The Role of Evidence in Policy Adoption
How to cite
Neilson, Christopher; Morales, Daniel; Rao, Gautam. 2018. "The Role of Evidence in Policy Adoption." AEA RCT Registry. AEARCTR-0003324. https://doi.org/10.1257/rct.3324-2.1
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
This study examines whether providing information on the evidence behind a program increases its adoption by government officials. To test this, we exploit the nationwide scale-up of an education campaign in the Dominican Republic and randomly vary whether school officials receive information on the existing evidence of the impact of the program. A treatment arm with financial incentives acts as a benchmark. Further, we analyze whether technical assistance and additional reminders increase take-up.
Experimental design
This study is based on the national roll-out of the Aprendiendo el Valor de la Educación (AVE) program. AVE is an at-scale evaluation of interventions that use videos to present accurate and clear information on the potential benefits and costs of schooling to 7th through 12th-grade students in the Dominican Republic. The program was piloted in 25 percent of all public middle schools in 2015 and was extended to 50 percent of all public middle schools in 2016. Berry, Coffman, Morales, and Neilson (2018) shows that AVE significantly increases test scores and reduces drop-out rates of participating schools.
The Ministry of Education of the Dominican Republic has decided to scale up the AVE program in the 2018-19 academic year. The scaled-up version will be similar to the one tested before, but its delivery to schools will change. While the school directors in the pilot were personally trained and incentivized to implement the program, the scale-up will rely on directors following an implementation protocol. All directors in the country will receive an email with links to the videos, and with precise implementation instructions. All school directors will also receive a phone call to confirm that they received the email.
The week of September 20-24 has been denominated the AVE week and has the aim of showing the AVE videos to every student enrolled in 7th grade or above. The idea is to introduce the program as part of the school’s regular acti
vities.
With the aim of maximizing the number of schools taking-up the program, we are interested in evaluating the performance of four interventions on top of the email.
In a first experiment, 50% of the directors will receive information about the results of the AVE pilot. This treatment is expected to encourage directors to implement AVE by highlighting the existing evidence of the impacts of the program. To measure spillover effects for this treatment, districts will be evenly divided into high share-districts (67%) and low share-districts (33%).
In a second experiment, 18.75% of the directors will receive a financial incentive in form of a lottery. This treatment is expected to serve as a benchmark and comprises the random distribution of 50 tablets among directors who show the videos to the students and upload a photo of the event to the AVE website (as specified in the instructions in the email).
In a third experiment, 25% of the directors will receive additional reminders from a call center organized by the Ministry of Education. This includes a second email and second call two weeks after the initial email was sent out. Finally, if the director did not download the videos by September 24, he will receive a third email and a third call on September 25, reminding him to show the videos. This intervention is denoted as “high-intensity”.
In a fourth experiment, 50% of the directors will receive information on the assistance that the ministry provides to help schools to implement the program. This includes the telephone number to a helpline and a link to a website with further information.
With the exception that the high-intensity group will never receive the financial incentive, the four treatments will be perfectly cross-randomized.
Design details
Cited by
No citation data available for this registration yet.