Faculty Research - Information Analytics Research Publications

Why do organizational decision-makers use social media information to assess hireability?: An empirical study of stated reasons
Journal of Management Scientific Reports
Jason Thatcher
Additional authors: Philip Bobko, Philip L. Roth, Rebecca A. Roth, Michael McDaniel
Nov. 2025, Vol. 3 Issue 3/4, p324-354

Abstract: Social media (SM) assessments are a new type of selection procedure. Their frequency of use in hireability decision-making is substantial, encompassing both utilitarian (e.g., LinkedIn) and hedonic (e.g., Facebook) platforms. Yet, there is limited evidence of SM assessment validity (prediction of job performance). This mismatch (use vs. lack of validity evidence) motivated our study of why decision-makers embrace such information. We used (i) the unified technology acceptance model and (ii) organizational justice theory to list positively framed facets of SM assessments (e.g., provide information about applicant skills, used consistently, fun to use). Decision-makers were assigned to one of three conditions—information retrieved from Facebook, LinkedIn, or structured interviews (as a baseline)—and were asked to rate how well these explanatory facets characterized information from those SM (and comparison) sources. Empirical composites (principal components) of these facets were developed. We found that all composites were positively related to use of SM across all conditions. These findings are disconcerting given the equivocal validity of SM assessments (and potential for much extraneous, non-job-related information). We urge continued research on reasons for frequent use of SM assessments, as well as implications of that use. Such research for other predictors is also encouraged.

An explainable framework for assisting the detection of AI-generated textual content
Decision Support Systems
David Dobolyi
Additional authors: Sen Yan, Zhiyi Wang
Sep. 2025, Vol. 196

Abstract: The recent development of generative AI (GenAI) algorithms has allowed machines to create new content in a realistic way, driving the spread of AI-generated content (AIGC) on the Internet. However, generative AI models and AIGC have exacerbated several societal challenges such as security threats (e.g., misinformation), trust issues, ethical concerns, and intellectual property regulation, calling for effective detection methods and a better understanding of AI-generated vs. human-written content. In this paper, we focus on AI-generated texts produced by large language models (LLMs) and extend prior detection methods by proposing a novel framework that combines semantic information and linguistic features. Based on potential semantic and linguistic differences in AI vs. human writing, we design our Semantic-Linguistic-Detector (SemLinDetector) framework by integrating a transformer-based semantic encoder and a linguistic encoder with parallel linguistic representations. By comparing a series of benchmark models on datasets collected from various LLMs and human writers in multiple domains, our experiments show that the proposed detection framework outperforms other benchmarks in a consistent and robust manner. Moreover, our model interpretability analysis showcases our framework's potential to help understand the reasoning behind prediction outcomes and identify patterns of differences in AI-generated and human-written content. Our research adds to the growing space of GenAI by proposing an effective and responsible detection system to address the risks and challenges of GenAI, offering implications for researchers and practitioners to better understand and regulate AIGC. • Proposes SemLinDetector, an explainable framework integrating both semantic and linguistic features to detect AI-generated texts. • Conducts a thorough benchmarking analysis, which shows SemLinDetector outperforms nine benchmark models across both in-sample and out-of-sample datasets, including essays, news, and mental health texts. • Employs XAI methods to interpret model decisions. • Investigates patterns for differentiating between AI-generated and human written texts in terms of vocabulary, structure, linguistic styles, etc. • Provides insights for regulators, educators, and others regarding AI-generated text detection.

Does Ransomware Make Investors "Wannacry"? On Investors' Divergent Reactions to Ransomware Hits and Near Misses
MIS Quarterly
Sebastian Schuetz
Additional authors: Yan Chen, Jens Forderer, Yusi Ma
Sep. 2025, Vol. 49 Issue 3, p1153-1168

Abstract: In recent years, ransomware has become one of the most dangerous cyber threats, with successful attacks causing severe operational disruptions and staggering damages. Rationally speaking, investors should react negatively to firms' ransomware disclosures, but this may not always be the case. Based on norm theory, we describe a paradoxical phenomenon wherein investors exhibit negative reactions to ransomware hits (i.e., events that led to operational disruptions) but positive reactions to near misses (i.e., events in which operational disruptions were narrowly avoided). The positive reactions occur due to an outcome bias in which near-miss events—events that are objectively negative but less severe than expected—are viewed positively instead of negatively. We tested these predictions by reporting on an investigation of stock market reactions to disclosures of ransomware hits vs. near misses. To do so, we assembled a comprehensive dataset of ransomware incidents disclosed by U.S. public firms. Using the event study method, we estimated abnormal stock market returns and found evidence in support of our predictions. First, in line with expectations, ransomware hits that led to the expected severe impact resulted in stock price drops of -4.40%. However, near misses, where disruptions were avoided, were rewarded with gains of 2.87%, confirming positive instead of negative reactions. This offers new insights into investors' biased responses to certain cybersecurity incidents. These positive reactions, however, represent a call for caution because, albeit seemingly favorable, they mask underlying risks.

Mapping Entrepreneurial Collaborative Economy Landscape: A Systematic Literature Review with Textometric Analysis
Economies
Ramiro Montealegre
Additional authors: Salvador Bueno, Eva M. Gallego, M. Dolores Gallego
Aug. 2025, Vol. 13 Issue 8, p246

Abstract: The collaborative economy is experiencing a remarkable surge, offering vast potential for growth. Consequently, this burgeoning movement has become a focal point of interest in the realm of entrepreneurship. However, numerous unexplored or inadequately addressed research gaps persist, leaving us without a well-defined paradigm for what we can term the entrepreneurial collaborative economy. In light of these challenges, this study embarks on a quest to bridge these gaps through a comprehensive systematic literature review. Two research objectives guided our endeavor: (1) mapping the literature related to the collaborative economy in the field of entrepreneurship to propose a research taxonomy, and (2) analyzing areas in this field that warrant further research. Our literature review, conducted using the PRISMA methodology, yielded 407 studies. Employing advanced textometric techniques, we uncovered a research taxonomy consisting of three distinct clusters within entrepreneurial collaborative economy studies. In particular, our investigation has unveiled that the entrepreneurial collaborative economy paradigm remains in a state of emergence within the academic literature. The paper concludes with thought-provoking discussions and key insights.

Generative AI and Empirical Research Methods in Operations Management
Journal of Operations Management
Jason Thatcher
Additional authors: Timofey Shalpegin, Tyson R. Browning, Ajay Kumar, Guangzhi Shang, Jan C. Fransoo, Matthias Holweg, Benn Lawson
Jul. 2025, Vol. 71 Issue 5, p578-587

Abstract: The article delves into the integration of Generative Artificial Intelligence (Gen-AI) in academic research, particularly in Operations Management (OM), discussing its potential benefits and challenges. It addresses ethical concerns like privacy, bias, and the importance of human judgment in decision-making processes. Experts highlight the risks of epistemic, methodological, and systemic failures that could impact the reliability of academic knowledge, emphasizing the necessity for transparency and accountability in the responsible use of Gen-AI. The document offers a comprehensive overview of recent AI research, covering applications in research methods, biomedicine, social science, and policy implications, while stressing the importance of research integrity and ethical publishing practices.

AI assistance improves people’s ability to distinguish correct from incorrect eyewitness lineup identifications
Proceedings of National Academy of Sciences (PNAS)
David Dobolyi
Additional authors: L.E. Kelso, J.H. Grabman
May 2025, 122 (21), e2503971122

Abstract: Mistaken eyewitness identification is one of the leading causes of false convictions. Improving law enforcement’s ability to identify correct identifications could have profound implications for criminal justice. Across two experiments, we show that AI-assistance can improve people’s ability to distinguish between accurate and inaccurate eyewitness lineup identifications. Participants (Experiment 1: N = 1,092, Experiment 2: N = 1,809) saw an eyewitness’s lineup identification, accompanied by the eyewitness’s verbal confidence statement (e.g., “I’m pretty sure”) and either a featural (“I remember his eyes”), recognition (“I remember him”), or familiarity (“He looks familiar”) justification. They then judged the accuracy of the eyewitness’s identification. AI-assistance (vs. no assistance) improved people’s ability to distinguish between correct identifications and misidentifications, but only when they evaluated lineup identifications based on recognition or featural justifications. Discrimination of identifications based on familiarity justifications showed little improvement with AI-assistance. This project is a critical step in evaluating human-algorithm interactions before widespread use of AI-assistance by law enforcement.

Self-organization and governance in digital platform ecosystems: An information ecology approach
MIS Quarterly
Jason Thatcher
Additional authors: Martin Engert, Andreas Hein, Likoebe M. Maruping, Helmut Krcmar
Mar. 2025, Vol. 49 Issue 1, p91-122

Abstract: This research investigates the interplay of top-down control and bottom-up self-organization within digital platform ecosystems (DPEs), focusing on the formation and management of complementor coalitions. Although these coalitions can increase a DPE's generativity, they can also threaten its integrity. We investigate this tension by employing information ecology (IE) theory, which allows us to examine complementor coalitions as holons that navigate between self-assertiveness and integration within the structural hierarchies of DPEs. Utilizing an inductive, embedded case-study approach, we analyze the interplay between top-down control exerted by platform owners and the bottom-up selforganization of complementors in two enterprise software platform ecosystems. Our findings identify three distinct interaction modes—mandated, supported, and autonomous self-organization—each presenting hierarchical trade-offs between platform owner control and complementor autonomy. Our findings extend the prevalent owner-centric theory of platform governance by highlighting the significant impact of bottom-up self-organization on the governance and evolution of DPEs. We propose an integrated theory that accommodates these new dynamics, suggesting soft power as an effective governance mechanism. This study contributes to a deeper understanding of the complexities in governing DPEs and offers practical insights for managing top-down control and bottom-up self-organization in the evolving landscape of enterprise software DPEs.

A Comparison Between Numeric Confidence Ratings and Verbal Confidence Statements
Journal of Experimental Pyschology
David Dobolyi
Additional authors: Travis M. Seale-Carlisle, Jesse H. Grabman, Chad S. Dodson
Mar. 2025, Vol. 31 Issue 1, p12-39

Abstract: Is confidence most diagnostic of accuracy when expressed in numbers or when expressed in words? This question bears immense importance in many real-world contexts especially within the confines of eyewitness identification. In an eyewitness identification task, we compared the diagnostic value of numeric confidence across rating scales that varied in grain size (3-point vs. 6-point vs. 21-point vs. 101-point rating scales). We also compared the diagnostic value of numeric confidence to verbal confidence statements using several machine-learning algorithms. We found that fine-grain ratings are more diagnostic of identification accuracy than coarse-grain ratings, which suggests that the former provides a closer correspondence to memory strength than the latter. Moreover, we found that verbal confidence statements capture diagnostic information about the likely accuracy of an identification that numeric confidence ratings do not capture. This suggests that verbal confidence statements and numeric confidence ratings reflect partially independent, nonoverlapping sources of information. These results shed light on the processes that provide diagnostic value to confidence. From an applied standpoint, these results suggest that verbal confidence statements and numeric confidence ratings ought to be collected from eyewitnesses after an identification decision. Collecting both captures more diagnostic information than either can capture in isolation. Public Significance Statement: People routinely express confidence in their decisions. High-confidence decisions usually yield high accuracy and much higher accuracy than low-confidence decisions. Yet, we know very little about the best method for collecting confidence. Is confidence most indicative of accuracy when expressed in numbers or when expressed in words? This is an important question that is especially relevant to investigators who must collect eyewitness identification evidence. In this article, we describe an eyewitness identification experiment that sought to answer this question. We compared coarse-grain and fine-grain numeric confidence ratings. We also compared numeric confidence ratings to verbal confidence statements. We found that confidence is most indicative of accuracy when both numeric confidence ratings and verbal confidence statements are considered. These results shed new light on the underlying processes that give rise to confidence and suggest that investigators collect eyewitness confidence in numbers and words following an identification decision.

Consumers' Opinion Orientations and Their Credit Risk: An Econometric Analysis Enhanced by Multimodal Analytics
Journal of the Association for Information Systems
Jason Thatcher
Additional authors: Qiping Wang, Raymond Yiu Keung Lau, Wai Ting Eric Ngai, Wei Xu
2024, Vol. 25 Issue 4, p1117-1156.

Abstract: The rise of financial technology (fintech) has motivated practitioners and researchers to explore alternative data sources and enhanced credit scoring methods for better assessment of consumers' credit risk. In this study, we examine whether deep-level diversity derived from consumers' multimodal social media posts (i.e., alternative data) can enhance credit risk assessment or not. First, we propose novel lifestyle-based risk constructs (e.g., opinion risk) to capture consumers' deep-level diversity. Second, we incorporate these lifestyle-based risk constructs into econometric models to empirically evaluate the relationship between consumers' deep-level diversity and their credit risk. Using a credit scoring dataset provided by a fintech firm listed on Nasdaq, our econometric analysis reveals that consumers' opinion risk constructs extracted from their multimodal social media posts are positively associated with their credit risk. Furthermore, our results show that the proposed opinion risk constructs can significantly improve the effectiveness of predicting consumers' credit risk. Interestingly, our empirical results also show that combining the opinion risk constructs derived from images and text can significantly improve the effectiveness in credit risk prediction. This work contributes to the fintech domain by proposing novel lifestyle-based risk constructs for decision support in the credit scoring context.

The Impact of Social Comparison on Turnover Among Information Technology Professionals
Journal of Management Information Systems
Jason Thatcher
Additional authors: Manuel Wiesche, Christoph Pflügler
2024, Vol. 41 Issue 1, p297-324.

Abstract: While IT workforce research often examines job- and organizational-related reasons for turnover, studies rarely connect the immediate social context of IT work to IT professionals' workplace behavior. To understand the influence of the immediate social context, we develop a social comparison model that connects the social influence of co-worker who have left an organization to turnover among IT professionals who remain. We test our model using a social network of 4,011 IT professionals employed in a large IT firm over four years. We complement our analysis of this data with a multiple-case study with five software development teams. Across the two studies, our results suggest that the departure of an IT professional increases the probability of turnover among remaining coworkers; further, we found that turnover is even more likely when the remaining IT professionals are similar in technical abilities and demographic attributes to the co-worker who left. Our results direct attention to the immediate social context as an influence on the turnover behavior of IT professionals and explain how similarity in domain-specific attributes shapes this turnover behavior. Practitioners should know that a single departure may cause a chain reaction in IT work teams and organizations and find suggestions for assigning new employees.

Crowdlending Behaviors in the Aftermath of a Crisis: Evidence From a Natural Experiment
Production & Operations Management
Zhiyi Wang
Additional authors: Lusi Yang, Varun Karamshetty, Jungpil Hahn
Jan. 2024, Vol. 33 Issue 1, p243-263.

Abstract: Natural disasters and disease outbreaks cause substantial social turbulence and economic damage. The survival and continued operation of local small businesses and entrepreneurs are critical to the development activities in post-disaster recovery. However, these small businesses and entrepreneurs face greater challenges in accessing funding through traditional channels during a crisis. Crowdlending, also known as peer-to-peer microfinancing, has been successfully used to bypass traditional channels and raise funds directly from crowd lenders. However, it is unclear if such platforms can also be effectively used in the aftermath of crises, given that disasters induce both prosocial motivations and risk considerations in lender responses. To understand the operational implications of crowdlending for small businesses, we examine how crowdlenders respond to loan requests during a crisis and what factors moderate their responses. Drawing on the literature on disaster management and crowdlending, we hypothesize that lenders respond positively to loan requests from crisis-affected areas, and such responses are moderated by fundraising objectives and the lender's national culture. With observational data from an influential crowdlending platform and the 2014 Ebola outbreak as the treatment in a natural experiment design, we find that, on average, lenders respond positively to loan requests from crisis-affected areas, and they tend to favor loan requests emphasizing economic rather than social objectives. Furthermore, lenders from collectivistic cultures are more likely to respond positively during a crisis than lenders from individualistic ones. Our study contributes to research and practice in disaster management, particularly small business operations management during crises, by showing that crowdlending can be a useful fundraising channel for small businesses, which is meaningful for post-disaster economic development and recovery. We also offer implications for the recent conversation on the coronavirus disease 2019 (COVID-19) pandemic by analyzing and discussing the similarities and differences between the Ebola outbreak and the COVID-19 pandemic.

Mobilising new frontiers in digital transformation research: A problematization review
Information Systems Journal
Jason Thatcher
Additional authors: Amir Ashrafi, Panos Constantinides, Nikolay Mehandijiev
May 2024, p1.

Abstract: In this paper, we mobilize new frontiers in digital transformation (DT) research by deconstructing the literature's underlying assumptions and analyzing their correspondence with current theory. To do so, we conduct a problematization review across the fields of IS, strategy and entrepreneurship, organization theory and management studies, to capture the multidimensionality of DT research. Unlike systematic literature reviews commonly found in DT research, a problematization review critically questions how theoretical contributions have been constructed in past research to develop novel theoretical questions. Our findings offer three contributions. First, we uncover five research trajectories, each with its own in‐house assumptions about the nature of digital technologies and how organizations, groups and individuals interact with those technologies and the data they generate. Second, we show how individual studies within the identified research trajectories position themselves against prior research, pointing at six distinct processes of constructing theoretical contributions. Finally, we mobilize new frontiers of research by questioning DT research field assumptions that cut across the five research trajectories. We conclude by discussing the theoretical implications of our problematization review for further DT research.

Metaverse: A real change or just another research area?
Electronic Markets
Jason Thatcher
Additional authors: Christian Peukert, Hamed Qahri-Saremi, Ulrike Schultze, Christy M. K. Cheung, Adeline Frenzel-Piasentin, Maike Greve, Christian Marr, Manuel Trenz, Ofir Turel 
May 10, 2024, Vol. 34 Issue 1, p1-10.

Abstract: The Metaverse, an evolving concept that fuses physical reality with digital virtuality, offers a dynamic environment for exploration. This paper reports the panel discussion on the Metaverse and its potential implications for individuals and research. This discussion was held at the Digitization of the Individual (DOTI) workshop at the International Conference on Information Systems in December 2022. Four scientists who have researched virtual reality, immersiveness, and corresponding user behavior were invited to the panel discussion. The panelists offered their perspectives on the unique characteristics of the Metaverse, how it differs from earlier digital worlds, and the implications that the Metaverse will bring for individuals. This paper provides an introduction to the emerging phenomenon of "Metaverse" and summarizes the discussion and expert perspectives on the topic. Furthermore, this paper links the discussion to the ongoing discourse in the literature, setting the stage for further investigations by providing explicit research avenues and questions.

Deconstructing Technostress: A Configurational Approach to Explaining Job Burnout and Job Performance
MIS Quarterly
Jason Thatcher
Additional authors: Katharina Pflügner, Christian Maier, Jens Mattke, Tim Weitzel
Jun. 2024, Vol. 48 Issue 2, p679-698.

Abstract: Understanding how technostressors lead to technostrain, such as high job burnout or low job performance, has become a core question in information systems (IS) research and practice. To unpack this relationship, we build on general systems theory to argue that the next step for technostress research is to go beyond examining the independent influences of technostressors and discuss how their interdependencies lead to technostrain. To illustrate our argument empirically, we use fuzzy-set qualitative comparative analysis (fsQCA) and identify four configurations of high- and low-intensity technostressors that lead to high job burnout and one that leads to low job performance. We show that three types of interdependencies among technostressors, i.e., complementarity, contingency, and substitution, form configurations that lead to technostrain. Within these configurations, high-intensity technostressors can mutually enhance their effects and low-intensity technostressors can buffer the impact of other high-intensity technostressors on technostrain. The results help to explain why organizational interventions that address independent technostressors may fail if they do not account for the interdependencies among technostressors. Our work provides evidence of the need to further develop theories that explain how and why interdependencies among technostressors lead to technostrain.

The Influence of Political Skill and Community Capabilities on Microtask Worker Hourly Wage: A Mixed Methods Study of Mechanical Turk
Journal of the Association for Information Systems
Jason Thatcher
Additional authors: Paul M. Di Gangi, Jack L. Howard, Charn P. McAllister
2024, Vol. 25 Issue 4, p890-935.

Abstract: Microlabor markets engage workers in temporary employment contracts to complete short-duration tasks for micropayments. Because microlabor platforms often preclude worker interaction, independent microtasking communities have emerged to allow workers to exchange ideas and interact to improve their work performance. Research has yet to take an in-depth look at how workers utilize microtasking communities to mitigate unpaid coordination costs to improve their financial productivity. The present study uses political skill as a theorizing lens to investigate how microtask workers utilize the capabilities of these communities that influence their ability to avoid financial marginalization. Using pseudo-ethnography and thematic analysis, we employed a sequential mixed methods design to identify how community capabilities and ideological beliefs influence worker performance. These insights then informed the design of an empirical study using survey data from 253 Amazon Mechanical Turk workers who use microtasking communities to test our research model. We found that politically skilled individuals use community capabilities, subsequently influencing their hourly wage. We also found that microtasking ideology weakens the effects of political skill on community capabilities and their influence on hourly wages. We discuss several contributions to the political skill and microtask literature.

Chatbot Interactions: How Consumption Values and Disruptive Situations Influence Customers' Willingness to Interact
Information Systems Journal
Jason Thatcher
Additional authors: Marco Meier, Christian Maier, Tim Weitzel
Sep. 2024, Vol. 34 Issue 5, p1579-1625.

Abstract: Chatbots offer customers access to personalized services and reduce costs for organizations. While some customers initially resisted interacting with chatbots, the COVID‐19 outbreak caused them to reconsider. Motivated by this observation, we explore how disruptive situations, such as the COVID‐19 outbreak, stimulate customers' willingness to interact with chatbots. Drawing on the theory of consumption values, we employed interviews to identify emotional, epistemic, functional, and social values that potentially shape willingness to interact with chatbots. Findings point to six values and suggest that disruptive situations stimulate how the values influence WTI with chatbots. Following theoretical insights that values collectively contribute to behavior, we set up a scenario‐based study and employed a fuzzy set qualitative comparative analysis. We show that customers who experience all values are willing to interact with chatbots, and those who experience none are not, irrespective of disruptive situations. We show that disruptive situations stimulate the willingness to interact with chatbots among customers with configurations of values that would otherwise not have been sufficient. We complement the picture of relevant values for technology interaction by highlighting the epistemic value of curiosity as an important driver of willingness to interact with chatbots. In doing so, we offer a configurational perspective that explains how disruptive situations stimulate technology interaction.

Exploring Contrasting Effects of Trust in Organizational Security Practices and Protective Structures on Employees' Security-Related Precaution Taking
Information Systems Research
Jason Thatcher
Additional authors: Malte Greulich, Sebastian Lins, Daniel Pienta, Ali Sunyaev
Dec. 2024, Vol. 35 Issue 4, p1586-1608.

Abstract: Encouraging employees to take security precautions is a vital strategy that organizations can use to reduce their vulnerability to information security (ISec) threats. This study investigates how the bright- and dark-side effects of trust in organizational information security impact employees' intention to take security precautions. Employees who trust organizational security practices are more committed to protecting the organization and are more willing to take security precautions. To foster trust in organizational security practices and security commitment, ISec managers should establish a trusting security climate to ensure that employees can speak freely about the security problems they face in their work and receive support to resolve those problems if needed. This study also alerts managers to the potential adverse consequences of employees' trust in the organization's protective structures. We find that employees' trust in the organization's protective structures can backfire, making employees complacent regarding security. Further analyses indicate that security mindfulness mediates the influence of security complacency and security commitment on precaution taking. This study contributes by exploring and verifying the bright- and dark-side effects of trust in organizational ISec. Employees' precautionary security behaviors are vital to the effective protection of organizations from cybersecurity threats. Despite substantial security training efforts, employees frequently do not take security precautions. This study draws from trust theory and mindfulness theory to investigate how the bright- and dark-side effects of two conceptualizations of trust in organizational information security impact employees' precaution taking. Insights drawn from a survey of 380 organizational employees suggest that employees who trust their organization's security practices are more committed and less complacent in protecting their organization and more likely to take security precautions. In contrast, we find evidence of the dark-side effect of employees' trust in organizational protective structures by showing that such trust can lead to complacency regarding security. Analyses indicate that security mindfulness mediates the influence of security complacency and security commitment on precaution taking. These results highlight the crucial roles of security commitment, security complacency, and security mindfulness in shaping employees' precaution taking. This study contributes to information security research by providing empirical evidence concerning the simultaneous bright- and dark-side effects of employees' trust in organizational information security, thereby creating valuable opportunities for researchers to theorize about the ways in which trusting beliefs shape employees' security behaviors. 

Rethinking Gamification Failure: A Model and Investigation of Gamified System Maladaptive Behaviors
Information Systems Research
Jason Thatcher
Additional authors: Shih-Lun "Allen" Tseng, Heshan Sun, Radhika Santhanam, Shuya Lu
Dec. 2024, Vol. 35 Issue 4, p1743-1765.

Abstract: Current studies show gamification, the integrating of game design elements into target systems, enhances user engagement and instrumental task outcomes. Despite its potential for improving behavioral outcomes, gamification can also lead to maladaptive behaviors, behaviors directed at misappropriating gamified systems. We conceptualized gamified system maladaptive behaviors (GSMB), which involve technology and gamified task maladaptations. We developed a model that depicts three drivers of GSMB from design elements, how they fulfill or frustrate psychological innate needs, which in turn drive GSMB, and how GSMB affect task performance. We tested how the three drivers of design elements affect GSMB in Study 1 by empirically examining users of a gamified system, Pocket Points. The results support our conceptualization of GSMB, and design issues as its antecedents. To further unpack this relationship, we then employed a within-subject experiment and a follow-up survey in Study 2. By manipulating the design issues, we found that GSMB adversely affect task performance, because these users may focus too intently on winning the game, at the expense of task performance. By assessing the fulfillment of psychological needs, our findings suggest that design in gamified systems may not uniformly fulfill the satisfaction of psychological needs and consequently triggers GSMB. Despite the increasing interest in gamified systems and excitement about their potential positive impact on user engagement, a few studies have started to note gamification failures, which can result from user maladaptive behaviors, or behaviors directed at misappropriating gamified systems. In this research, we examine how such maladaptive behaviors can result from design issues of gamified systems and how such behaviors impact task performance. To date, little is known about design issues which may drive users to maladapt, and why they maladapt gamified systems. We systematically conceptualize gamified system maladaptive behaviors (GSMB) as having two dimensions: technology maladaptation and gamified task maladaptation. Based on goal-setting theory and self-determination theory, we develop a research model of GSMB. The model depicts three drivers of GSMB: game-task goals misalignment, game-task complexity, and gamification structure injustice, and how they fulfill or frustrate psychological innate needs (i.e., needs for autonomy, competence, and relatedness), which in turn drive GSMB. We conducted two studies using different contexts. We tested the model with Study 1 empirically examining users of a gamified system, Pocket Points. With Study 2, we employed a within-subject experiment. By manipulating the design issues, we assessed the fulfillment of psychological needs induced by the gamified system. The results largely support our conceptualization of GSMB and the research model, highlighting the design issues as the main drivers of GSMB, and that the greater the GSMB, the greater the negative impact on task performance. Findings from this research have implications for both information systems research and gamification practices.

Large language models present new questions for decision support
International Journal of Information Management
Kai Larsen
Additional authors: Abram Handler, Richard Hackathorn
Dec. 2024, Vol. 79, 102811.

Abstract: Large language models (LLMs) have proven capable of assisting with many aspects of organizational decision making, such as helping to collect information from databases and helping to brainstorm possible courses of action ahead of making a choice. We propose that broad adoption of these technologies introduces new questions in the study of decision support systems, which assist people with complex and open-ended choices in business. Where traditional study of decision support has focused on bespoke tools to solve narrow problems in specific domains, LLMs offer a general-purpose decision support technology which can be applied in many contexts. To organize the wealth of new questions which result from this shift, we turn to a classic framework from Herbert Simon, which proposes that decision making requires collecting evidence, considering alternatives, and finally making a choice. Working from Simon's framework, we describe how LLMs introduce new questions at each stage of this decision-making process. We then group new questions into three overarching themes for future research, centered on how LLMs will change individual decision making, how LLMs will change organizational decision making, and how to design new decision support technologies which make use of the new capabilities of LLMs. • We discuss how language models might change decision support systems. • We argue large language models are more versatile than prior decision support tools. • We propose that such versatility introduces new questions during decision making. • We organize new questions using a classic three-stage model of decision making. • We also propose three future directions for research in decision support.