Why Research (Does Not) Affect Policy?: Experimental Evidence on the Role of Political Bias

How to cite

Hombrados, Jorge Garcia; Biel, Pedro Rey; Ozcan, Berkay; Jansen, Marcel; Roldán, Antonio. 2022. "Why Research (Does Not) Affect Policy?: Experimental Evidence on the Role of Political Bias." AEA RCT Registry. AEARCTR-0008967. https://doi.org/10.1257/rct.8967-1.7

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Jorge Garcia Hombrados* , Pedro Rey Biel , Berkay Ozcan , Marcel Jansen , Antonio Roldán

* Primary investigator

Registry ID
AEARCTR-0008967
DOI
https://doi.org/10.1257/rct.8967-1.7
First registered
Last updated
Status
Completed
Trial dates
2022-05-25 – 2023-05-31
Countries
Spain ()
JEL codes
D72, F78
Registry record
http://www.socialscienceregistry.org/trials/8967

Abstract

One of the critical issues in bridging the divide between scientific evidence and policy is the "perceived credibility" of the evidence. Policymakers may be more or less inclined to follow scientific evidence if they believe the messenger, i.e. who presents the scientific evidence, has an ideological agenda. This may be true even in policy areas with a pressing need for scientific evidence and even when the evidence is produced with the highest scientific standards. This study tests whether the divergence in political biases between who send the evidence and the politician who receives it affects policy adoption using a Randomized Controlled Trial (RCT). Exploiting a sample of all touristic municipalities in Spain, we send policymakers in these municipalities either a policy brief or a link to a newspaper article describing a high-quality scientific-evidence on a very low-cost and effective and uncontroversial policy to increase tourism. Furthermore, we create random variation in the identity of the source of the information to signal a possible ideological bias. We combine survey and administrative data to define our sample of treated and control municipalities and the random assignment to the different treatment arms. Ex post, we then compare different measures of interest in the scientific evidence and the recommended policy. Specifically, we investigate whether policy makers pay attention and read the recommended policy/ evidence and whether they eventually adopt the recommended policy.

Intervention

See pre-analysis plan

Experimental design

See pre-analysis plan

Design details

Randomization method
Randomization done in office by a computer. See pre-analysis plan for details.
Randomization unit
Municipality level. See pre-analysis plan for details.
Clusters
5677 municipalities in the sample (No cluster in the analysis)
Observations
5677 municipalities in the sample
Arms
See pre-analysis plan for details:
Primary outcomes
See pre-analysis plan
Secondary outcomes
See pre-analysis plan

Cited by

No citation data available for this registration yet.

Keywords