AEA RCT Registry currently lists 12932 studies with locations in 174 countries.
Using an experiment, we intend to study the causal effects of providing professional forecaster estimates of inflation on incentivized investor behavior (i.e., asset pricing and portfolio allocation).
AI is increasingly used to advise hiring and pay decisions. We run a large-scale experiment in which human employers set the maximum wage they would pay to hire each of a series of workers — first on their own, then with wage recommendations from a commercial AI chatbot, and finally again without it. We vary across subjects whether the hiring is real or simulated, and whether the worker is hired for a math or literature task. We test the effect of AI assistance on (i) the wages employers offer, (ii) the employer and worker surplus, and (iii) the gender gap in those offers, as well as (iv) whether these effects persist after the AI is withdrawn and (v) how the surplus gain, if any, arises. We also explore how employers' behavior and the effect of AI assistance vary across employer and wo...
Artificial intelligence may change how many workers firms need in different occupations. This study asks how households form beliefs about the labor-market effects of AI adoption, and whether these beliefs matter for their economic decisions. We ask whether learning what firms themselves expect about AI-driven changes in employment for an occupation changes workers' beliefs about their own job security, earnings, and outside options, and their expectations of national unemployment. We then ask whether these belief changes affect household spending and saving, and demand for information about jobs and AI skills.
Whether people act on collective problems depends not only on what they know, but on whether they see those problems as real, theirs, and solvable. Most higher education still presents social issues as abstract exercises, so it is unclear if authentic contact with live problems and named partners changes civic dispositions or merely makes problems feel more interesting. This randomized trial tests whether these dispositions change when a real-world problem is attached to a named external partner, rather than presented only as a “real problem,” and when peer feedback comes from more or less disciplinary mixed groups.
Millions of children worldwide are impacted by school-related gender-based violence (SRGBV), which worsens academic performance, reduces educational attainment, and harms mental, physical, and reproductive health. In this study, we evaluate an intervention in Nigerian upper primary school classes which trains pupils and teachers simultaneously to increase reporting, enable mitigation, and ultimately decrease SRGBV. We test multiple arms across 365 schools in Oyo State, including the full intervention which trains both pupils and teachers at the same school; a teacher-only training arm; and a cascading teacher training arm, in which teachers are trained with the teacher training, and additionally trained to deliver the pupil curriculum to pupils, in order to understand how the interventi...
This randomized online study examines how prediction-market odds and polling information affect voters’ candidate preferences, intended political behaviour, and electoral beliefs. Eligible adults residing in several US States including Michigan, Ohio, Georgia, North Carolina, and Texas who intend to vote in the relevant 2026 U.S. Senate election will participate. Participants will be randomly assigned within each state to three arms providing either no information, odds from prediction market, or information from polls. The primary outcomes are candidate preference, turnout intention, and intended donations to the preferred candidate’s campaign. Secondary measures include perceived winning probabilities, expected margins of victory and weekly news sources.
This study evaluates a production-grade Generative AI tool that offers real-time reply suggestions to licensed insurance brokers at Ant Insurance, China's largest insurance brokerage, during live customer advisory conversations. The tool was introduced in 2025. In the first stage, a subset of brokers received access to the tool ahead of the wider release. In the second stage, access was randomized at the customer level over a 20-day window: each customer who contacted the platform was assigned to a treatment group, in which the tool was active in the broker's workspace for that customer's conversations, or to a control group, in which the tool was disabled. The primary outcomes are realized premium revenue per customer and the conversion rate within 15 days of first contact. Secondary ...
Post-harvest grain losses severely limit food security and income among smallholder farmers in Sub-Saharan Africa. Although hermetic storage bags (HBs) effectively reduce losses, adoption remains constrained by liquidity shortages at harvest. While deferred payment schemes increase uptake, high repayment rates are needed for scalability. This randomized controlled trial in Kakamega County, Kenya, evaluates the impact of different options to increase repayment rates, including community-led repayment structures, payment in instalments, and SMS nudges. The overall goal is to identify sustainable models for scaling agricultural technologies in general, and hermetic storage bags in particular.
Do decision-makers account for how people select into opportunities when evaluating members of different groups? In hiring, employers observe applicants rather than everyone who could have applied. Workers may apply based on private information about their suitability for a job. When this selection differs across groups, applying can reveal different information about an applicant’s likely productivity. This study asks whether employers overlook that information and consequently misjudge productivity differences between groups of applicants. It examines how these beliefs affect wage offers and whether making the implications of selection explicit improves judgments and decisions.
This study examines social and behavioral factors that shape workers' willingness to report violations of their workplace rights. Using a series of randomized survey experiments with approximately 10,000 hourly workers in the U.S. retail and food-service industries, we investigate how perceptions of labor standards violations, social norms, and beliefs about the costs and benefits of reporting influence workers' decisions to take action.