Voter Behavior in Regional Elections in Indonesia: A Conjoint Experiment

How to cite

Neo, Hui-Yuan; Dienes, Harry; Paul, Eitan; Afrimadona, Afrimadona. 2024. "Voter Behavior in Regional Elections in Indonesia: A Conjoint Experiment." AEA RCT Registry. AEARCTR-0014524. https://doi.org/10.1257/rct.14524-1.0

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

Hui-Yuan Neo* , Harry Dienes , Eitan Paul , Afrimadona Afrimadona

* Primary investigator

Registry ID
AEARCTR-0014524
DOI
https://doi.org/10.1257/rct.14524-1.0
First registered
Last updated
Status
In development
Trial dates
2024-10-20 – 2024-10-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/14524

Abstract

Indonesia has a weak political party system; most voters do not have a strong partisan identity and many political parties are not built on clear ideological platforms and do not take specific or consistent policy positions. Moreover, coalitions for regional executive elections in Indonesia often include ideologically distant parties that are opponents at the national level. Existing research similarly shows that a mix of factors, including policy platform and candidate quality (Fossati 2016) contribute equally to their vote choice. How, then, do voters decide which regional executive candidate to support when the coalitions backing them feature parties that are rivals at the national level? We answer this research question through a conjoint experiment that tests which policy, party, or candidate characteristics matter the most in affecting vote choice in regional executive elections – including whether what candidates wear affects voter choice. The experiment presents respondents, who are Indonesian voters, with different hypothetical profiles of candidates for district governor (bupati) or mayor (walikota) with randomized demographic, political, and career and family backgrounds. The study contributes to a growing literature on subnational political behavior in Indonesia. It will present a systematic test of characteristics that have been identified by the literature as meaningful, across the spectrum of policy positions, candidate characteristics, and party identity.

Intervention

We will field a conjoint experiment that features 10 attributes, each with between two to twelve levels. The 10 attributes are: appearance, party name, party backing the candidate, family background, education level, economic policy, social policy, career background, whether they were appointed by the national government, and democratic attitude. The experiment then presents respondents, who are Indonesian voters, with different hypothetical profiles of candidates for regent (bupati) or mayor (walikota) that rotate through this randomized set of demographic and political characteristics.

Experimental design

We will present voters with 12 pairs of hypothetical candidate profiles to choose between in a conjoint experiment. We also include a series of demographic questions in the survey.

Design details

Randomization method
Conjoint randomization of attributes by Qualtrics
Randomization unit
individual
Clusters
1500
Observations
1500
Arms
4500
Primary outcomes
Vote choice

Cited by

No citation data available for this registration yet.

Keywords