantha, S. S. Modelling human resource dimension of international project risk management. J. Global Oper. Strateg. Sourcing 14 (2), 261–290 (2021).

Wang, L. et al. Applying social network analysis to genetic algorithm in optimizing project risk response decisions. Inf. Sci. 512 , 1024–1042 (2020).

Marx-Stoelting, P. et al. A walk in the PARC: developing and implementing 21st century chemical risk assessment in Europe. Arch. Toxicol. 97 (3), 893–908 (2023).

Awais, M., Salahuddin, T. & Muhammad, S. Evaluating the thermo-physical characteristics of non-Newtonian Casson fluid with enthalpy change. Thermal Sci. Eng. Prog. 42 , 101948 (2023).

Article   CAS   Google Scholar  

Awais, M., Salahuddin, T. & Muhammad, S. Effects of viscous dissipation and activation energy for the MHD Eyring-Powell fluid flow with Darcy-Forchheimer and variable fluid properties. Ain Shams Eng. J. 15 (2), 102422 (2024).

Yang, L., Lou, J. & Zhao, X. Risk response of complex projects: Risk association network method. J. Manage. Eng. 37 (4), 05021004 (2021).

Acebes, F. et al. Project risk management from the bottom-up: Activity Risk Index. Cent. Eur. J. Oper. Res. 29 (4), 1375–1396 (2021).

Siyal, S. et al. They can’t treat you well under abusive supervision: Investigating the impact of job satisfaction and extrinsic motivation on healthcare employees. Rationality Society 33 (4), 401–423 (2021).

Chen, D., Wawrzynski, P. & Lv, Z. Cyber security in smart cities: A review of deep learning-based applications and case studies. Sustain. Cities Soc. 66 , 102655 (2021).

Zhao, M. et al. Pythagorean fuzzy TODIM method based on the cumulative prospect theory for MAGDM and its application on risk assessment of science and technology projects. Int. J. Fuzzy Syst. 23 , 1027–1041 (2021).

Suresh, K. & Dillibabu, R. A novel fuzzy mechanism for risk assessment in software projects. Soft Comput. 24 , 1683–1705 (2020).

Akhavan, M., Sebt, M. V. & Ameli, M. Risk assessment modeling for knowledge based and startup projects based on feasibility studies: A Bayesian network approach. Knowl.-Based Syst. 222 , 106992 (2021).

Guan, L., Abbasi, A. & Ryan, M. J. A simulation-based risk interdependency network model for project risk assessment. Decis. Support Syst. 148 , 113602 (2021).

Vujović, V. et al. Project planning and risk management as a success factor for IT projects in agricultural schools in Serbia. Technol. Soc. 63 , 101371 (2020).

Muñoz-La Rivera, F., Mora-Serrano, J. & Oñate, E. Factors influencing safety on construction projects (FSCPs): Types and categories. Int. J. Environ. Res. Public Health 18 (20), 10884 (2021).

Article   PubMed   PubMed Central   Google Scholar  

Nguyen, P. T. & Nguyen, P. C. Risk management in engineering and construction: A case study in design-build projects in Vietnam. Eng. Technol. Appl. Sci. Res 10 , 5237–5241 (2020).

Nguyen PT, Le TT. Risks on quality of civil engineering projects-an additive probability formula approach//AIP Conference Proceedings. AIP Publishing, 2798(1) (2023).

Nguyen, P.T., Phu, P.C., Thanh, P.P., et al . Exploring critical risk factors of office building projects. 8 (2), 0309–0315 (2020).

Nguyen, H. D. & Macchion, L. Risk management in green building: A review of the current state of research and future directions. Environ. Develop. Sustain. 25 (3), 2136–2172 (2023).

He, S. et al. Risk assessment of oil and gas pipelines hot work based on AHP-FCE. Petroleum 9 (1), 94–100 (2023).

Asadullah, M. et al. Evaluation of machine learning techniques for hypertension risk prediction based on medical data in Bangladesh. Indones. J. Electr. Eng. Comput. Sci. 31 (3), 1794–1802 (2023).

Testorelli, R., de Araujo, F., Lima, P. & Verbano, C. Fostering project risk management in SMEs: An emergent framework from a literature review. Prod. Plan. Control 33 (13), 1304–1318 (2022).

Download references

Author information

Authors and affiliations.

Institute of Policy Studies, Lingnan University, Tuen Mun, 999077, Hong Kong, China

Xuying Dong & Wanlin Qiu

You can also search for this author in PubMed   Google Scholar

Contributions

Xuying Dong and Wanlin Qiu played a key role in the writing of Risk Assessment of Scientific Research Projects and the Relationship between Related Factors Based on Naive Bayes Algorithm. First, they jointly developed clearly defined research questions and methods for risk assessment using the naive Bayes algorithm at the beginning of the research project. Secondly, Xuying Dong and Wanlin Qiu were responsible for data collection and preparation, respectively, to ensure the quality and accuracy of the data used in the research. They worked together to develop a naive Bayes algorithm model, gain a deep understanding of the algorithm, ensure the effectiveness and performance of the model, and successfully apply the model in practical research. In the experimental and data analysis phase, the collaborative work of Xuying Dong and Wanlin Qiu played a key role in verifying the validity of the model and accurately assessing the risks of the research project. They also collaborated on research papers, including detailed descriptions of methods, experiments and results, and actively participated in the review and revision process, ensuring the accuracy and completeness of the findings. In general, the joint contribution of Xuying Dong and Wanlin Qiu has provided a solid foundation for the success of this research and the publication of high-quality papers, promoted the research on the risk assessment of scientific research projects and the relationship between related factors, and made a positive contribution to the progress of the field.

Corresponding author

Correspondence to Wanlin Qiu .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary information., rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Dong, X., Qiu, W. A case study on the relationship between risk assessment of scientific research projects and related factors under the Naive Bayesian algorithm. Sci Rep 14 , 8244 (2024). https://doi.org/10.1038/s41598-024-58341-y

Download citation

Received : 30 October 2023

Accepted : 27 March 2024

Published : 08 April 2024

DOI : https://doi.org/10.1038/s41598-024-58341-y

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Naive Bayesian algorithm
• Scientific research projects
• Risk assessment
• Factor analysis
• Probability estimation
• Decision support
• Data-driven decision-making

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

This paper is in the following e-collection/theme issue:

Published on 4.4.2024 in Vol 8 (2024)

Exploring the Implementation of Shared Decision-Making Involving Health Coaches for Diabetes and Hypertension Self-Management: Qualitative Study

Authors of this article:

Original Paper

• Sungwon Yoon 1 , PhD   ; 
• Chao Min Tan 2 , MScR   ; 
• Jie Kie Phang 2 , MPH   ; 
• Venice Xi Liu 2 , BSc   ; 
• Wee Boon Tan 2 , BSc   ; 
• Yu Heng Kwan 1 , PhD   ; 
• Lian Leng Low 2 , MBBS, MMed  

1 Duke-NUS Medical School, Singapore, Singapore

2 Centre for Population Health Research and Implementation, SingHealth Regional Health System, Singapore, Singapore

Corresponding Author:

Sungwon Yoon, PhD

Duke-NUS Medical School

8 College Rd

Singapore, 169857

Phone: 65 65167666

Email: [email protected]

Background: An emerging focus on person-centered care has prompted the need to understand how shared decision-making (SDM) and health coaching could support self-management of diabetes and hypertension.

Objective: This study aims to explore preferences for the scope of involvement of health coaches and health care professionals (HCPs) in SDM and the factors that may influence optimal implementation of SDM from the perspectives of patients and HCPs.

Methods: We conducted focus group discussions with 39 patients with diabetes and hypertension and 45 HCPs involved in their care. The main topics discussed included the roles of health coaches and HCPs in self-management, views toward health coaching and SDM, and factors that should be considered for optimal implementation of SDM that involves health coaches. All focus group discussions were audio recorded, transcribed verbatim, and analyzed using thematic analysis.

Results: Participants agreed that the main responsibility of HCPs should be identifying the patient’s stage of change and medication education, while health coaches should focus on lifestyle education, monitoring, and motivational conversation. The health coach was seen to be more effective in engaging patients in lifestyle education and designing goal management plans as health coaches have more time available to spend with patients. The importance of a health coach’s personal attributes (eg, sufficient knowledge of both medical and psychosocial management of disease conditions) and credentials (eg, openness, patience, and empathy) was commonly emphasized. Participants viewed that addressing the following five elements would be necessary for the optimal implementation of SDM: (1) target population (newly diagnosed and less stable patients), (2) commitment of all stakeholders (discrepancy on targeted times and modality), (3) continuity of care (familiar faces), (4) philosophy of care (person-centered communication), and (5) faces of legitimacy (physician as the ultimate authority).

Conclusions: The findings shed light on the appropriate roles of health coaches vis-à-vis HCPs in SDM as perceived by patients and HCPs. Findings from this study also contribute to the understanding of SDM on self-management strategies for patients with diabetes and hypertension and highlight potential opportunities for integrating health coaches into the routine care process.

Introduction

Lifestyle factors, including engaging in adequate physical activity and consuming a healthy diet, are important in the management of diabetes and hypertension [ 1 - 3 ]. However, many patients with diabetes and hypertension do not achieve optimal lifestyle targets [ 4 - 6 ]. Effective interventions to help patients with diabetes and hypertension should be tailored for each patient, and this may be achieved through engaging in shared decision-making (SDM). In line with the growing emphasis on person-centered care, SDM is recognized as a way to empower patients with chronic diseases such as diabetes and hypertension. SDM involves patients and health care professionals (HCPs) collaboratively making a health care decision after discussing the treatment, management, and support packages and considering the patient’s preferences, priorities, and goals [ 7 - 9 ]. Although there is limited research on the effect of SDM on clinical outcomes for diabetes and hypertension [ 10 ], studies invariably suggest that SDM makes a positive difference to patients in their care. This includes better treatment adherence, increased patient coping, improved knowledge attainment, higher levels of patient satisfaction, and greater empowerment [ 10 - 12 ].

In addition to situations requiring treatment decisions, SDM may be useful in supporting healthy behavior change [ 13 , 14 ]. A feasibility study on decision tools in primary care to help initiate lifestyle change among patients with or at risk of coronary heart disease has shown the potential beneficial effects of paper-based tools for SDM in initiating behavior change [ 15 ]. Another feasibility study of an internet-based decision aid to encourage lifestyle change and adherence among people at moderate or high risk of coronary heart disease was found to increase participants’ ability to make clear decisions about making changes [ 16 , 17 ]. However, it was suggested that further impact may have been achieved if more comprehensive implementation strategies had been available for the interventions [ 16 , 17 ].

Despite the evidence supporting SDM, it is not widely practiced in clinical settings due to several reasons, such as low patient self-efficacy, a power imbalance between patients and physicians, and HCP’s limited time and knowledge [ 12 , 18 - 20 ]. To overcome these communication and resource barriers, several studies proposed the inclusion of health coaches to facilitate the SDM process in chronic care through continuous counseling dialogue with patients and exploration of patients’ situations and preferences in order to make informed decisions together on treatment and lifestyle [ 21 , 22 ]. Health coaches are individuals who aid patients in gaining the knowledge and confidence necessary to become engaged in their care and promote communication and collaborative decisions between patients and HCPs [ 23 ]. The practice of health coaching can differ in the type of coach, their training, and their level of involvement [ 24 ]. Nonetheless, randomized controlled trials have shown that health coaching can lead to enhanced self-management of diabetes and hypertension [ 25 , 26 ]. Furthermore, the experiences of patients and HCPs were found to be largely positive [ 27 - 29 ].

Although literature has documented the effectiveness and feasibility of SDM and health coaching, the evidence primarily comes from patients and HCPs who were willing to take part in or complete interventions [ 30 - 32 ]. There are fewer studies that have examined the viewpoints of potential end users’ perspectives regarding their preferences for and expectations of SDM [ 33 , 34 ], as well as the role and relationships of health coaches in patient care practice [ 35 ]. Obtaining the buy-in of patients and HCPs is crucial when developing a robust care model. To this end, we conducted a study to gather the viewpoints of patients and HCPs to gain insight on developing strategies for SDM programs that incorporate nurse-trained health coaches in primary care.

The aim of this formative study was to explore the perspectives and preferences of HCPs and patients with diabetes and hypertension concerning the respective professional roles of health coaches and HCPs in SDM, as well as the factors that should be considered for optimal implementation of a SDM model that involves patients, health coaches, and HCPs.

This was a qualitative study, reported following the Consolidated Criteria for Reporting Qualitative Research (COREQ) Guidelines [ 36 ].

Setting and Participants

This study was conducted in Singapore, a multiethnic city state where the majority of the population (80%) obtains health care from the public health care system [ 37 ]. Participants were mainly from SingHealth Cluster, which is the largest regional health care system in Singapore, offering a complete range of medical care for patients, including those with diabetes and hypertension. Eligible patient participants were those aged 40 years and older, diagnosed with diabetes and hypertension, and attending public primary care clinics. We identified eligible patients from a study cohort that investigates the clinical and cost-effectiveness of a behavioral intervention delivered through mobile health [ 38 ]. The participants were then approached through a phone call with the study aim and methods explained. On the other hand, eligible HCP participants were those responsible for managing patients with diabetes and hypertension in public primary care clinics, step-down care, and secondary or tertiary centers with at least 1 year of experience. We approached potential HCP participants through email and provided background information. Purposive sampling was adopted in terms of age (patients) and clinical experience (HCPs) to maximize diversity of perspectives.

Data Collection

We conducted focus group discussions (FGDs) with participants between February 2022 and May 2022. A semistructured topic guide was developed and subsequently pilot-tested to facilitate discussions on the roles of health coaches and HCPs in self-management, views toward health coaching and SDM, and factors that should be considered for optimal implementation of SDM that involves health coaches. All FGDs were carried out through web-based videoconferencing by facilitators (CMT and WBT) who were trained in social sciences and qualitative research and did not have a personal relationship with the participants. Each FGD session lasted approximately 90 minutes for patient participants and 60 minutes for HCP participants. No repeat interviews were conducted, and transcripts were not returned to participants for further input. Data collection and analysis were an iterative process that continued until no new themes emerged. Field notes were taken to support the contextual interpretation of the data.

Data Analysis

All FGDs were audio-recorded and transcribed verbatim. Transcripts were checked for accuracy and thematically analyzed. A total of 2 coders (CMT and WBT) were assigned to code the patient FGDs, while 2 other coders (JKP and VXL) were assigned to code the HCP FGDs. The team adopted the 6 steps to thematic analysis suggested by Braun and Clarke [ 39 ], in which the coders first familiarize themselves with the data and generate initial codes independently before collating the codes into potential themes together. The themes were constantly reviewed, refined, and reclassified to ensure the best fit of themes to the data. Discrepancies between coders were resolved through consensus meetings involving all study team members. The FGDs were conducted until thematic saturation occurred at the 17th and 18th FGDs with patients and HCPs, respectively. Storing and managing data during data analysis was done using NVivo (version 12; QSR International).

Ethical Considerations

This study was approved by the SingHealth Centralized Institutional Review Board (2019/2468). Participants provided verbal informed consent before the study began. The study team maintained data confidentiality by redacting personally identifiable information from interview transcripts and generating unique study identifiers, which were linked to participant identifiable information only through a password-protected file. Participants were reimbursed SG $60 (US $44.70) to defray the cost of their participation in this research.

Participant Characteristics

Out of 89 patients approached, 39 were recruited for the study, with the most frequent reasons for the decline being difficulties in participating in web-based interviews or schedule unavailability. The recruited patients participated in 17 FGDs. Their age ranged from 43 to 68 years, with 74% (29/39) being male candidates and 64% (25/39) being Chinese. Concurrently, we approached 52 HCPs and recruited 45 HCPs who participated in 18 FGDs, with schedule unavailability as the main reason for the decline. Approximately 65% (29/45) were clinicians, followed by 16% (7/45) being nurses. The range of clinical experience in managing chronic diseases of the HCPs was from 1 year to 28 years ( Table 1 ).

a HCP: health care professional.

b FGD: focus group discussion.

c N/A: not applicable.

Through our analysis, we identified three themes that represent (1) the participants’ perspectives concerning the professional roles of health coaches and HCPs in SDM, (2) the perceived importance of health coaches’ credentials and attributes, and (3) a total of 5 essential elements to be considered for optimal implementation of SDM. Figure 1 presents a visual summary that suggests how SDM involving health coaches could be applied in clinical settings to facilitate diabetes and hypertension self-management, based on the findings.

Perceived Preference for Professional Roles of HCP and Health Coach in SDM

While there were some commonalities in the roles of HCPs and health coaches in the SDM model of care, a distinction in the extent of their responsibilities was evident.

Patient Education

Many of the patient participants and HCP participants recognized the shared responsibilities of HCPs and health coaches in improving patients’ understanding of their conditions in order to decide on the self-management strategies that fit best according to their individual situations and capacities. However, their expectations of the specific roles that HCPs and health coaches would play in patient education differed. The health coach was seen to be more effective in engaging patients in lifestyle education and informing patients on healthy lifestyle choices, while the HCPs are expected to educate patients on medication and alternative treatment options for their conditions. The participants explained that the difference in expectations was based on the perceived amount of time availability the professionals have with the patients and the background of the professionals.

[The health coach] is a single point of contact that I can refer to, who is an expert in this area, and I can leverage on that to achieve the goals that I want. Knowing that there is somebody associated with you, and you can engage with it helps a lot. Whether it’s about physical activity, food intake, [or] the discipline you need to get in order to achieve the goals, if I know I can reach out to someone to talk about it, it will definitely make a difference. [Patient 37, Patient FGD #13]

I will tell her what the best option as a physician is, based on our guidelines. However, I will tell her other possible options if let’s say she doesn’t want the recommended option. As a physician, it’s our duty to tell patients the options they have and the pros and cons of each option. [HCP 36, HCP FGD #12]

Goal Setting

Both patient participants and HCP participants agreed that setting actionable goals would be crucial to improving clinical outcomes. However, the brief consultation sessions in primary care settings were inadequate for patients to develop personalized care plans with their HCPs. Thus, patient participants saw the value of involving the health coach to work with them to set actionable goals while taking into consideration their personal circumstances in order to devise an appropriate self-management plan that aligns with the expectations of the HCPs. At the same time, HCPs suggested that the health coach plays an important role in identifying any lapses and bridging the gaps between the treatment offered by HCPs and the health care preferences and goals of the patients.

A doctor’s goals may be different from a patient’s goals. Sometimes it’s hard for us to assess the ideas, values, and preferences during our short 10 to 15 minutes [consultations]. If the health coach informs us, that will be good so that the patient, doctor, and health coach can be on the same page to help the patient achieve his or her goals. [HCP 36, HCP FGD #12]

[A health coach should] connect with the patient to discover what the problems are and also be aware of the patient’s environment. Then align these with the expectations that HCPs might have of the patient. So, the health coach’s duty is basically to identify all of these, in order to make things easier. [Patient 25, Patient FGD #10]

Patient Empowerment

According to patients, “feeling empowered” entailed remaining motivated to engage in behaviors that promote their health. To this end, they preferred to work with the health coaches and mutually establish achievable goals related to behaviors such as diet and exercise. The partnership that patients form with their health coach allows patients to feel supported throughout their self-management journey, which motivates them to be more engaged and adherent to the chosen care plan. Likewise, HCP participants recognized the significance of such a partnership, in which the patients are free to express their expectations and wishes for care while deciding on a care plan that would be tailored to their individual needs and preferences.

It would be more of a motivating aspect because the health coach can help set specific Agendas, target them, and communicate with patients…In this way, we are more motivated to try hard to meet the targets. Because when seeing the doctor, you know, he/she will just tell you to lose weight but [the] health coach can motivate you for sure. [Patient 25, Patient FGD #10]

In an ideal setting, [health coaches] need to understand what their targets are, and what the patient thinks [of] their health conditions. From there, see what the patient is willing and able to do and what their plans are going forward. [HCP 09, HCP FGD #1]

Importance of Health Coach’s Credentials and Attributes

Both patient participants and HCP participants emphasized the importance of health coaches’ credentials and attributes that would influence their acceptance of health coaching as part of routine patient care. Most participants mentioned that a desirable health coach would need to possess sufficient knowledge of both medical and psychosocial management of diabetes and hypertension so that health coaches could offer appropriate guidance to patients.

Health coach should be medically trained to give correct advice. I mean, apart from medication and disease management [which is the ambit of HCPs], the health coach should be able to provide psychosocial counseling. [HCP 38, HCP FGD #8]

A health coach’s personal attributes were equally stressed; most participants noted that a health coach should demonstrate positive personality traits such as openness, patience, and empathy to effectively improve a patient’s willingness to consider the recommended health practices as suggested by the health coach.

How do I become open to the health coach? Firstly, I think the health coaches themselves must be very caring and full of empathy, to disarm all the unhappiness of the patient, maybe then patients will be willing to tell the coach about their story. [Patient 51, Patient FGD #17]

5 Essential Elements for Optimal Implementation of SDM

A total of 5 elements have been identified that participants believe should be considered for optimal implementation of the SDM model of care involving health coaches.

Target Population

Most participants (patients and HCPs) shared that SDM should target a certain segment of patients to maximize its benefits. They believed that individuals with newly diagnosed chronic conditions and those with poorly controlled diabetes and hypertension would benefit more from the proposed SDM model of care involving health coaches. This is because these patient groups may have inadequate knowledge or support to effectively manage their conditions. Thus, the education, guidance, and discussion provided by the health coach could prepare them with the necessary knowledge to begin their self-management journey effectively.

Maybe this will help for those who just are diagnosed with diabetes, and you know, those at a loss or don’t know what to do. But if you’re working with a “seasoned” patient, they know what to do, what to expect and all. [Patient 17, Patient FGD #3]

Commitment of all Stakeholders

Participants asserted that in order to effectively facilitate SDM and achieve the mutually established goals of improving self-management, all stakeholders involved in this communication process should be equally committed. They expressed commitment to engaging in open dialogues and establishing partnerships among all 3 parties, including patients, HCPs, and health coaches, at targeted times and through preferred modalities. While patient participants expressed a preference for frequent check-ins with the health coach (eg, monthly), HCP participants felt that meetings between health coaches and HCPs should be arranged on a case-by-case basis, depending on the urgency and complexity of the patient’s conditions due to their high workload. Most participants were open to the varied modes of communication, including in-person and web-based means to facilitate the SDM to cater to different situational needs (eg, web-based call for a brief check-in and an in-person call for an in-depth dialogue).

I think doctors need to focus on the very complex cases for communication with health coaches. We can’t put too much effort into handling every single chronic patient because the workload will be too high. [HCP 02, HCP FGD #2]

Continuity of Care

Care continuity was identified by participants as a critical factor in facilitating SDM. Both HCP participants and patient participants expressed a preference for having the same health coach for follow-up appointments to foster a sense of rapport, continuity of care, and motivate patients to carry out the plan of care. Additionally, patient participants felt that following up with the same health coach would help them build trust with their health coach and disclose their concerns, strengths, and limitations in their self-management journey.

In a polyclinic setting, the doctor will change every appointment. If the same health coach can provide consistent support and counseling to the patient, and if the patient has someone who is checking on him, he will want to take better care of his chronic conditions. [HCP 02, HCP FGD #2]

I think the health coach should be someone that at least I know and have some level of good relationship…so I can treat the coach like a friend and open up. [Patient 32, Patient FGD #9]

Philosophy of Care

Participants highlighted the importance of open communication and person-centered inquiry to facilitate mutual understanding among all parties involved in SDM. They valued strategies that could allow patients to set personal goals, negotiate, and discuss challenges. Thus, the philosophy of care was focused on supporting patients to make informed choices and engaging patients in discussion to develop a care plan that is tailored to individuals, as opposed to simply offering generic health advice that may not be as effective in motivating patients.

Ask the patient what they want first, because if it’s not something that they want, it’s not likely that they will cooperate with us [HCPs or health coaches] even though it is what we want from them. [HCP 49, HCP FGD #18]

Faces of Legitimacy

Notably, our interviews revealed that patients generally prioritized the advice given by their physicians because they perceived their physician’s advice to be more important and reliable than that of health coaches and other HCPs (eg, dieticians, nurses, and physiotherapists). This could pose a challenge to SDM when consensus cannot be built about the preferred self-management plan among all parties involved and patients are less receptive to exploring other treatment options and recommendations unless they are endorsed by a physician.

I’ll still take the final instructions from the doctor. Frankly speaking, the health coach might be knowledgeable in terms of some medical information, but they are still not reliable. [Patient 17, Patient FGD #3]

Sometimes it also depends on which healthcare provider is approaching the patient. A lot of times, our patients defer to what we intended so if the doctors don’t say like “Oh you need to do this” then they won’t really cooperate because doctor’s recommendations take precedence over whatever other professionals are seeking to help. [HCP 49, HCP FGD #18]

To mitigate this, some of our HCP participants, who are physicians, suggested that reinforcement and endorsing the advice from the health coaches and other HCPs during follow-up appointments would be important in increasing patients’ trust in other providers and promoting effective communication in SDM.

One thing that physicians like me could do is to reinforce what the health coach has taught the patients. Then the patient would realize that, Oh yes, that [advice given by health coach] is very important. [HCP 49, HCP FGD #18]

Principle Findings

This study explored the preferences and perspectives of both patients and HCPs on how SDM involving health coaches could help patients make informed decisions about their health and improve self-management of their diabetes and hypertension. While some perspectives varied across patients and HCPs, we identified three unified themes, including (1) the perceived preference for and expectation of the roles of HCPs and health coaches in SDM, (2) the importance of health coaches’ credentials and attributes, and (3) the 5 elements necessary for effective implementation of SDM. The findings gained from this study offered key insights to support efforts to optimally implement SDM involving health coaches for patients with chronic conditions.

The lack of patient education [ 40 , 41 ] and psychosocial support [ 42 - 45 ] can hinder patients’ ability to self-manage their diabetes and hypertension, which eventually results in suboptimal control and negative health outcomes. In this study, patient participants and HCP participants agreed that the primary responsibility of a health coach is to educate patients on healthy lifestyle choices and provide several self-management options before setting actionable goals that align with the patient’s needs and preferences, while an HCP is expected to provide medication education and offer alternative treatment options for their conditions. In this regard, SDM provides a platform for patients, health coaches, and HCPs to engage in conversations that enable information to be shared and address each party’s expectations for care [ 13 ]. In addition, the involvement of health coaches in SDM has been shown to improve self-management by fostering greater patient involvement in their care and designing care plans that take into account their unique treatment goals and preferences [ 46 ]. When patients have a better understanding of their options and have the autonomy to express their preferences and wishes for care, they are more likely to be satisfied with the eventual plan of care and adhere to it [ 47 ]. Moreover, our findings showed that the involvement of a health coach would offer patients a sense of support through their self-management journey and motivate them to take charge of their diabetes and hypertension self-care. This finding reflects previous studies that found integrative health coaching improved patients’ psychosocial outcomes, resulting in reduced perceived barriers to self-management, enhanced perceptions of social support, and ultimately improved clinical outcomes in patients with diabetes and hypertension [ 48 - 50 ]. The results of this study offer valuable insights into the distinct responsibilities of health coaches and HCPs in SDM and chronic disease management and highlight potential areas of emphasis in patient coaching, workflow, and collaborative efforts.

A common theme running through the FGDs was participants’ keen interest in the credentials and characteristics of health coaches. Participants stressed the importance of a health coach’s positive attitude and knowledge in both the medical and psychosocial aspects of disease management in order to engage in a partner relationship in SDM. Indeed, a health coach’s professional expertise and personal traits, such as openness and empathy, can improve the therapeutic relationship and ultimately enhance self-management skills [ 30 , 51 ]. Therefore, it is essential for health coaches to receive appropriate clinical training and education as well as possess a strong capacity for empathy [ 52 ]. Beyond the credentials and characteristics of health coaches, studies on SDM also stressed the importance of a supportive and caring environment with adequate interaction time as key aspects of the patient-provider partnership in chronic care [ 53 , 54 ]. Many of the physicians interviewed in this study often mentioned that they are unable to cover all aspects of the patient’s self-care due to the brief consultation sessions they have with the patients. The time required for information sharing and clarifying patients’ values, needs, and preferences could impact the already-pressured clinical setting [ 13 ]. Therefore, it was suggested that the involvement of health coaches in SDM would help to prioritize discussions about specific aspects of diabetes and hypertension self-management (ie, medical and lifestyle) and allow patients to benefit from the enhanced support from their health coaches, who have more time to work with them on modifying their lifestyle and achieving better control of their conditions. Our findings underscore the significance of health coaches’ competencies to ensure that health coaches can fulfill their core responsibilities and the potential benefit of involving health coaches in SDM to support patients further in their self-management journey.

Lastly, our participants believed that newly diagnosed or less stable patients could benefit the most from SDM involving health coaches and emphasized the importance of continuity of care through the same coach. They also recognized that open communication and person-centered inquiry would be crucial for improving the quality of SDM. Indeed, previous studies demonstrated that open communication and consistent coaching improved decision quality, knowledge, and risk perception among patients with diabetes and hypertension [ 10 , 55 ]. Despite these findings supporting the use of SDM to support healthy behavior change among patients with chronic disease, patients in this study still held a strong belief in the traditional approach of “doctor knows best” (faces of legitimacy), with many patients relying disproportionately on physicians for decisions [ 56 ]. Patients’ reliance on physicians for decisions may pose a challenge to the SDM process involving health coaches since patients may prioritize the advice of their physicians over that of health coaches. For health care institutions that wish to implement SDM for chronic disease management, we suggest distinguishing the roles of HCPs and health coaches in chronic disease management to ensure successful implementation of SDM. Institutions can also consider educating patients about the unique and valuable contributions that health coaches can make in their care [ 57 ] to reinforce their trust in health coaches. As observed in this study, the health coaches’ involvement in SDM was important in offering personalized support to patients to modify their lifestyle and self-management in order to achieve behavioral change and gain better control of their diabetes and hypertension. Future research should aim to identify factors that affect patients’ engagement and trust with health coaches to enable successful implementation of this SDM model for chronic disease management.

Limitations

This study has a few limitations. The perspectives of health coaches were not included, which may limit the comprehensiveness of the results. Furthermore, participants’ preferences and expectations were not examined by subgroups such as HCP’s professional roles or patients’ confidence levels in self-management and cultural backgrounds. Further research focusing on these aspects may prove useful for a richer understanding of the SDM implementation. Despite these limitations, this study provided valuable insights into the SDM model of care, highlighting how patients, HCPs, and health coaches can collaborate and the factors needed to be considered for robust implementation of the SDM for patients with diabetes and hypertension.

Conclusions

Our findings examined the viewpoints of potential end users’ perspectives regarding their preferences for and expectations of SDM from patients and HCPs. Our analysis identified the appropriate roles of health coaches vis-à-vis HCPs in SDM and underscored the importance of a health coach’s credentials and personal attributes. At the same time, the five elements for optimal implementation of SDM can be used to guide future efforts to contextualize SDM and integrate health coaches into routine primary care to support diabetes and hypertension treatment.

Acknowledgments

The authors thank the participants who generously provided their time and shared their perspectives. We would also like to thank Cassandra and Valen from the Centre for Population Health Research and Implementation for their support for the study. This study is supported by the National Research Foundation, Singapore, under its AI Singapore Programme (AISG-GC-2019-001-2A). This study is also supported by the Singapore Ministry of Health’s National Medical Research Council under its HPHSR Clinician Scientist Award (HCSAINV21jun-0004) and the Ministry of Education’s Academic Research Fund Tier 1 Funding (2022-MOET1-0005).

Data Availability

The data sets generated during and analyzed during this study are available from the corresponding author on reasonable request.

Conflicts of Interest

None declared.

Consolidated criteria for reporting qualitative research (COREQ): 32-item checklist.

• Valenzuela PL, Carrera-Bastos P, Gálvez BG, Ruiz-Hurtado G, Ordovas JM, Ruilope LM, et al. Lifestyle interventions for the prevention and treatment of hypertension. Nat Rev Cardiol. 2021;18(4):251-275. [ CrossRef ] [ Medline ]
• Colberg SR, Sigal RJ, Yardley JE, Riddell MC, Dunstan DW, Dempsey PC, et al. Physical activity/exercise and diabetes: a position statement of the American Diabetes Association. Diabetes Care. 2016;39(11):2065-2079. [ FREE Full text ] [ CrossRef ] [ Medline ]
• American Diabetes Association. 4. Lifestyle management. Diabetes Care. 2016;40(Suppl 1):S33-S43. [ FREE Full text ] [ CrossRef ]
• Gant CM, Binnenmars SH, Harmelink M, Soedamah-Muthu SS, Bakker SJL, Navis G, et al. Real-life achievement of lipid-lowering treatment targets in the DIAbetes and LifEstyle Cohort Twente: systemic assessment of pharmacological and nutritional factors. Nutr Diabetes. 2018;8(1):24. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Fine LJ, Philogene GS, Gramling R, Coups EJ, Sinha S. Prevalence of multiple chronic disease risk factors. 2001 National Health Interview Survey. Am J Prev Med. 2004;27(Suppl 2):18-24. [ CrossRef ] [ Medline ]
• Linardakis M, Smpokos E, Papadaki A, Komninos ID, Tzanakis N, Philalithis A. Prevalence of multiple behavioral risk factors for chronic diseases in adults aged 50+, from eleven European countries—the SHARE study (2004). Prev Med. 2013;57(3):168-172. [ CrossRef ] [ Medline ]
• Elwyn G, Frosch D, Thomson R, Joseph-Williams N, Lloyd A, Kinnersley P, et al. Shared decision making: a model for clinical practice. J Gen Intern Med. 2012;27(10):1361-1367. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Charles C, Gafni A, Whelan T. Shared decision-making in the medical encounter: what does it mean? (or it takes at least two to tango). Soc Sci Med. 1997;44(5):681-692. [ CrossRef ] [ Medline ]
• Kunneman M, Montori VM, Castaneda-Guarderas A, Hess EP. What is shared decision making? (and what it is not). Acad Emerg Med. 2016;23(12):1320-1324. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Kashaf MS, McGill ET, Berger ZD. Shared decision-making and outcomes in type 2 diabetes: a systematic review and meta-analysis. Patient Educ Couns. 2017;100(12):2159-2171. [ CrossRef ] [ Medline ]
• Shay LA, Lafata JE. Where is the evidence? A systematic review of shared decision making and patient outcomes. Med Decis Making. 2015;35(1):114-131. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Peek ME, Gorawara-Bhat R, Quinn MT, Odoms-Young A, Wilson SC, Chin MH. Patient trust in physicians and shared decision-making among African-Americans with diabetes. Health Commun. 2013;28(6):616-623. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Politi MC, Wolin KY, Légaré F. Implementing clinical practice guidelines about health promotion and disease prevention through shared decision making. J Gen Intern Med. 2013;28(6):838-844. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Resnicow K, Catley D, Goggin K, Hawley S, Williams GC. Shared decision making in health care: theoretical perspectives for why it works and for whom. Med Decis Making. 2022;42(6):755-764. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Cupples ME, Cole JA, Hart ND, Heron N, McKinley MC, Tully MA. Shared decision-making (SHARE-D) for healthy behaviour change: a feasibility study in general practice. BJGP Open. 2018;2(2):bjgpopen18X101517. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Sheridan SL, Draeger LB, Pignone MP, Keyserling TC, Simpson RJ, Rimer B, et al. A randomized trial of an intervention to improve use and adherence to effective coronary heart disease prevention strategies. BMC Health Serv Res. 2011;11:331. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Sheridan SL, Draeger LB, Pignone MP, Rimer B, Bangdiwala SI, Cai J, et al. The effect of a decision aid intervention on decision making about coronary heart disease risk reduction: secondary analyses of a randomized trial. BMC Med Inform Decis Mak. 2014;14:14. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Moleman M, Regeer BJ, Schuitmaker-Warnaar TJ. Shared decision-making and the nuances of clinical work: concepts, barriers and opportunities for a dynamic model. J Eval Clin Pract. 2021;27(4):926-934. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Gravel K, Légaré F, Graham ID. Barriers and facilitators to implementing shared decision-making in clinical practice: a systematic review of health professionals' perceptions. Implement Sci. 2006;1:16. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Pollard S, Bansback N, Bryan S. Physician attitudes toward shared decision making: a systematic review. Patient Educ Couns. 2015;98(9):1046-1057. [ CrossRef ] [ Medline ]
• De Sutter A, De Maeseneer J, Boeckxstaens P. Empowering patients to determine their own health goals. Eur J Gen Pract. 2013;19(2):75-76. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Sieber W, Newsome A, Lillie D. Promoting self-management in diabetes: efficacy of a collaborative care approach. Fam Syst Health. 2012;30(4):322-329. [ CrossRef ] [ Medline ]
• Alders I, Smits C, Brand P, van Dulmen S. Does patient coaching make a difference in patient-physician communication during specialist consultations? a systematic review. Patient Educ Couns. 2017;100(5):882-896. [ CrossRef ] [ Medline ]
• Racey M, Jovkovic M, Alliston P, Ali MU, Sherifali D. Diabetes health coach in individuals with type 2 diabetes: a systematic review and meta analysis of quadruple aim outcomes. Front Endocrinol (Lausanne). 2022;13:1069401. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Miyamoto S, Henderson S, Fazio S, Saconi B, Thiede E, Greenwood DA, et al. Empowering diabetes self-management through technology and nurse health coaching. Diabetes Educ. 2019;45(6):586-595. [ CrossRef ] [ Medline ]
• Young RJ, Taylor J, Friede T, Hollis S, Mason JM, Lee P, et al. Pro-active Call Center Treatment Support (PACCTS) to improve glucose control in type 2 diabetes: a randomized controlled trial. Diabetes Care. 2005;28(2):278-282. [ FREE Full text ] [ CrossRef ] [ Medline ]
• McGill B, O'Hara BJ, Phongsavan P. Participant perspectives of a 6-month telephone-based lifestyle coaching program. Public Health Res Pract. 2018;28(2):27451705. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Almulhim AN, Hartley H, Norman P, Caton SJ, Doğru OC, Goyder E. Behavioural change techniques in health coaching-based interventions for type 2 diabetes: a systematic review and meta-analysis. BMC Public Health. 2023;23(1):95. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Nelson K, Pitaro M, Tzellas A, Lum A. Practice profile. Transforming the role of medical assistants in chronic disease management. Health Aff (Millwood). 2010;29(5):963-965. [ CrossRef ] [ Medline ]
• Thom DH, Wolf J, Gardner H, DeVore D, Lin M, Ma A, et al. A qualitative study of how health coaches support patients in making health-related decisions and behavioral changes. Ann Fam Med. 2016;14(6):509-516. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Djuric Z, Segar M, Orizondo C, Mann J, Faison M, Peddireddy N, et al. Delivery of health coaching by medical assistants in primary care. J Am Board Fam Med. 2017;30(3):362-370. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Dubé K, Willard-Grace R, O'Connell B, DeVore D, Prado C, Bodenheimer T, et al. Clinician perspectives on working with health coaches: a mixed methods approach. Fam Syst Health. 2015;33(3):213-221. [ CrossRef ] [ Medline ]
• Shay LA, Lafata JE. Understanding patient perceptions of shared decision making. Patient Educ Couns. 2014;96(3):295-301. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Zeuner R, Frosch DL, Kuzemchak MD, Politi MC. Physicians' perceptions of shared decision-making behaviours: a qualitative study demonstrating the continued chasm between aspirations and clinical practice. Health Expect. 2015;18(6):2465-2476. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Liddy C, Johnston S, Nash K, Ward N, Irving H. Health coaching in primary care: a feasibility model for diabetes care. BMC Fam Pract. 2014;15:60. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Tong A, Sainsbury P, Craig J. Consolidated Criteria for Reporting Qualitative Research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care. 2007;19(6):349-357. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Tan CC, Lam CSP, Matchar DB, Zee YK, Wong JEL. Singapore's health-care system: key features, challenges, and shifts. Lancet. 2021;398(10305):1091-1104. [ CrossRef ] [ Medline ]
• Kwan YH, Yoon S, Tan CS, Tai BC, Tan WB, Phang JK, et al. EMPOWERing patients with diabetes using profiling and targeted feedbacks delivered through smartphone app and wearable (EMPOWER): protocol for a randomized controlled trial on effectiveness and implementation. Front Public Health. 2022;10:805856. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol. 2006;3(2):77-101. [ CrossRef ]
• Tan JP, Cheng KKF, Siah RCJ. A systematic review and meta-analysis on the effectiveness of education on medication adherence for patients with hypertension, hyperlipidaemia and diabetes. J Adv Nurs. 2019;75(11):2478-2494. [ CrossRef ] [ Medline ]
• Ampofo AG, Khan E, Ibitoye MB. Understanding the role of educational interventions on medication adherence in hypertension: a systematic review and meta-analysis. Heart Lung. 2020;49(5):537-547. [ CrossRef ] [ Medline ]
• McCoy MA, Theeke LA. A systematic review of the relationships among psychosocial factors and coping in adults with type 2 diabetes mellitus. Int J Nurs Sci. 2019;6(4):468-477. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Suriyawong W, Kao TSAK, Robbins LB, Ling J, Malete L. Psychosocial determinants of recommended lifestyle behaviors among hypertensive patients: an integrative literature review. West J Nurs Res. 2023;45(5):455-468. [ CrossRef ] [ Medline ]
• Young-Hyman D, de Groot M, Hill-Briggs F, Gonzalez JS, Hood K, Peyrot M. Psychosocial care for people with diabetes: a position statement of the American Diabetes Association. Diabetes Care. 2016;39(12):2126-2140. [ FREE Full text ] [ CrossRef ] [ Medline ]
• McCoy MA. Examining relationships among person factors, psychosocial factors, and coping patterns in people living with type 2 diabetes mellitus in North Central Appalachia. West Virginia University. 2021. URL: https://researchrepository.wvu.edu/cgi/viewcontent.cgi?article=11189&context=etd [accessed 2024-03-08]
• Langford AT, Williams SK, Applegate M, Ogedegbe O, Braithwaite RS. Partnerships to improve shared decision making for patients with hypertension—health equity implications. Ethn Dis. 2019;29(Suppl 1):97-102. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Coylewright M, Montori V, Ting HH. Patient-centered shared decision-making: a public imperative. Am J Med. 2012;125(6):545-547. [ CrossRef ] [ Medline ]
• Wolever RQ, Dreusicke M, Fikkan J, Hawkins TV, Yeung S, Wakefield J, et al. Integrative health coaching for patients with type 2 diabetes: a randomized clinical trial. Diabetes Educ. 2010;36(4):629-639. [ CrossRef ] [ Medline ]
• Wong-Rieger D, Rieger FP. Health coaching in diabetes: empowering patients to self-manage. Can J Diabetes. 2013;37(1):41-44. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Kivelä K, Elo S, Kyngäs H, Kääriäinen M. The effects of health coaching on adult patients with chronic diseases: a systematic review. Patient Educ Couns. 2014;97(2):147-157. [ CrossRef ] [ Medline ]
• DeJesus RS, Clark MM, Rutten LJF, Hathaway JC, Wilson PM, Link SM, et al. Wellness coaching to improve lifestyle behaviors among adults with prediabetes: patients' experience and perceptions to participation. J Patient Exp. 2018;5(4):314-319. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Snyder S. Health coaching education: a conversation with pioneers in the field. Glob Adv Health Med. 2013;2(3):12-24. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Bidmead C, Cowley S. A concept analysis of partnership with clients. Community Pract. 2005;78(6):203-208. [ Medline ]
• Wiggins MS. The partnership care delivery model: an examination of the core concept and the need for a new model of care. J Nurs Manag. 2008;16(5):629-638. [ CrossRef ] [ Medline ]
• Jolles MP, Richmond J, Thomas KC. Minority patient preferences, barriers, and facilitators for shared decision-making with health care providers in the USA: a systematic review. Patient Educ Couns. 2019;102(7):1251-1262. [ CrossRef ] [ Medline ]
• Joseph-Williams N, Elwyn G, Edwards A. Knowledge is not power for patients: a systematic review and thematic synthesis of patient-reported barriers and facilitators to shared decision making. Patient Educ Couns. 2014;94(3):291-309. [ CrossRef ] [ Medline ]
• Olsen JM. Health coaching: a concept analysis. Nurs Forum. 2014;49(1):18-29. Wiley Online Library. [ CrossRef ] [ Medline ]

Abbreviations

Edited by A Mavragani; submitted 14.08.23; peer-reviewed by T Karhula, A Rahn; comments to author 06.02.24; revised version received 26.02.24; accepted 04.03.24; published 04.04.24.

©Sungwon Yoon, Chao Min Tan, Jie Kie Phang, Venice Xi Liu, Wee Boon Tan, Yu Heng Kwan, Lian Leng Low. Originally published in JMIR Formative Research (https://formative.jmir.org), 04.04.2024.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Formative Research, is properly cited. The complete bibliographic information, a link to the original publication on https://formative.jmir.org, as well as this copyright and license information must be included.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• HHS Author Manuscripts

Peer Influences on Adolescent Decision Making

Dustin albert.

Duke University

Jason Chein

Temple University

Laurence Steinberg

Research efforts to account for elevated risk behavior among adolescents have arrived at an exciting new stage. Moving beyond laboratory studies of age differences in “cool” cognitive processes related to risk perception and reasoning, new approaches have shifted focus to the influence of social and emotional factors on adolescent neurocognition. We review recent research suggesting that adolescent risk-taking propensity derives in part from a maturational gap between early adolescent remodeling of the brain's socio-emotional reward system and a gradual, prolonged strengthening of the cognitive control system. At a time when adolescents spend an increasing amount of time with their peers, research suggests that peer-related stimuli may sensitize the reward system to respond to the reward value of risky behavior. As the cognitive control system gradually matures over the course of the teenage years, adolescents grow in their capacity to coordinate affect and cognition, and to exercise self-regulation even in emotionally arousing situations. These capacities are reflected in gradual growth in the capacity to resist peer influence.

“…it seems like people accept you more if you're, like, a dangerous driver or something. If there is a line of cars going down the road and the other lane is clear and you pass eight cars at once, everybody likes that. […] If my friends are with me in the car, or if there are a lot of people in the line, I would do it, but if I'm by myself and I didn't know anybody then I wouldn't do it. That's no fun.'” Anonymous teenager, as reported in The Culture of Adolescent Risk-Taking ( Lightfoot, 1997 ; p.10)

It is well-established that adolescents are more likely than children or adults to take risks, as evinced by elevated rates of experimentation with alcohol, tobacco, and drugs, unprotected sexual activity, violent and nonviolent crime, and reckless driving ( Steinberg, 2008 ). Early research efforts to identify the distinguishing cognitive immaturity underlying adolescents' heightened risk-taking propensity bore little fruit. A litany of carefully-controlled laboratory experiments contrasted adolescent and adult capacities to perceive and process fundamental components of risk information, but found that adolescents possess the knowledge, values, and processing efficiency to evaluate risky decisions as competently as adults ( Reyna & Farley, 2006 ).

If adolescents are so risky in the real world, why do they appear so risk-averse in the lab? We propose that the answer to this question is nicely illustrated by the American teenager quoted above: “ …if I'm by myself and I didn't know anybody then I wouldn't do it. That's no fun .” If adolescents made all of their decisions involving drinking, driving, dalliances, and delinquency in the cool isolation of an experimenter's testing room, those decisions would likely appear as risk-averse as those of adults. But therein lies the rub: teenagers spend a remarkable amount of time in the company of other teenagers. This paper describes a new wave of research on the neurobehavioral substrates of adolescent decision making in peer contexts suggesting that the company of other teenagers fundamentally alters the calculus of adolescent risk taking.

Peer Influences on Adolescent Risk Behavior

Consistent with self-reports of lower resistance to peer influence among adolescents than adults ( Steinberg & Monahan, 2007 ), observational data point to the role of peer influences as a primary contextual factor contributing to adolescents' heightened tendency to make risky decisions. For instance, crime statistics indicate that adolescents typically commit delinquent acts in peer groups, whereas adults more frequently offend alone ( Zimring, 1998 ). Furthermore, one of the strongest predictors of delinquent behavior in adolescence is affiliation with delinquent peers, an association that has been attributed in varying proportions to peer socialization (e.g., “deviancy training”; Dishion, Bullock, & Granic, 2002 ) and friendship choices, wherein risk-taking adolescents naturally gravitate toward one another (e.g., Bauman & Ennett, 1996 ). Given the difficulty of distinguishing between these causal alternatives with correlational data, our lab has pursued a program of experimental research directly comparing the behavior of adolescents and adults when making decisions either alone, or in the presence of their peers.

In the first experimental study to examine age differences in the effect of peer context on risky decision making ( Gardner & Steinberg, 2005 ), early adolescents (mean age = 14), late adolescents (mean age = 19), and adults (mean age = 37) were tested on a computerized driving task, called the “Chicken Game,” which challenges the driver to advance a vehicle as far as possible on the driving course, while avoiding a crash into a wall that could appear, without warning, at any point on the course. Peer context was manipulated by randomly assigning each group of three participants to play the game either individually (alone in the room), or with two same-aged peers in the room. When tested alone, the participants from each of the three age groups engaged in a comparable amount of risk taking. In contrast, early adolescents scored twice as high on an index of risky driving when tested with their peers in the room than when alone, whereas late adolescents were approximately 50% riskier in groups, and adults showed no difference in risky driving related to social context. The ongoing goal of our research program is to further specify the behavioral and neural mechanisms of this peer effect on adolescent risk taking.

A Neurodevelopmental Model of Peer Influences on Adolescent Decision Making

Building on extensive evidence demonstrating maturational changes in brain structure and function occurring across the second decade of life (and frequently beyond), we have advanced a neurodevelopmental account of heightened susceptibility to peer influence among adolescents ( Steinberg, 2008 ; Albert & Steinberg, 2011 ). In brief, we propose that, among adolescents more than adults, the presence of peers “primes” a reward-sensitive motivational state that increases the subjective value of immediately available rewards and thereby increases the probability that adolescents will favor the short-term benefits of a risky choice over the long-term value of a safe alternative. Although a comprehensive presentation of the behavioral and neuroscientific evidence underlying this hypothesis is beyond our current scope (but see Albert & Steinberg, 2011 ), a brief review of three fundamental assumptions of this model will set the stage for a description of our peer influence studies.

First, decisions are a product of both cognitive and affective input, even when affect is unrelated to the choices under evaluation

Research based on adult populations has identified several pathways by which affect influences decision making ( Loewenstein, Weber, Hsee, & Welch, 2001 ). For instance, the anticipated emotional outcome of a behavioral option -- how one expects to feel after making a given choice --contributes to one's cognitive assessment of its expected value. Indeed, affective states may influence decision processing even when the source of the affect is not directly related to the choices under evaluation. Such incidental affective influences are apparent in experiments demonstrating the effect of pre-existing or experimentally elicited affective states on adult perception, memory, judgment, and behavior ( Winkielman, Knutson, Paulus, & Trujillo, 2007 ). One experiment illustrating this effect found that elicitation of incidental positive emotion (via presentation of masked happy faces) caused participants to pour and drink more of a novel beverage than participants who had viewed angry faces, despite no differences in self-reported emotion between the two groups ( Winkielman, Berridge, & Wilbarger, 2005 ). Consistent with evidence for extensive overlap in the neural circuitries implicated in the evaluation of socio-emotional and choice-related incentive cues (e.g., ventral striatum, ventromedial prefrontal cortex; for a recent review, see Falk, Way, & Jasinska, 2012 ), Winkielman and colleagues describe this priming effect as an instance of approach sensitization . That is, neural responses to positively valenced socio-emotional stimuli – in this case, responses not even reaching the level of conscious awareness – may sensitize approach (i.e., “GO”) responding to unrelated incentive cues. As we describe below, several characteristics of adolescent neurobehavioral functioning suggest that this approach sensitization effect could be a particularly powerful influence on adolescent decision making in peer contexts.

Second, adolescents exhibit stronger “bottom-up” affective reactivity than adults in response to socially relevant stimuli

Whereas some controversy remains regarding the degree to which adolescents are more or less sensitive than children and adults to non-social reward cues ( Galvan, 2010 ; Spear, 2009 ), few scholars now dispute that adolescence is a period of peak neurobehavioral sensitivity to social stimuli ( Burnett, Sebastian, Kadosh, & Blakemore, 2011 ; Somerville, this issue ). Puberty-related increases in gonadal hormones have been linked to a proliferation of receptors for oxytocin within subcortical and limbic circuits, including the amygdala and striatum ( Spear, 2009 ). Oxytocin neurotransmission has been implicated in a variety of social behaviors, including facilitation of social bonding and recognition and memory for positive social stimuli (Insel & Fernald, 2004). Alongside concurrent changes in dopaminergic function within neural circuits broadly implicated in incentive processing ( Spear, 2009 ), these puberty-related increases in gonadal hormones and oxytocin receptor density contribute to changes in a constellation of social behaviors observed in adolescence.

Peer relations are never more salient than in adolescence. In addition to a puberty-related spike in interest in opposite-sex relationships, adolescents spend more time than children or adults interacting with peers, report the highest degree of happiness when in peer contexts, and assign greatest priority to peer norms for behavior ( Brown & Larson, 2009 ). This developmental peak in affiliation motivation appears highly conserved across species: Adolescent rats also spend more time than younger or older rats interacting with peers, while showing evidence that such interactions are highly rewarding ( Doremus-Fitzwater, Varlinskaya, & Spear, 2010 ). Moreover, several developmental neuroimaging studies indicate that, relative to children and adults, adolescents show heightened neural activation in response to a variety of social stimuli, such as facial expressions and social feedback ( Burnett et al., 2011 ). For instance, one of the first longitudinal neuroimaging studies of early adolescence demonstrated a significant increase from ages 10 to 13 in ventral striatal and ventral prefrontal reactivity to facial stimuli ( Pfeifer et al., 2011 ). Together, this evidence for hypersensitivity to social stimuli suggests that adolescents may be more likely than adults to generate a baseline state of heightened approach motivation when exposed to positively valenced peer stimuli in a decision-making scenario, thus setting the stage for an exaggerated approach sensitization effect of peer context on risky decision making.

Third, adolescents are less capable than adults of “top-down” cognitive control of impulsive behavior

In contrast to the relatively sudden changes in social processing that occur around the time of puberty, cognitive capacities supporting efficient self-regulation mature in a gradual, linear pattern over the course of adolescence. In developmental parallel with structural brain changes thought to support neural processing efficiency (e.g., increased axonal myelination), adolescents show continued gains in response inhibition, planned problem solving, flexible rule use, impulse control, and future orientation ( Steinberg, 2008 ). Indeed, evidence is growing for a direct link between structural and functional brain maturation during adolescence and concurrent improvements in cognitive control. In addition to studies correlating white matter maturation with age-related cognitive improvements ( Schmithorst & Yuan, 2010 ), developmental neuroimaging studies utilizing tasks requiring response inhibition (e.g., Go-No/Go, Stroop, flanker tasks, ocular antisaccade) describe relatively inefficient recruitment by adolescents of the core neural circuitry supporting cognitive control (e.g., lateral prefrontal and anterior cingulate cortex) ( Luna, Padmanabhan, & O'Hearn, 2010 ). Moreover, research on age differences in control-related network dynamics demonstrate adolescent immaturity in the functional integration of neural signals deriving from specialized cortical and subcortical “hub” regions ( Stevens, 2009 ). This immature capacity for functional integration may contribute to adolescent difficulties in simultaneously evaluating social, affective, and cognitive factors relevant to a given decision, particularly when social and emotional considerations are disproportionately salient.

Identification of Mechanisms Underlying Peer Influences on Adolescent Decision Making

In an effort to further specify the neurodevelopmental vulnerability underlying adolescent susceptibility to peer influence, we have conducted a series of behavioral and neuroimaging experiments comparing adolescent and adult decision making in variable social contexts. Specifically, we have sought to determine whether the presence of peers biases adolescent decision making by (a) modulating responses to incentive cues, consistent with the approach sensitization hypothesis, (b) disrupting inhibitory control, or (c) altering both of these processes. As a first step in addressing this question, we conducted an experiment that randomly assigned late adolescents (ages 18 and 19) to complete a series of tasks either alone or in the presence of two same-age, same-sex peers. Risk-taking propensity was assessed using the “Stoplight” game, a first-person driving game wherein participants must advance through a series of intersections to reach a finish line as quickly as possible to receive a monetary reward ( Figure 1 ). Each intersection is marked by a stoplight that turns yellow (and sometimes red) as the car approaches, and participants must decide to either “hit the brakes” (and lose time while waiting for the light to turn green) or run the light (and risk crashing while crossing through an intersection). We also administered tasks for cognitive control (Go/No-Go) and preference for immediate-over-delayed rewards (Delay Discounting). Whereas no group differences were evident on the Go/No-Go index of inhibitory control, adolescents in the peer presence condition took more risks on the Stoplight task ( Albert et al., 2009 ) and indicated stronger preference for immediate-over-delayed rewards ( O'Brien, Albert, Chein, & Steinberg, 2011 ), relative to adolescents who completed the tasks alone.

In the Stoplight driving game, participants are instructed to attempt to reach the end of a straight track as quickly as possible. At each of 20 intersections, participants render a decision to either stop the vehicle (STOP) or to take a risk and run the traffic light (GO). Stops result in a short delay. Successful risk taking results in no delay. Unsuccessful risk taking results in a crash, and a relatively long delay. Summary indices of risk taking include (a) the proportion of intersections in which the participant decides to run the light, and (b) the total number of crashes.

Findings from a recent follow-up experiment suggest that peer observation influences adolescents' decision making even when the peer is anonymous and not physically present in the same room. Utilizing a counterbalanced, repeated-measures design, we assessed the task performance of late adolescents once in a standard “alone” condition, and once in a deception condition that elicited the impression that their task performance was being observed by a same-age peer in an adjoining room. As predicted, participants exhibited stronger preference for immediate rewards on a delay discounting task when they believed they were being observed, relative to alone ( Weigard, Chein, & Steinberg, 2011 ). Peer observation also resulted in a higher rate of monetary gambles on a probabilistic gambling task, but only for participants with relatively lower self-reported resistance to peer influence ( Smith, Chein, & Steinberg, 2011 ). Along similar lines, Segalowitz et al. (2012) report that individuals high in self-reported sensation seeking are especially susceptible to the peer effect on risk taking. Considered together, these behavioral results suggest that peer presence increases adolescents' risk taking by increasing the salience (or subjective value) of immediately available rewards, and that some adolescents are more susceptible to this effect than others.

Our recent work has utilized brain imaging to more directly examine the neural dynamics underlying adolescent susceptibility to peer influences. In the first of these studies, we scanned adolescents and adults while they played the Stoplight game, again utilizing a counterbalanced within-subjects design ( Chein, Albert, O'Brien, Uckert, & Steinberg, 2011 ). All subjects played the game in the scanner twice -- once in a standard “alone” condition, and once in a “peer” condition, wherein they were made aware that their performance was being observed on a monitor in a nearby room by two same-age, same-sex peers who had accompanied them to the experiment. As predicted, adolescents but not adults took significantly more risks when observed by peers than when alone ( Figure 2 ). Furthermore, analysis of neural activity during the decision-making epoch showed greater activation of brain structures implicated in reward valuation (ventral striatum and orbitofrontal cortex) for adolescents in the peer relative to the alone scans, an effect that was not apparent for adults ( Figure 3 ). Indeed, the degree to which participants (across all ages) evinced peer-greater-than- alone activation in the ventral striatum was inversely correlated with self-reported resistance to peer influence ( Figure 4 ). This study represents the first evidence that peer presence accentuates risky decision making in adolescence by modulating activity in the brain's reward valuation system.

Mean (a) percentage of risky decisions and (b) number of crashes for adolescent, young adult, and adult participants when playing the Stoplight driving game either alone or with a peer audience. Error bars indicate the standard error of the mean.

(a) Brain regions exhibiting an age by social condition interaction included the right ventral striatum (VS, MNI peak coordinates: x = 9, y = 12, z = -8) and left orgitofrontal cortex (OFC, MNI peak coordinates: x = -22, y = 47, z = -10). (b) Mean estimated BOLD signal change (beta coefficient) from the four peak voxels of the VS and the OFC in adolescents (adols.), young adults (YA), and adults under ALONE and PEER conditions. Error bars indicate standard errors of the mean. Brain images are shown by radiological convention (left on right), and thresholded at p < .01 for presentation purposes.

Resistance to Peer Influence correlated with Stoplight-related activity in the right ventral striatum (VS). Estimated activity was extracted from an average of the four peak voxels in the VS region of interest. Scatterplot of activity in the VS indicating an inverse linear correlation between self-reported resistance to peer influence (RPI) and the neural peer effect (i.e., the difference in average VS activity in peer relative to alone conditions, or βpeer − βalone).

Conclusions and Future Directions

Although our work to date has indicated that the effect of peers on adolescent risk taking is mediated by changes in reward processing, we recognize that the distinction between risk taking that is attributable to heightened arousal of the brain's reward system versus that which is due to immaturity of the cognitive control system is somewhat artificial, since these brain systems influence each other in a dynamic fashion. Consistent with this notion, in a comparison of children, adolescents, and adults on a task that requires participants to either produce or inhibit a motor response to pictures of calm or happy faces, Somerville, Hare, and Casey (2011) not only found elevated ventral striatal activity for adolescents in response to happy faces, which the authors discuss as an “appetitive” cue, but also a corresponding increase in failures to inhibit motor responses to the happy relative to the calm facial stimuli. Thus, adolescents' exaggerated response to positively-valenced social cues is shown here to directly undermine their capacity to inhibit approach behavior. Translated to the peer context, this finding suggests that adolescents may not only be particularly sensitive to the reward-sensitizing effects of social stimuli, but that this sensitization may further undermine their capacity to “put the brakes on” impulsive responding.

Despite the promise of this conceptual model, further work is needed to specify the neurodevelopmental dynamics underlying adolescent susceptibility to peer influence, and to translate this understanding to the design of effective prevention programs. In an effort to “decompose” the peer effect, we are currently examining age differences in the influence of social cues on neural activity underlying performance on tasks specifically tapping reward processing and response inhibition. In addition, we are investigating whether conditions known to diminish cognitive control (e.g., alcohol intoxication) might exacerbate the influence of peer context on risky decision making. Finally, as a first step toward our ultimate goal of utilizing this research to improve the efficacy of risk-taking prevention programs, we are examining whether targeted training designed to promote earlier maturation of cognitive control skills might attenuate the influence of peers on adolescent decision making.

Contributor Information

Dustin Albert, Duke University.

Jason Chein, Temple University.

Laurence Steinberg, Temple University.

Recommended Readings

• Albert D, Steinberg L. A comprehensive presentation of the neurodevelopmental model of peer influences on adolescent decision making 2011 See References. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Burnett S, Sebastian C, Kadosh K, Blakemore An up-to date review of the social neuroscience of adolescence 2011 See References. [ Google Scholar ]
• Chein J, Albert D, O'Brien L, Uckert K, Steinberg L. An empirical report of peer influences on adolescent risk taking and neural activity 2011 See References. [ Google Scholar ]
• Falk EG, Way BM, Jasinska AJ. A recent review highlighting promising new directions for neuroscientific research on social influence across the lifespan 2012 See References. [ Google Scholar ]
• Spear LP. A thorough and accessible textbook survey of neuroscientific research on adolescent development 2009 See References. [ Google Scholar ]
• Albert D, O'Brien L, DiSorbo A, Uckert K, Egan DE, Chein J, Steinberg L. Peer influences on risk taking in young adulthood; Poster session presented at the biennial meeting of the Society for Research in Child Development; Denver, CO. 2009. Apr, [ Google Scholar ]
• Albert D, Steinberg L. Peer influences on adolescent risk behavior. In: Bardo MT, Fishbein DH, Milich &R, editors. Inhibitory Control and Drug Abuse Prevention: From Research to Translation. New York: Springer; 2011. [ Google Scholar ]
• Bauman KE, Ennett SE. On the importance of peer influence for adolescent drug use: Commonly neglected considerations. Addiction. 1996; 91 :185–198. [ PubMed ] [ Google Scholar ]
• Brown BB, Larson J. Handbook of adolescent psychology, Vol 2: Contextual influences on adolescent development. 3rd. Hoboken, NJ US: John Wiley & Sons Inc; 2009. Peer relationships in adolescents; pp. 74–103. [ Google Scholar ]
• Burnett S, Sebastian C, Kadosh K, Blakemore The social brain in adolescence: Evidence from functional magnetic resonance imaging and behavioural studies. Neuroscience and Biobehavioral Reviews. 2011; 35 :1654–1664. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Chein J, Albert D, O'Brien L, Uckert K, Steinberg L. Peers increase adolescent risk taking by enhancing activity in the brain's reward circuitry. Developmental Science. 2011; 14 :F1–F10. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Dishion TJ, Bullock BM, Granic I. Pragmatism in modeling peer influence: Dynamics, outcomes, and change processes. Development and Psychopathology. 2002; 14 (4):969–981. [ PubMed ] [ Google Scholar ]
• Doremus-Fitzwater TL, Varlinskaya EI, Spear LP. Motivational systems in adolescence: possible implications for age differences in substance abuse and other risk-taking behaviors. Brain and cognition. 2010; 72 :114–23. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Falk EG, Way BM, Jasinska AJ. An imaging genetics approach to understanding social influence. Frontiers in Human Neuroscience. 2012; 6 :1–13. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Galvan A. Adolescent development of the reward system. Frontiers in Human Neuroscience. 2010; 4 :1–9. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Gardner M, Steinberg L. Peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: An experimental study. Developmental Psychology. 2005; 41 :625–635. [ PubMed ] [ Google Scholar ]
• Lightfoot C. The culture of adolescent risk-taking. New York: Guilford Press; 1997. [ Google Scholar ]
• Loewenstein G, Weber EU, Hsee CK, Welch N. Risk as feelings. Psychological Bulletin. 2001; 127 :267–286. [ PubMed ] [ Google Scholar ]
• Luna B, Padmanabhan A, O'Hearn K. What has fMRI told us about the development of cognitive control through adolescence? Brain and Cognition. 2010; 72 :101–113. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• O'Brien L, Albert D, Chein J, Steinberg L. Adolescents prefer more immediate rewards when in the presence of their peers. Journal of Research on Adolescence. 2011; 21 :747–753. [ Google Scholar ]
• Pfeifer JH, Masten CL, Moore WE, Oswald TM, Iacoboni M, Mazziotta JC, Dapretto M. Entering adolescence: Resistance to peer influence, risky behavior, and neural changes in emotional reactivity. Neuron. 2011; 69 :1029–1036. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Reyna VF, Farley F. Risk and rationality in adolescent decision making - Implications for theory, practice, and public policy. Psychological Science in the Public Interest. 2006; 7 :1–44. [ PubMed ] [ Google Scholar ]
• Schmithorst VJ, Yuan WH. White matter development during adolescence as shown by diffusion MRI. Brain and Cognition. 2010; 72 :16–25. [ PubMed ] [ Google Scholar ]
• Segalowitz SJ, Santesso DL, Willoughby T, Reker DL, Campbell K, Chalmers H, Rose-Krasnor L. Adolescent peer interaction and trait surgency weaken medial prefrontal cortex responses to failure. Social, Cognitive, and Affective Neuroscience. 2012; 7 :115–124. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Smith A, Chein J, Steinberg L. Developmental differences in reward processing in the presence of peers; Paper presented at the annual meeting of the Society for Neuroscience; Washington. 2011. Nov, [ Google Scholar ]
• Somerville LH. this issue [ Google Scholar ]
• Somerville LH, Hare TA, Casey BJ. Frontostriatal maturation predicts cognitive control failure to appetitive cues in adolescents. Journal of Cognitive Neuroscience. 2011; 23 :2123–2134. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Spear LP. The behavioral neuroscience of adolescence. New York: W. W. Norton; 2009. [ Google Scholar ]
• Steinberg L. A social neuroscience perspective on adolescent risk-taking. Developmental Review. 2008; 28 :78–106. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Steinberg L, Monahan KC. Age differences in resistance to peer influence. Developmental Psychology. 2007; 43 :1531–1543. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Stevens MC. The developmental cognitive neuroscience of functional connectivity. Brain and Cognition. 2009; 70 :1–12. [ PubMed ] [ Google Scholar ]
• Weigard A, Chein J, Steinberg L. Influence of anonymous peers on risk-taking behavior in adolescents; Paper presented at the Eleventh Annual Stanford Undergraduate Psychology Conference; Stanford, CA. 2011. May, [ Google Scholar ]
• Winkielman P, Berridge KC, Wilbarger JL. Unconscious affective reactions to masked happy versus angry faces influence consumption behavior and judgments of value. Personality and Social Psychology Bulletin. 2005; 31 :121–135. [ PubMed ] [ Google Scholar ]
• Winkielman P, Knutson B, Paulus MP, Trujillo JL. Affective influence on judgments and decisions: Moving towards core mechanisms. Review of General Psychology. 2007; 11 :179–192. [ Google Scholar ]
• Zimring F. American Youth Violence Studies in Crime and Public Policy. New York: Oxford: Oxford University Press; 1998. [ Google Scholar ]

Search Cornell

Class Roster

Section menu.

• Toggle Navigation
• Summer 2024
• Spring 2024
• Winter 2024
• Archived Rosters

Last Updated

• Schedule of Classes - April 9, 2024 7:33PM EDT
• Course Catalog - April 9, 2024 7:07PM EDT

COGST 4250 Translational Research on Decision Making

Course description.

Course information provided by the Courses of Study 2023-2024 . Courses of Study 2024-2025 is scheduled to publish mid-June.

Introductory laboratory-based course focusing on basic foundations in translational research on decision making across the lifespan. The course introduces students to hands-on applications of research skills in the context of research on decision making, spanning basic and applied research in law, medicine, behavioral economics, and policy. It focuses on such topics as human subjects protection, working with populations across the lifespan (e.g., children, seniors), database development, working with external partners and stakeholders (e.g., schools, hospitals), and basic concepts and techniques in decision research. Students participate in weekly laboratory meetings in small teams focused on specific projects as well as monthly meetings in which all teams participate. During laboratory meetings, students discuss ongoing research, plans for new studies, and interpretations of empirical findings from studies that are in progress or have been recently completed. New students work closely with experienced students and eventually work more independently. In order to fully grasp how the research projects fit into the broader field, students read relevant papers weekly and write reaction responses. Because several projects are ongoing at all times, students have the opportunity to be involved in more than one study and are assigned multiple tasks such as piloting research paradigms, subject recruitment, data collection, data analysis, and data entry. Students attend a weekly lab meeting for 1.5 hours per week, read pertinent papers, write reaction responses, and work 10.5 hours per week in the laboratory completing tasks that contribute to ongoing research studies.

When Offered Fall.

Prerequisites/Corequisites Prerequisite: HD 1150 or HD 1170 or PSYCH 1101 also HD 2830 and HD 4750 and HD 4760.

Distribution Category (SCD-AS)

• Be able to know and evaluate evidence-based hypotheses.

View Enrollment Information

  Regular Academic Session.   Combined with: HD 4250

Credits and Grading Basis

4 Credits Stdnt Opt (Letter or S/U grades)

Class Number & Section Details

 5509 COGST 4250   LAB 401

Meeting Pattern

• M 2:00pm - 4:30pm To Be Assigned
• Aug 26 - Dec 9, 2024

Instructors

To be determined. There are currently no textbooks/materials listed, or no textbooks/materials required, for this section. Additional information may be found on the syllabus provided by your professor.

For the most current information about textbooks, including the timing and options for purchase, see the Cornell Store .

Additional Information

Instruction Mode: In Person

Instructor Consent Required (Add)

Or send this URL:

Available Syllabi

About the class roster.

The schedule of classes is maintained by the Office of the University Registrar . Current and future academic terms are updated daily . Additional detail on Cornell University's diverse academic programs and resources can be found in the Courses of Study . Visit The Cornell Store for textbook information .

Please contact [email protected] with questions or feedback.

If you have a disability and are having trouble accessing information on this website or need materials in an alternate format, contact [email protected] for assistance.

Cornell University ©2024