The Impact of Evidence on Policy Demand

Participants will be provided with information about social policies in Mexico

This study is a randomized online survey experiment that examines how exposure to information about the impacts of a well-known program (Mexico's Conditional Cash Transfer Program (CCT), Progresa) influences beliefs about program effectiveness and policy preferences for CCTs and evidence generation. Respondents are first asked about their knowledge of the program and their personal or household exposure to it. Participants are then randomly assigned to receive summary evidence of program impacts. As secondary treatments a random subset will receive narratives about beneficiary experiences and another one will be asked to search the information by themselves. After the intervention, respondents answer questions on perceptions of program effectiveness across multiple domains (poverty reduction, children’s health and education, long-term prospects, corruption), support for conditional cash transfers and similar programs, and views on the role of evidence in policymaking.

2025-09-02

2025-09-30

The primary outcomes are beliefs about program effectiveness, demand for conditional cash transfer policies, and demand for evidence generation of social programs.

Comparison between narrative-based and evidence-based information, comparison between control and search arm, whether effects persist over time, and whether evidence closes polarization gaps.

Participants will be randomly exposed to different types of information about social programs in Mexico.

Participants are randomly assigned to one of four groups. Group 1 (Impact Evidence) receives an easy-to-understand summary of academic findings explaining that Progresa was rigorously evaluated using an experimental design and detailing its significant effects on children’s education (, health, nutrition, and long-term household income. Group 2 (Recipient Narratives) reads short, accessible quotes from beneficiaries reflecting on how Progresa affected their education, health, nutrition, and long-term well-being. Group 3 (Information Search) is instructed to independently search for credible, objective information on Progresa’s impacts across education, health, nutrition, and long-term outcomes, with an incentive: participants will receive a reward if their later survey responses accurately reflect the information and include citations from the sources they consulted. Finally, Group 4 (Control Group) reads a brief, neutral passage on healthy habits—such as nutrition, sleep, and well-being—that is unrelated to the main topic of the survey.

randomization done via qualtrics surveys

individual

No

4,000 individuals (main analysis will encompass 1,000 treated and 2,000 control; secondary analyses will encompass 1,000 treated and 1,000 control)

4,000 individuals (main analysis will encompass 1,000 treated and 2,000 control; secondary analyses will encompass 1,000 treated and 1,000 control)

1000 treated and 2,000 control for primary outcomes, 1,000 and 1,000 control for outcomes measured using follow-up survey.

For 1,000 treated vs. 2,000 control, MDE is 0.11. For 1,000 treated vs. 1,000 control MDE is 0.13 (assumes mean 0, sd 1, alpha 0.05, power 0.8)