AEA RCT Registry currently lists 12464 studies with locations in 171 countries.
We study whether telling loan applicants what other borrowers ask for changes how much they themselves request. Does it matter whether the information is about male peers, female peers, or borrowers in general? And do men and women react differently? We work with a microfinance institution in Kenya, and the study is embedded in the institution's routine loan-application phone call. Applicants are randomly assigned to hear the average requested amount of a random sample of similar past borrowers (same area and age bracket): either male applicants, female applicants, or applicants in general. A control group receives no information.
The aim of this project is to investigate the impact of trade cost uncertainty on the expectations of Swiss firms and more specifically on their turnover and prices. To this end, an online survey experiment is conducted as part of the KOF Business Survey, in which participating firms receive a vignette - a hypothetical trade cost scenario.
We will explore whether strengthening local identity can influence cooperation between host-community members and migrants in a Colombian community that has experienced substantial Venezuelan immigration. Using a lab-in-the-field experiment, participants will engage in incentivized economic games that measure different dimensions of prosocial behavior, including volunteerism, altruism, cooperation, and common-pool resource management. We vary group composition by migration status and expose participants to contextual cues designed to increase or decrease the salience of local identity. We investigate whether these cues affect cooperation and collective action in mixed groups and whether their effects vary with group composition. The project contributes to understanding how local identit...
How well do people anticipate the effects of social observation on their own behavior? We test this in an online real-effort experiment in which participants complete four within-subject 5-minute rounds whose output funds a charitable donation. The rounds vary along two dimensions: (i) whether the participant's effort is observed live by another person via video, and (ii) whether the participant receives a personal piece-rate bonus on top of the donation. Each participant is assigned to one of eight counterbalanced orders that nest the observability manipulation (Anonymous vs. Video) either in the first or second half of the experiment and the piece-rate manipulation (standard vs. higher rate) in the other half. Before any round begins, we elicit participants' predictions of their own e...
This study tests whether an affordable, scalable mental health program can reduce psychological barriers to work in a socially stratified labor market. I conduct a two-arm randomized controlled trial among working-age adults in India with moderate to severe depression. Participants in the treatment group receive a guided self-help program adapted from the World Health Organization's Step-by-Step protocol, delivered through a chatbot with weekly support calls from trained guides. The control group receives care-as-usual. Randomization is stratified by gender, baseline income, and baseline mental health. Outcomes are measured at 3-, 6-, and 12-month follow-ups and cover mental health, economic, and social outcomes, as well as the mechanisms linking mental health to economic participation....
This experiment studies electronics repair prices in 12 different countries. We are interested in the difference in prices between when the customer states the required repair (an ORDinary good) and when the customer states they have no idea what is wrong (a CREdence good). Each repair store will conduct an ORD and CRE repair, where the order of these repairs will be randomized at the store level. Repair stores will be randomly selected from a list scraped using Google Maps, but are ‘excluded’ if they cannot complete both repairs. The sample size will be 30 stores (60 total repairs) per country, or 720 repairs overall. Note: The data collection is conducted by a commercial partner. They have begun the data collection, but have not shared any provisional data with any of the authors a...
Poor sleep is common among adults living and working in low-income urban settings, but there is little evidence on whether improving sleep changes health or economic outcomes in these contexts. Existing interventions have mostly tried to increase how long people sleep. This study instead evaluates a program that targets sleep quality. The program is adapted from Cognitive Behavioral Therapy for Insomnia (CBT-I), the standard non-drug treatment for insomnia in high-income countries, where it is typically delivered by trained clinicians over seven or more individual sessions. Our adaptation condenses the core content into four weekly group sessions, delivered through pre-recorded video modules and trained lay facilitators rather than health professionals, and is designed for low-cost deli...
Many experiments have compared the behavior of players in finitely repeated games and indefinitely repeated games, but to our knowledge no experiments have examined the impact of mixed-continuation games where some some players repeat the game indefinitely while others cycle through for a single round. In this experiment, we study a mixed-continuation game based on plea bargaining in the United States criminal justice system. In this system prosecuting district attorneys (DAs) and public defenders (PDs) will generally interact frequently across many cases. On the other hand, defendants typically only interact with these attorneys for a single case. This incentivized experiment tests whether the long term relationships between DAs and PDs harm clients or reduce the fairness of the cour...
The goal of this study is to examine whether perceived destination safety is a barrier to rural–urban migration and geographically broad job search, particularly for women. The study population consists of rural jobseekers aged 18–35 living in Amathole District, Eastern Cape, South Africa. Participants will be assigned to receive either credible, location-specific information about safety conditions and safety-related amenities in potential migration destinations or to receive no destination-safety information. We observe impacts of information on beliefs about safety and where respondents choose to migrate to (both intentions and realised outcomes) and/or search for work.
A large consulting firm is conducting an experiment with its employees and will share the resulting data with us so that we can study the treatment effects. In brief, employees will have the opportunity to create and train their own AI coworkers: AI agents trained specifically on an employee’s knowledge. The main treatment arm randomizes incentives: half of the employees receive credit for the AI coworkers they create through recognition, visibility, and promotion opportunities. A secondary, cooperative treatment arm randomizes whether the company combines similar AI coworkers.