People supported by AI-powered decision support tools frequently overrely on the AI: they accept ... more People supported by AI-powered decision support tools frequently overrely on the AI: they accept an AI's suggestion even when that suggestion is wrong. Adding explanations to the AI decisions does not appear to reduce the overreliance and some studies suggest that it might even increase it. Informed by the dual-process theory of cognition, we posit that people rarely engage analytically with each individual AI recommendation and explanation, and instead develop general heuristics about whether and when to follow the AI suggestions. Building on prior research on medical decision-making, we designed three cognitive forcing interventions to compel people to engage more thoughtfully with the AI-generated explanations. We conducted an experiment (N=199), in which we compared our three cognitive forcing designs to two simple explainable AI approaches and to a no-AI baseline. The results demonstrate that cognitive forcing significantly reduced overreliance compared to the simple explainable AI approaches. However, there was a trade-off: people assigned the least favorable subjective ratings to the designs that reduced the overreliance the most. To audit our work for intervention-generated inequalities, we investigated whether our interventions benefited equally people with different levels of Need for Cognition (i.e., motivation to engage in effortful mental activities). Our results show that, on average, cognitive forcing interventions benefited participants higher in Need for Cognition more. Our research suggests that human cognitive motivation moderates the effectiveness of explainable AI solutions. CCS Concepts: • Human-centered computing → Interaction design.
Philosophical Transactions of the Royal Society B: Biological Sciences, 2021
Reputation is a fundamental feature of human sociality as it sustains cooperative relationships a... more Reputation is a fundamental feature of human sociality as it sustains cooperative relationships among unrelated individuals. Research from various disciplines provides insights on how individuals form impressions of others, condition their behaviours based on the reputation of their interacting partners and spread or learn such reputations. However, past research has often neglected the socio-ecological conditions that can shape reputation systems and their effect on cooperation. Here, we outline how social environments, cultural values and institutions come to play a crucial role in how people navigate reputation systems. Moreover, we illustrate how these socio-ecological dimensions affect the interdependence underlying social interactions (e.g. potential recipients of reputational benefits, degree of dependence) and the extent to which reputation systems promote cooperation. To do so, we review the interdisciplinary literature that illustrates how reputation systems are shaped by ...
This introductory chapter places the contributions in this volume in the larger picture of resear... more This introductory chapter places the contributions in this volume in the larger picture of research on governance in markets and organizations and highlights the structure of the volume. We argue that including embeddedness arguments in a model for purposive behavior is a fruitful ...
The present study analyzes whether and how different types of progress indicators affect the tend... more The present study analyzes whether and how different types of progress indicators affect the tendency of respondents to continue filling out a web survey, focusing on whether the progress indicators effects depend on the position of the respondent in the questionnaire. Using a sample of 2460 respondents of a Dutch online access panel, we analyze three kinds of progress indicators (linear, fast-then-slow, slow-then-fast, and a control condition) using survival analysis. The results show that the effect of the indicators on the completion rate is either negative or nonexistent, depending on the questionnaire length. Moreover, the effect of an indicator does not depend on the position of the respondent in the answering process. We interpret our findings in terms of the implicit narrative between survey designer and respondent.
In this codebook, we collect the questionnaires and summary statics of all the variables from all... more In this codebook, we collect the questionnaires and summary statics of all the variables from all for studies. Chapter 2 provides the questionnaires with the variable names included in the questionnaires. Chapter 3 presents the summary statistics for all the variables. Chapter 4 ...
International Journal of Electronic Commerce, 2009
Previous research on reputation systems has primarily focused on the trust building function of r... more Previous research on reputation systems has primarily focused on the trust building function of reputation systems. The current research also addresses the trust rebuilding function of reputation systems, specifically, the role of the short text comments given in reaction to negative feedback. Online markets are noisy environments; rebuilding trust is therefore often necessary. The results of two experimental studies among actual eBay users show that the text comments accompanying negative feedback are influential towards trustworthiness judgments. It also
The Oxford Handbook of Gossip and Reputation, 2019
Potentially, reputation systems in online markets are ways of safeguarding hazardous single-shot ... more Potentially, reputation systems in online markets are ways of safeguarding hazardous single-shot transactions between traders, by artificially creating a network of connections between all users of an online marketplace. Given that online markets exist and use such reputation systems, this has triggered the question how large the value of reputation is. This question has been analyzed in previous research with mixed results. After introducing the issue in some more detail, the chapter posits two arguments that may put research into the value of reputation in a somewhat different light: (1) the fact that the most often used “hedonic regression” method, which considers actual sales only, does not lead to estimates that can be straightforwardly connected to the value of reputation, neither for the seller nor for the buyer, and (2) the empirical evidence from the experimental literature on the effects of semantic feedback, which suggests effect sizes that may well be an order of magnitu...
This book deals with one of the most important problems facing modern societies the question of s... more This book deals with one of the most important problems facing modern societies the question of social cohesion. The author offers a comprehensive overview of past and present theories about the role of trust as a means of creating solidarity. She shows how in sociological ...
In a previous experiment, we have shown that risk assessments of purchasing experts are certainly... more In a previous experiment, we have shown that risk assessments of purchasing experts are certainly not better than that of subjects untrained in purchasing, and worse than the decisions made by formal models (J. Purchas. Supply Manage. 9 (2003) 191-198). Since both these results are rather counterintuitive, we conducted a series of experiments geared at replication and extension of these findings. These new experiments show that our previous results are robust, and reveal an additional finding that is both worrying and puzzling. It actually seems to be the case that for the purchasing decision tasks in our experiments, experts perform worse with growing experience. It therefore seems that, at least for the kinds of purchasing decisions under study, it does not make much sense to use expert judgments at all. However, we show that there is a way in which expert judgments can be used in combination with formal models to improve the predictive accuracy of purchasing predictions. In our case, superior predictions are made when we combine the prediction of a formal model with the prediction of the 'average expert', thereby combining the robust linear trends as encapsulated in the formal model with the more intuitive configural rules used by experts. We provide several explanations for this phenomenon.
This chapter addresses social embeddedness effects on ex ante management of economic transactions... more This chapter addresses social embeddedness effects on ex ante management of economic transactions. We focus on dyadic embeddedness, that is the history of prior transactions between business partners and the anticipation of future transactions. Ex ante management ...
People supported by AI-powered decision support tools frequently overrely on the AI: they accept ... more People supported by AI-powered decision support tools frequently overrely on the AI: they accept an AI's suggestion even when that suggestion is wrong. Adding explanations to the AI decisions does not appear to reduce the overreliance and some studies suggest that it might even increase it. Informed by the dual-process theory of cognition, we posit that people rarely engage analytically with each individual AI recommendation and explanation, and instead develop general heuristics about whether and when to follow the AI suggestions. Building on prior research on medical decision-making, we designed three cognitive forcing interventions to compel people to engage more thoughtfully with the AI-generated explanations. We conducted an experiment (N=199), in which we compared our three cognitive forcing designs to two simple explainable AI approaches and to a no-AI baseline. The results demonstrate that cognitive forcing significantly reduced overreliance compared to the simple explainable AI approaches. However, there was a trade-off: people assigned the least favorable subjective ratings to the designs that reduced the overreliance the most. To audit our work for intervention-generated inequalities, we investigated whether our interventions benefited equally people with different levels of Need for Cognition (i.e., motivation to engage in effortful mental activities). Our results show that, on average, cognitive forcing interventions benefited participants higher in Need for Cognition more. Our research suggests that human cognitive motivation moderates the effectiveness of explainable AI solutions. CCS Concepts: • Human-centered computing → Interaction design.
Philosophical Transactions of the Royal Society B: Biological Sciences, 2021
Reputation is a fundamental feature of human sociality as it sustains cooperative relationships a... more Reputation is a fundamental feature of human sociality as it sustains cooperative relationships among unrelated individuals. Research from various disciplines provides insights on how individuals form impressions of others, condition their behaviours based on the reputation of their interacting partners and spread or learn such reputations. However, past research has often neglected the socio-ecological conditions that can shape reputation systems and their effect on cooperation. Here, we outline how social environments, cultural values and institutions come to play a crucial role in how people navigate reputation systems. Moreover, we illustrate how these socio-ecological dimensions affect the interdependence underlying social interactions (e.g. potential recipients of reputational benefits, degree of dependence) and the extent to which reputation systems promote cooperation. To do so, we review the interdisciplinary literature that illustrates how reputation systems are shaped by ...
This introductory chapter places the contributions in this volume in the larger picture of resear... more This introductory chapter places the contributions in this volume in the larger picture of research on governance in markets and organizations and highlights the structure of the volume. We argue that including embeddedness arguments in a model for purposive behavior is a fruitful ...
The present study analyzes whether and how different types of progress indicators affect the tend... more The present study analyzes whether and how different types of progress indicators affect the tendency of respondents to continue filling out a web survey, focusing on whether the progress indicators effects depend on the position of the respondent in the questionnaire. Using a sample of 2460 respondents of a Dutch online access panel, we analyze three kinds of progress indicators (linear, fast-then-slow, slow-then-fast, and a control condition) using survival analysis. The results show that the effect of the indicators on the completion rate is either negative or nonexistent, depending on the questionnaire length. Moreover, the effect of an indicator does not depend on the position of the respondent in the answering process. We interpret our findings in terms of the implicit narrative between survey designer and respondent.
In this codebook, we collect the questionnaires and summary statics of all the variables from all... more In this codebook, we collect the questionnaires and summary statics of all the variables from all for studies. Chapter 2 provides the questionnaires with the variable names included in the questionnaires. Chapter 3 presents the summary statistics for all the variables. Chapter 4 ...
International Journal of Electronic Commerce, 2009
Previous research on reputation systems has primarily focused on the trust building function of r... more Previous research on reputation systems has primarily focused on the trust building function of reputation systems. The current research also addresses the trust rebuilding function of reputation systems, specifically, the role of the short text comments given in reaction to negative feedback. Online markets are noisy environments; rebuilding trust is therefore often necessary. The results of two experimental studies among actual eBay users show that the text comments accompanying negative feedback are influential towards trustworthiness judgments. It also
The Oxford Handbook of Gossip and Reputation, 2019
Potentially, reputation systems in online markets are ways of safeguarding hazardous single-shot ... more Potentially, reputation systems in online markets are ways of safeguarding hazardous single-shot transactions between traders, by artificially creating a network of connections between all users of an online marketplace. Given that online markets exist and use such reputation systems, this has triggered the question how large the value of reputation is. This question has been analyzed in previous research with mixed results. After introducing the issue in some more detail, the chapter posits two arguments that may put research into the value of reputation in a somewhat different light: (1) the fact that the most often used “hedonic regression” method, which considers actual sales only, does not lead to estimates that can be straightforwardly connected to the value of reputation, neither for the seller nor for the buyer, and (2) the empirical evidence from the experimental literature on the effects of semantic feedback, which suggests effect sizes that may well be an order of magnitu...
This book deals with one of the most important problems facing modern societies the question of s... more This book deals with one of the most important problems facing modern societies the question of social cohesion. The author offers a comprehensive overview of past and present theories about the role of trust as a means of creating solidarity. She shows how in sociological ...
In a previous experiment, we have shown that risk assessments of purchasing experts are certainly... more In a previous experiment, we have shown that risk assessments of purchasing experts are certainly not better than that of subjects untrained in purchasing, and worse than the decisions made by formal models (J. Purchas. Supply Manage. 9 (2003) 191-198). Since both these results are rather counterintuitive, we conducted a series of experiments geared at replication and extension of these findings. These new experiments show that our previous results are robust, and reveal an additional finding that is both worrying and puzzling. It actually seems to be the case that for the purchasing decision tasks in our experiments, experts perform worse with growing experience. It therefore seems that, at least for the kinds of purchasing decisions under study, it does not make much sense to use expert judgments at all. However, we show that there is a way in which expert judgments can be used in combination with formal models to improve the predictive accuracy of purchasing predictions. In our case, superior predictions are made when we combine the prediction of a formal model with the prediction of the 'average expert', thereby combining the robust linear trends as encapsulated in the formal model with the more intuitive configural rules used by experts. We provide several explanations for this phenomenon.
This chapter addresses social embeddedness effects on ex ante management of economic transactions... more This chapter addresses social embeddedness effects on ex ante management of economic transactions. We focus on dyadic embeddedness, that is the history of prior transactions between business partners and the anticipation of future transactions. Ex ante management ...
... The report “Categorizing Business Goals for Software Architectures” [28] offers an overview o... more ... The report “Categorizing Business Goals for Software Architectures” [28] offers an overview on ... rather than code-related measures, which are predominantly used in cost-saving value ... requires marketing expertise and long experience in subjective data estimation across different ...
Uploads
Books by C. Snijders
Papers by C. Snijders