AEA RCT Registry currently lists 12695 studies with locations in 171 countries.
Women-owned/led small- and medium-sized enterprises (WSMEs) are estimated to earn less than one percent of procurement spend by large corporations and governments worldwide. There is a lack of rigorous quantitative evidence on interventions designed to support WSMEs’ participation in procurement. This study uses an RCT to estimate the impact of capacity-building solutions for WSMEs --- comprising online and in-person trainings, and related activities --- on their procurement-readiness.
The Japanese government suspended the Go To Travel campaign that it had launched in 2020 because of the spread of COVID-19. However, the number of new cases of infection nationwide continues to decrease in Japan. In this study, we examine the effects of the Go To Travel campaign on the Oita-prefecture tourist decision making. Oita-prefecture is a famous tourist spot in Japan. To investigate this, we will conduct an online survey experiment on tourist decision making, under certain conditions including the nudge-based text, lack of COVID-19 vaccination certificate, reduced business hours of bars, and declaration of a state of emergency. Based on the theoretical framework in Allcott and Kessler (2019, American Economic Journal: Applied Economics), we estimate the social welfare effect i...
This study examines how the presentation of delivery-time predictions affects user behavior and satisfaction on a large digital logistics platform. Online platforms often provide estimated delivery times, but these predictions are inherently uncertain. Platforms must decide how precisely to present this information—for example, as a specific time, a broader time window, or with additional shipment checkpoint updates. In collaboration with a large logistics information platform, we conduct a randomized controlled field experiment in which users are randomly assigned to different versions of the shipment-tracking interface. The experiment varies three main dimensions: the type of prediction model used, whether intermediate shipment nodes are displayed, and the precision of the predicted d...
Transgender people experience worse labor market outcomes compared to similar cisgender peers; recently, research has found causal evidence of discrimination against this group in multiple settings. I propose a field experiment and a set of survey experiments involving fictitious A.I.-generated headshots, where the extent to which individuals “pass” as cisgender is experimentally manipulated, as is whether applicants indirectly discloses their transgender identity via a male-to-female name change. I aim to measure (1) discrimination against transgender women, (2) whether not "passing" as cisgender exacerbates discrimination (i.e., is there "passing privilege"), and (3) what mechanisms may be driving discrimination.
Across the globe, significant economic disparities prevail along gender lines. Economists have pointed to a variety of explanations for persistent gender gaps, including economic, cultural, and norms-based forces. Here, we explore the role of gender-based differences in the desire to exert influence over others. In this study, we conduct three incentivized experiments in which subjects have the ability to interfere in and dictate the choice sets and economic outcomes of future male and female participants. Accordingly, we examine the extent to which people interfere in the choices of others, and why, using a field experiment in rural Madhya Pradesh, India.
Digital credit through mobile money has expanded rapidly in Mozambique. Yet many users do not know the basic terms of the products available to them, including the service fee, the repayment deadline, the automatic deduction of wallet inflows while a debt is unpaid, and the consequences of non-payment. This study tests whether accurate information about Txuna, the digital credit product offered through M-Pesa, changes how market micro-entrepreneurs in Maputo understand and use digital credit. We interview approximately 1,400 micro-entrepreneurs aged 21 or older who hold an active M-Pesa account registered in their own name, in the 27 markets of the official municipal register of Maputo markets that meet our inclusion criteria. Within each market, respondents are approached using a rando...
Artificial intelligence (AI) is often described as the fuel that will propel a fourth industrial revolution, but we are still learning whether its widespread adoption will be embraced by individuals or whether it will be met with fear and backlash. In this project, we use RCT to seek to understand the effects of AI adoption on political preferences.
To exit poverty, income streams of the poor must align with their liquidity needs (e.g., large investments, consumption-smoothing, saving for shocks, and smaller expenses). Traditional cash transfer programs lack flexibility, offering only fixed payment structures. This project will measure the demand for, and the welfare impacts of, the choice of a fully flexible payment schedule (frequency, timing, and amounts) in partnership with the Government of Ghana's Livelihood Against Poverty (LEAP) Program. First, we will randomize 800 (of 1,500) households into a control group (C) and 700 to a treatment group will receive cash transfers (T). Next, we will elicit beneficiary preferences over preferred times and amounts of cash transfers and measure respondents' willingness to accept (WTA)...
The first three years of life are crucial for early childhood development (ECD), with research in Western contexts highlighting the effectiveness of video feedback (VF) interventions in ECD programming. An illustrative example is the Filming Interactions to Nurture Development (FIND) program, which has consistently improved child development and caregiver outcomes. Nevertheless, FIND has not undergone evaluation in a non-Western and low-income setting, where its potential to impact caregiver and child outcomes could be significant. Therefore, we will conduct a 10-week randomized controlled trial in rural villages across Hubei Province. Villages are randomized 1:1 to the FIND intervention or control. Trained coaches deliver 10 home-based sessions, filming caregiver-child interactions and...
This study evaluates whether access to a natural-language large language model assistant improves consumer decision-making and scam avoidance in simulated online second-hand marketplaces. Adult participants recruited through Prolific will complete nine marketplace choice tasks. In each task, participants will evaluate five listings for a standardized used consumer electronic product and must select the listing they perceive to be the best available purchase. Participants will be randomly assigned at the individual level to one of two conditions: an AI-assistance treatment condition or a no-assistance control condition. Participants in the treatment condition will be permitted to consult an LLM-based assistant of their choosing while evaluating the listings. Participants in the contro...