How to Increase Demand for Evidence in Policymaking: Experimental Evidence from 3,500 Elected Officials

How to cite

Zelizer, Adam; Pereira, Miguel. 2026. "How to Increase Demand for Evidence in Policymaking: Experimental Evidence from 3,500 Elected Officials." AEA RCT Registry. AEARCTR-0019143. https://doi.org/10.1257/rct.19143-1.0

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

Adam Zelizer* , Miguel Pereira

* Primary investigator

Registry ID
AEARCTR-0019143
DOI
https://doi.org/10.1257/rct.19143-1.0
First registered
Last updated
Status
Completed
Trial dates
2024-04-01 – 2025-01-01
Countries
Switzerland (), Germany (), United Kingdom of Great Britain and Northern Ireland ()
Registry record
http://www.socialscienceregistry.org/trials/19143

Abstract

Despite the growing availability of policy evidence, politicians seldom use impact evidence. We argue that demand-side constraints explain part of this gap. In a survey of 3,521 local politicians, we document two key constraints on demand for evidence: (1) knowledge about evidence quality and application, and (2) political incentives. However, these obstacles can be partially overcome. We fielded an experiment that provided randomly-selected politicians with a training course on incorporating evidence in policymaking. The training increased statistical literacy and willingness-to-pay for evidence, though political constraints continue to mediate how politicians incorporate evidence.

Intervention

We fielded an intensive training intervention that provided randomly-selected politicians in Germany, Switzerland and the United Kingdom with a capacity building course on incorporating evidence in policymaking. The course was developed in conjunction with the Joint Research Center of the European Commission. A pre-treatment survey asked about potential obstacles to the incorporation of evidence in policymaking including political and knowledge constraints.

Experimental design

We partnered with the Joint Research Center (JRC) of the European Union (EU) to evaluate a program on statistical literacy for policymakers. The JRC is a department of the European Commission responsible for providing independent advice and support to EU policy. In 2023, the JRC commissioned a training program for local elected officials aimed at better connecting scientific knowledge and policymaking. The JRC released the training materials in 2024. We embedded a streamlined version of this training in an RCT with 150 local politicians recruited in three countries: Germany, Switzerland, and the United Kingdom.\footnote

To isolate the effects of attending the JRC training, we implemented a placebo design. Officials were invited to register for ``a series of online training courses on evidence-based policy making and participatory governance for local elected officials.'' When registering, officials could select one of seven training sessions, held on different dates and times. Notably, they could not choose whether they wanted the evidence-based policymaking or participatory governance training; they could only choose a date and time and were not told which session was being held. Five of the sessions were conducted in English and two in German, to accommodate officials in Switzerland and Germany, though they could also register for the English sessions.

After officials registered for a training, we randomly assigned participants to either the JRC training or the placebo training. The placebo training had the same duration (two hours) and centered on how to organize citizens' assemblies and other forms of participatory governance. It did not discuss incorporating policy evidence or expertise. The placebo sessions were conducted at the same dates and times as the JRC trainings, which means we held 14 training sessions in total.

Design details

Randomization method
Done in office by computer.
Randomization unit
Individual elected official.
Clusters
NA
Observations
150
Arms
150
Primary outcomes
Survey and behavioral data from participating officials.

Cited by

No citation data available for this registration yet.

Keywords