Policy Diffusion, Popularity Cues, and Ideological Alignment: An RCT with Spanish Municipalities

How to cite

Capozza, Francesco; Rey-Biel, Pedro; Ozcan, Berkay; Schjetnan, Sergio. 2026. "Policy Diffusion, Popularity Cues, and Ideological Alignment: An RCT with Spanish Municipalities." AEA RCT Registry. AEARCTR-0018667. https://doi.org/10.1257/rct.18667-2.0

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

Francesco Capozza* , Pedro Rey-Biel , Berkay Ozcan , Sergio Schjetnan

* Primary investigator

Registry ID
AEARCTR-0018667
DOI
https://doi.org/10.1257/rct.18667-2.0
First registered
Last updated
Status
In development
Trial dates
2026-06-02 – 2026-11-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/18667

Abstract

We study whether ideological alignment and descriptive-norm (popularity) cues drive policy diffusion among Spanish municipalities. Municipalities receive an email offering a free Wikipedia update service and information about prior adopters. We experimentally vary (i) whether highlighted prior adopters are ideologically aligned or cross-partisan relative to the recipient, and (ii) whether the email implies high or low adoption rates. The 2×2 factorial design (plus two popularity-only arms and a pure control) allows causal identification of alignment effects, popularity-cue effects, and their interaction. The pre-analysis plan is attached as a supplementary document.

Experimental design

Seven-arm experiment: a 2 (Aligned vs. Cross-partisan ideology cue) × 2 (High vs. Low popularity cue) factorial, plus two popularity-only arms (High only, Low only), and a pure control. Municipalities are assigned with equal probability (≈809 per arm). Assignment uses stratified block rerandomization (Morgan & Rubin, 2012) on four dimensions: governing-party ideology, population tertile, tourism-rate tertile, and prior treatment history. Of 1,000 candidate randomizations, the draw minimizing the maximum absolute standardized mean difference (max-ASMD < 0.10) across 13 covariates was selected (achieved max-ASMD = 0.053).

Design details

Randomization method
Computer
Randomization unit
Municipalities are assigned with equal probability (≈809 per arm). Assignment uses stratified block rerandomization (Morgan & Rubin, 2012) on four dimensions: governing-party ideology, population tertile, tourism-rate tertile, and prior treatment history. Of 1,000 candidate randomizations, the draw minimizing the maximum absolute standardized mean difference (max-ASMD < 0.10) across 13 covariates was selected (achieved max-ASMD = 0.053).
Clusters
na
Observations
5,666 Spanish touristic municipalities
Arms
5,666 Spanish touristic municipalities with verified contact details, excluding 184 that responded in a prior round (Round 1). The universe is drawn from FEMP records and public municipal websites, restricted to municipalities meeting at least one tourism criterion (FITUR participation, coastal/border location, tourism office, registered landmark, SEGITTUR membership, or CaixaBank tourism data).
Primary outcomes
Email authorization: whether the municipality responds authorizing a Wikipedia update (click on the instruction link in the email). Baseline rate: 10.54% (García-Hombrados et al., 2024). Wikipedia implementation: whether the municipality's Wikipedia page is actually updated following authorization. Baseline rate: 2.55%.

Cited by

No citation data available for this registration yet.

Keywords