|
on Experimental Economics |
| By: | Tetsuya Kaji |
| Abstract: | We study the design of multi-arm experiments that identify local average treatment effects under imperfect compliance. With three or more arms, conventional randomization fails to identify margin-specific effects even under strong extensions of monotone compliance. A choice-theoretic analysis unifies these extensions and motivates a two-stage design: subjects report a preferred arm and are randomized conditional on it. If treatment is confined to the assigned or reported arm, one-sided noncompliance is restored conditional on the report, identifying the effect between the two for their compliers. We discuss the design's applicability and feasibility in relation to the choice model. |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2610.25667 |
| By: | Antoine Malezieux (BSB - Burgundy School of Business (BSB) - ESC Dijon-Bourgogne, CEREN - Centre de Recherche sur les Entreprises - BSB - Burgundy School of Business (BSB) - ESC Dijon-Bourgogne); Eli Spiegelman (BSB - Burgundy School of Business (BSB) - ESC Dijon-Bourgogne, CEREN - Centre de Recherche sur les Entreprises - BSB - Burgundy School of Business (BSB) - ESC Dijon-Bourgogne) |
| Abstract: | Over the last four decades, a broad stream of experimental literature has been published using the Common Pool Resource (CPR) game to study how people react to congestible resources, and how to keep such resources from socially harmful overexploitation. With the goal of providing guidance to future work on this still-important paradigm, we provide a narrative review of the literature, summarizing the results for several key aspects of the experimental operationalization. We classify these aspects into two broad categories. The first describes ‘environmental' assumptions on the modeled resource problem itself. This refers to aspects of the experimental environment reflecting factors such as group size, resource size and asymmetry of access, which generally constitute the nature of the dilemma. The second category involves ‘institutional' issues related to how people might solve the problem, such as user communication between subjects, information about previous subjects' choices, and regulatory measures. |
| Keywords: | Climate change, Common pool, Commons dilemma, Experiment, Harvesting |
| Date: | 2025–05–01 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-05797893 |
| By: | Dominik Atella-Suri (University of Bonn); Annalena Tetzner (University of Bonn); Sebastian Kube (University of Bonn); Jan Börner (University of Bonn) |
| Abstract: | Negative income shocks may increase households' reliance on common-pool resources and thereby undermine sustainable resource management. We study this in a preregistered laboratory experiment with 120 participants playing a dynamic resource extraction game framed around a community forest. Participants in the Control condition play the baseline game three times, whereas participants in the Shock condition experience a temporary loss of exogenous income in the second part before income is restored in the third. This design allows us to distinguish responses to the income shock from common changes associated with repeated play and to observe behavior after the shock ends. We find that forest extraction increases in Part 2 for the Shock treatment relative to the ongoing baseline in the Control condition. After income is restored in Part 3, extraction decreases and the forest partially recovers. However, the equivalence tests do not clearly show whether behavior fully returns to its pre-shock level or whether some of the shock effect remains. Our findings support previous evidence that economic hardship can increase pressure on shared natural resources. |
| Keywords: | Resource extraction game; Forest commons; Income shock; Lab experiment; Cooperation |
| JEL: | D70 Q20 Q23 Q50 Q56 |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:432 |
| By: | Hanna Brosch; Christina Langer; Philipp Lergetporer; Sandra Mueller; Harald Pfeifer; Nele Tschoepe |
| Abstract: | Does generative artificial intelligence (AI) expand the set of tasks that middle-skilled, early-career workers can successfully perform? In a field experiment with 673 final-year IT apprentices at twelve German vocational schools, we randomly assign access to AI for completing occupation-specific tasks. To identify task expansion, we study tasks both within the apprenticeship curriculum and deliberately beyond it. AI raises productivity substantially across all task types, with gains tending to be larger on tasks where performance without AI is lowest. Crucially, productivity increases by 26–31 percentage points on tasks beyond apprentices’ formal training, providing direct evidence of AI-enabled task expansion. These productivity gains do not come at the cost of reduced immediate task comprehension. Treatment effects are generally larger among apprentices with above-median AI literacy. |
| Keywords: | generative artificial intelligence, field experiment, productivity, task expansion, middle-skilled workers, apprenticeships |
| JEL: | J24 O33 C93 I21 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_13045 |
| By: | Rose Deperrois; Adélaïde Fadhuile; Julie Subervie; Sophie Thoyer |
| Abstract: | Peer-to-peer knowledge diffusion is increasingly recognised as a key mechanism in various economic contexts and may play a crucial role in fostering the adoption of new agricultural practices among farmers, including some that relate to agroecology. However, the conditions for effective peer learning remain poorly understood, particularly regarding the role of the injection point – the first individual to receive information – within a peer group. We conduct a randomized controlled trial (RCT) with roughly 850 voluntary French farmers, randomly assigned to peer groups on a digital communication platform. Over an 18-month period, one farmer per treated group receives information on agroecology, serving as the injection point for diffusion. Treatment varies according to whether the injection point is an early adopter of agroecology or an ordinary peer. A benchmark group receives direct broadcasting of the information, so that it is directly accessible for all members. Results from the experiment show that injection points relay only a small fraction of the information to their peer groups, which ultimately results in better performance of the broadcasting diffusion strategy. However, recipients are more likely to read a post shared by an injection point than the same post when broadcasted. This highlights a trust effect coming from peer endorsement, a promising finding for the future of peer-to-peer learning policies. |
| Keywords: | Information Diffusion, Social Learning, Agricultural Practices |
| JEL: | D8 O33 Q16 Q57 |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:gbl:wpaper:2026-04 |
| By: | Matsuda, Norihiko; Fasih, Tazeen; Nomura, Shinsaku; Tas, Emcet Oktay |
| Abstract: | In the presence of gender-specific barriers, certain workplace environments and amenities may be particularly important for women, yet difficult to observe during job search due to information frictions. A large-scale field experiment on Pakistan's online job board publicized these job characteristics—namely, supervisor gender, coworker gender composition, work flexibility, commuting support, and overnight travel requirements. Information on these characteristics was randomly disclosed in job advertisements and email alerts. Analysis at the job-alert level shows that female jobseekers respond to supervisor gender, remote work, and commuting support. At the job level, when these characteristics were disclosed, the number of female applicants increased as the jobs were more favorable to women. |
| Date: | 2026–10–15 |
| URL: | https://d.repec.org/n?u=RePEc:wbk:wbrwps:11484 |
| By: | Garcia, Thomas; Lambert, Claudia; Pancaro, Cosimo; Rimbaud, Claire; Soldà, Alice |
| Abstract: | Holding limits are a central design feature of proposed retail central bank digital currencies (CBDC). While such limits mechanically prevent individuals from exceeding a maximum balance, they may also generate behavioral responses whose directions are a priori unclear. We conducted an incentivized online experiment with participants from four euro-area countries, exogenously varying the presence and level of a holding limit to causally separate direct mechanical effects from behavioral effects. We find that behavioral responses increase holdings of the means of payment subject to the holding limit, consistent with an attraction effect. This effect is material only when the limit is high enough. Enabling reverse waterfall —automatic conversion into this limited means of payment— increases adoption while reducing holdings. These results suggest that combining holding limits with automatic funding mechanisms can support CBDC use while containing holdings, thereby helping to mitigate potential financial stability risks. JEL Classification: E42, C91, D14, E58 |
| Keywords: | behavioral economics, CBDC, digital currency, experimental economics, holding limits |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:ecb:ecbwps:20263300 |
| By: | Kotsadam, Andreas (Ragnar Frisch Centre for Economic Research); Pieters, Janneke (Wageningen University); Rossi, Pauline (CREST-École Polytechnique); Villanger, Espen (Chr. Michelsen Institute); Zipfel, Céline (House of Sustainable Society (HoSS)); Getahun, Tigabu (EconInsight Center for Development Research) |
| Abstract: | We conducted an experiment with multiple factories in Ethiopia to test the widespread hypothesis that formal employment among women reduces fertility. Tracking respondents over eight years, we find that women who were randomly offered a job had more children than the control group. The effect on fertility emerged once the effect on wage employment faded away, around three years after baseline. There are clear effects on household income and savings, but not on bargaining power or marriage. These patterns are consistent with a Beckerian model in which the income effect dominates the substitution effect and households are credit constrained. |
| Keywords: | Female employment; fertility; industrialization; sub-Saharan Africa |
| JEL: | J13 J16 J22 O15 O55 |
| Date: | 2026–10–08 |
| URL: | https://d.repec.org/n?u=RePEc:hhs:hamisu:2026_006 |
| By: | Alberto Chong (Andrew Young School of Policy Studies, Georgia State University and Universidad del Pacifico, Peru); Daniela Valdivia (Inter-American Development Bank) |
| Abstract: | Deceased-donor registries are chronically undersupplied. This is particularly true in developing countries. We ask whether putting a face on potential donors changes behavior. In a laboratory experiment in Peru, sessions were randomly assigned to a text-only registration request or to one accompanied by a photograph of a prospective recipient. Since treatment is assigned by session, we take the session as the unit of inference. When registering is costly, a photograph raises registration by 7.0 percentage points on a control mean of 36.2 percent; the effect is concentrated in the initial costly block and is not detected in the subsequent costless block. Given that block order is confounded with cost, this is consistent with a cost-dependent treatment effect. Every subject also ranked photographs of two ethnicities; subjects register more often in the rounds showing the photograph they themselves ranked more favorably, while the same variable constructed for control subjects is null. |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:ays:ispwps:paper2630 |
| By: | Sofia Amaral; Girija Borker; Nishith Prakash; Maria Micaela Sviatschi |
| Abstract: | Frontline state agents exercise substantial discretion over citizens’ access to public services, yet little is known about whether the attitudes and beliefs shaping this discretion can be changed among incumbent bureaucrats. Partnering with the Bihar Police - one of India’s largest states that operates under comparatively severe policing capacity constraints, we conduct a cluster randomized controlled trial across 417 police stations and roughly 2, 500 active duty senior field police officers to evaluate a three-day expressive arts intervention, grounded in the Theatre of the Oppressed, designed to change police responses to gender-based violence (GBV). Approximately 7 months after baseline, treated officers exhibit more favorable attitudes toward GBV, greater empathy, and stronger legal knowledge. These changes translate into observed policing behavior: in standardized victim audits conducted one year later, officers in treated stations are 15 percentage points less likely to blame victims, dismiss fewer complaints, and provide more referrals, while administrative records show a 41 percent reduction in registration delays and, over a longer horizon, increased GBV reporting by citizens. Exploiting variation generated by officer transfers, we find suggestive evidence that these effects persist when treated officers relocate and diffuse to untreated officers working alongside treated colleagues. The effects also extend to the workplace: junior female officers working with treated senior officers report lower harassment and anxiety. These findings show that the attitudes shaping bureaucratic discretion are malleable and that changing how incumbent state agents exercise authority can improve both citizen-facing service delivery and the broader workplace environment. |
| Keywords: | police, gender-based violence, state capacity, bureaucratic discretion, gender attitudes, workplace spillovers |
| JEL: | J16 K42 J45 D73 D91 C93 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:ces:ceswps:_13050 |
| By: | Wojciech Zawadzki (University of Warsaw, Faculty of Economic Sciences); Henrik Andersson (Swedish National Road and Transport Research Institute (VTI); University of Toulouse Capitole, Toulouse School of Economics); Mikołaj Czajkowski (University of Warsaw, Faculty of Economic Sciences); Arne Risa Hole (Universitat Jaume I) |
| Abstract: | This study investigates how the design of the cost attribute shapes stated preference estimates of the value of a statistical life (VSL). Using a discrete choice experiment (DCE) on willingness to pay (WTP) for cardiovascular mortality risk reductions, we implement a randomized design that varies the cost vector across three treatments, allowing us to compare monetary vectors that differ jointly in their range and spacing. Complementary diagnostics examine scope sensitivity and potential anchoring effects induced by information disclosure and feedback mechanisms. Point estimates of VSL differ by up to roughly 25% across treatments, but the reported between-treatment comparisons are not statistically significant. These results establish neither a detectable VSL effect nor equivalence across treatments. This contrast carries direct policy implications, since cost-benefit analyses and health regulations typically rely on point estimates of VSL rather than full distributions. Secondary diagnostics indicate less-than-proportional WTP scaling over the tested risk increments, while anchoring stimuli (revealing all attribute levels upfront, ordering cost levels from extremes, or providing feedback on comprehension questions) exert no detectable influence on choices or WTP. Taken together, the results highlight the need for clearer methodological guidance on cost attribute selection in stated preference research and underscore that seemingly modest design choices can shift welfare estimates in ways that are consequential for policy appraisal, while accounting for estimation uncertainty. |
| Keywords: | cost vector effects, value of statistical life, behavioral biases, discrete choice experiment, mortality risk reduction |
| JEL: | I12 D01 D61 Q51 C83 C93 |
| Date: | 2026–10–15 |
| URL: | https://d.repec.org/n?u=RePEc:war:wpaper:2026-40 |
| By: | Bernardo Candia; Olivier Coibion; Dimitris Georgarakos; Yuriy Gorodnichenko; Maarten van Rooij |
| Abstract: | CORRECT ORDER OF AUTHORS: van Rooij, Coibion, Georgarakos, Candia, Gorodnichenko. How does perceived relative income affect household behavior and well-being? We use a randomized information experiment in a representative panel of Dutch households to generate exogenous variation in beliefs about peers' income. Lower perceived relative standing shifts expenditure toward durable goods but we find little evidence of an increase in total spending. Higher perceived relative income increases socializing and substantially raises happiness, while lower relative income increases subsequent employment. Relative standing also shapes views about inequality and redistribution. A complementary vignette experiment provides supporting evidence on mechanisms. Overall, peer comparisons matter for behavior and well-being even when total spending responses are small. |
| Keywords: | peer effects; randomized controlled trial; surveys |
| JEL: | D1 D3 D6 E21 |
| Date: | 2026–10–09 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedcwq:103918 |
| By: | Alberto Ortega; Pietro Pellerito; Patrick S. Turner |
| Abstract: | More than 23 million adults in the United States lack a high school diploma, and most who return to school drop out before graduating. We evaluate a randomized intervention of adult high school students in which coaches delivered text messages that encouraged persistence. We find that assignment to treatment increased graduation by 4.9 percentage points, and the effect of message receipt among compliers is 9.3 percentage points, a 41 percent increase relative to the control mean. The intervention operated on the intensive margin by increasing the number of courses completed, but did not affect student enrollment. The effects of the intervention are concentrated among students assigned to experienced coaches, who customized their messages more extensively. We contextualize the economic impact of the low-cost intervention by comparing labor market earnings of graduates to non-graduates, which implies an earnings benefit of $149 per treated student over four years. |
| JEL: | C93 D91 I21 I26 J24 |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35923 |
| By: | Shruti Jha (University of California, San Diego); John A. List (The University of Chicago Department of Economics and Australian National University and also NBER); Rebecca C. Royer (The University of Chicago Department of Economics); Anya Samek (University of California, San Diego Rady School of Management and NBER) |
| Abstract: | Do early investments in parents pay off even when treatment effects for children appear to fade out quickly? We answer this question by leveraging a field experiment in which households with 3-5 year-old children were randomized into a parenting program, a preschool, or a control group, combined with a natural experiment (Covid-19 school closures) that occurred ten years later. Prior to the closures, treatment effects on academic achievement had faded to zero across all groups. Yet, with Covid-19 school closures, effects of the parenting program re-emerge: test scores in control and preschool groups drop by roughly 0.25-0.30 SD, whereas scores in the parenting group drop by only 0.11 SD (difference: p = 0.01). Our findings suggest that parenting programs can generate latent human capital that only becomes visible when complementary school inputs are removed. This result is consonant with parental and school investments acting as substitutes in skill production. Because this latent parental capacity pays off precisely when school inputs fail, standard measurement tools for early childhood programs can miss crucial state-contingent value. |
| JEL: | C93 I26 |
| Date: | 2026 |
| URL: | https://d.repec.org/n?u=RePEc:bfi:wpaper:2026-131 |
| By: | Daniel Evans (University of Bonn); Georg Schneider (University of Bonn) |
| Abstract: | Many important decisions are made jointly by groups. While groups can draw on and discuss diverse perspectives, they may also suffer from counterproductive dynamics, such as herding and gridlock. To investigate this tension, we study three features that distinguish joint from individual decision-making: composition, communication, and collaborative aggregation. Simple comparisons between joint and individual decisions cannot distinguish their effects. To separate them, we design an experiment in which subjects construct portfolios to maximize returns and minimize volatility, first individually and then in one of three decision modes; each mode differs from adjacent modes by a single feature. We find that composition drives the benefits of joint decision-making: simply pooling two individuals' independent submissions together reduces portfolio concentration and volatility. Communication attenuates these gains by inducing partners to converge on similar portfolios and to prioritize maximizing returns over managing risk. Collaborative aggregation has little effect on average, but is beneficial when partners hold differing beliefs. Disentangling these features reveals an offsetting cost of communication that limits the ability of groups to make full use of members' inputs. |
| Keywords: | Group decision-making, Wisdom of crowds, Communication, Portfolio choice |
| JEL: | C92 D70 D91 G41 M54 |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:ajk:ajkdps:436 |
| By: | Bryan Stuart; Evan Taylor |
| Abstract: | This paper develops a framework that allows researchers to describe a wide class of quasi-experimental treatment effect estimates using direct analogs of the features that make a randomized controlled trial transparent. The outcome weight measures how much each observation’s outcome contributes to an estimate, and its sign defines effective treatment and control groups. We show that estimates take a Wald form, equal to the weighted outcome difference between these groups divided by an effective first stage that quantifies the identifying treatment contrast. Our framework also allows researchers to assess covariate balance, examine how differences throughout the outcome distribution contribute to mean impacts, and identify influential observations. The treatment effect weight — the outcome weight times treatment — instead measures how much each observation’s unobserved treatment effect contributes to an estimate and exactly decomposes an estimate into subgroup-specific components. When the implied weighting of an estimate is undesirable, we show how to construct alternative estimates that satisfy researcher-specified criteria. We illustrate the framework with published analyses of democracy and growth, Chinese import competition, and returns to schooling. |
| Keywords: | Treatment effect estimation; transparency |
| JEL: | C13 C81 C82 |
| Date: | 2026–10–13 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedpwp:103922 |
| By: | Francis Annan; Seth Garz |
| Abstract: | Expanding access to financial services may not be enough to develop markets in low-income settings: providers also need to be held accountable to consumers. We test this in a field experiment across 195 rural Sierra Leonean communities, upgrading Village Savings and Loan Associations (VSLAs) into mobile money and bank agents, cross-randomized with a monthly consumer feedback mechanism. Upgrading expands access, but feedback drives large operational improvements (+20.5 pp, 85%), particularly in privacy and consumer protection, raising usage and revenues. Upgraded VSLAs also show higher savings digitization and higher incomes, evidence that disciplining agent actions benefits not only consumers but also the principal that finances and stands behind its agent. Interpreted through a framework of delegated intermediation, we show accountability, not infrastructure alone, aligns provider incentives with consumer welfare and is central to designing scalable financial intermediaries. |
| JEL: | C93 D22 D82 G21 O17 |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:nbr:nberwo:35912 |
| By: | Bernardo Candia |
| Abstract: | Using multiple randomized controlled trials (RCTs) implemented through surveys of US households, I show that household spending responses to exogenous changes in inflation expectations depend on the inflation environment. In times of low inflation, higher inflation expectations arising from information treatments lead households to reduce their consumption of durable goods (the negative income/uncertainty effect dominates). In contrast, in times of high inflation, I find a sharp positive effect on durable goods spending after an exogenous shock to inflation expectations (intertemporal substitution effect dominates). |
| Keywords: | inflation expectations; surveys; randomized controlled trials (RCTs) |
| JEL: | E3 E4 E5 |
| Date: | 2026–10–06 |
| URL: | https://d.repec.org/n?u=RePEc:fip:fedcwq:103916 |
| By: | Løken, Katrine V. (Dept. of Economics, Norwegian School of Economics and Business Administration); Silliman, Mikko (Dept. of Economics, Aalto University); Willén, Alexander (Dept. of Economics, Norwegian School of Economics and Business Administration); Xu, Yaling (Dept. of Economics, Norwegian School of Economics and Business Administration) |
| Abstract: | The division of labor is fundamental to how firms organize production. Since Adam Smith, economists have recognized that dividing tasks more finely increases productivity through specialization but also raises coordination costs—a tradeoff formalized by Becker and Murphy (1992). Yet we have little systematic evidence on how this tradeoff operates within firms. Moreover, this classical perspective overlooks a second, dynamic tradeoff: firms not only produce output, but also develop a central input into future production—their workers’ human capital. A finer division of labor allows workers to develop greater expertise through repeated performance of narrowly defined tasks, but may also reduce opportunities to learn from coworkers by limiting interactions and knowledge exchange. We study both tradeoffs using population-wide Norwegian employer–employee records, linked workplace surveys, quasi-experimental variation in coworker availability, and randomized experiments with workers and managers. We find that more specialized firms exhibit higher productivity but also greater coordination costs. Their workers report more learning-by-doing but substantially less coworker learning, resulting in lower overall learning and slower wage growth. To establish the importance of coworker learning for human capital accumulation, we exploit extended coworker absences and show that losing access to colleagues reduces subsequent wage growth, particularly in occupations characterized by greater task overlap and coworker learning. To isolate the role of the division of labor from other workplace characteristics, we conduct randomized experiments that vary task allocation in otherwise identical workplaces. Workers and managers anticipate greater coordination demands, weaker coworker interactions, and lower skill accumulation under greater specialization. The experiments then ask workers to choose their preferred workplace and managers how they would organize production. Both favor greater task overlap, citing coworker interactions and learning opportunities as key reasons for their choices. Together, our findings show that the division of labor shapes not only the efficiency of current production but also the accumulation of human capital and future productive capacity. |
| Keywords: | division of labor; learning; firm organization |
| JEL: | D21 J24 L25 |
| Date: | 2026–10–16 |
| URL: | https://d.repec.org/n?u=RePEc:hhs:nhheco:2026_016 |
| By: | Mohamed Amir Omezzine (EESC-GEM - Grenoble Ecole de Management); Maissa Obay |
| Abstract: | This study investigates how adaptive learning systems, powered by Bayesian Knowledge Tracing and augmented with physiological monitoring, regulate cognitive effort and engagement across digital and non-digital quiz formats. We introduce the Cognitive Elasticity Model, extending Flow Theory and Cognitive Load Theory, to conceptualize how adaptive systems stretch learner effort toward an optimal zone of engagement. A mixed-methods design combines a controlled experiment (N = 180) with Q methodology to map subjective learner profiles across three quiz modalities: adaptive Moodle, static Moodle, and paper-based. Physiological indicators—EEG (frontal theta/alpha ratio) and ECG (heart rate variability)—complement self-reports of flow, cognitive load, and stress. Q-sorting reveals distinct engagement-effort configurations, offering new insight into how learners perceive and respond to adaptive systems. Our contribution is both theoretical—refining non-linear models of cognitive regulation—and pedagogical—proposing a replicable framework for mediating scientific insight into classroom decision-making through neuro-instrumentation and learner-centered profiling. |
| Date: | 2026–05–20 |
| URL: | https://d.repec.org/n?u=RePEc:hal:journl:hal-05789339 |
| By: | Mohamed Amir Omezzine (EESC-GEM - Grenoble Ecole de Management); Maissa Obay |
| Abstract: | This study investigates how adaptive learning systems, powered by Bayesian Knowledge Tracing and augmented with physiological monitoring, regulate cognitive effort and engagement across digital and non-digital quiz formats. We introduce the Cognitive Elasticity Model, extending Flow Theory and Cognitive Load Theory, to conceptualize how adaptive systems stretch learner effort toward an optimal zone of engagement. A mixed-methods design combines a controlled experiment (N = 180) with Q methodology to map subjective learner profiles across three quiz modalities: adaptive Moodle, static Moodle, and paper-based. Physiological indicators—EEG (frontal theta/alpha ratio) and ECG (heart rate variability)—complement self-reports of flow, cognitive load, and stress. Q-sorting reveals distinct engagement-effort configurations, offering new insight into how learners perceive and respond to adaptive systems. Our contribution is both theoretical—refining non-linear models of cognitive regulation—and pedagogical—proposing a replicable framework for mediating scientific insight into classroom decision-making through neuro-instrumentation and learner-centered profiling. |
| Date: | 2026–05–20 |
| URL: | https://d.repec.org/n?u=RePEc:hal:gemptp:hal-05789339 |
| By: | Varun Gupta; Vijay Kamble |
| Abstract: | When using a large language model (LLM) to simulate a user population, how should we combine individual behavioral histories to predict the fraction of users who take a particular action? We study this question in e-commerce, predicting observed product-level conversion rates from browsing sessions without task-specific model training. We vary two design choices, each involving a bias-variance tradeoff. Summarizing a product's sessions before inference reduces cost but can discard predictive detail, adding bias, whereas predicting outcomes for individual sessions and averaging afterward preserves that detail but requires many calls to reduce variance. Evaluating products together in one prompt can help the model infer what is typical in the setting, reducing bias, but also makes each prediction depend on which other products appear alongside it, adding noise. We evaluate both choices across five LLMs and two datasets. Our main finding is that predictions from individual sessions outperform predictions from summaries when products are evaluated separately, but summaries lead when products are evaluated together. Comparative summaries achieve the highest Spearman correlation with observed rates in 12 of 15 settings, with no alternative significantly better. Comparison thus helps summaries far more than session predictions, overcoming the advantage of behavioral detail. We interpret this difference through a Bayesian model of the second tradeoff, which suggests that comparison helps more when the inputs in a prompt reveal more about differences across products, consistent with our results for four types of comparative context. Comparative summaries also incur a small fraction of the inference cost of the session methods. With the best model for each outcome, they attain 66-89% of the correlation of the best supervised benchmark trained on thousands of labeled products. |
| Date: | 2026–10 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2610.23934 |