From the Pulpit to the People: How Sermons Shape Beliefs and Behavior

How to cite

Ahsan, Muhammad Adil. 2024. "From the Pulpit to the People: How Sermons Shape Beliefs and Behavior." AEA RCT Registry. AEARCTR-0012715. https://doi.org/10.1257/rct.12715-2.2

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

Muhammad Adil Ahsan*

* Primary investigator

Registry ID
AEARCTR-0012715
DOI
https://doi.org/10.1257/rct.12715-2.2
First registered
Last updated
Status
On going
Trial dates
2024-09-01 – 2025-08-01
Countries
Pakistan (Punjab)
JEL codes
Z12
Registry record
http://www.socialscienceregistry.org/trials/12715

Abstract

Religion serves as an important source of beliefs for many across the world yet we know little about how religion and religious leaders contribute to belief formation. In this study, I exploit a unique partnership with the Punjab Auqaf Department in Pakistan, responsible for running state religious properties, to experimentally vary sermon content in 300 mosques and evaluate the causal impact of religious sermons on mosque-goers' social beliefs and behavior. By randomizing imams to sermon scripts emphasizing the importance of community rights and women's rights, I explore the importance of religious messaging for prosociality towards neighbors, community engagement, and attitudes towards women's social and economic participation.

Intervention

In the presence of uncertainty, beliefs drive decision-making. For example, beliefs about the importance of community, women’s appropriate role in society, and prevailing social norms all factor into decisions such as whether to act prosocially towards neighbors or support women’s labor market participation. Previous work has demonstrated how changing beliefs can engender positive social change (Bursztyn et al., 2020). However, belief formation, and the role of past experiences, religion, media, schooling, and local leaders therein remains understudied.

In this project, I focus on religious leaders and study the importance of one source of information in Muslim society for belief formation: sermons. Delivered before the Friday congregational prayer by an imam, the sermon, or khutbah, is often considered one of the most important sources of spiritual, social, and moral guidance for Muslims. While such messaging may partially help explain the strong correlation between moralistic religion and prosociality observed globally (Caicedo et al., 2023), concern has been raised about the potential for sermons to preserve conservative social norms but evidence is lacking. Since the information in sermons is public, the audience often exclusively male, and the source a local religious leader, a causal evaluation of Friday sermons offers a unique opportunity to study belief formation in the presence of others, intra-household belief transmission, the sources of religious leaders' authority, and the potential for sermons to serve as policy tools for information transmission.

To this end, I work with the Punjab Auqaf Department, which is responsible for running state mosques, to randomize sermon content and evaluate effects on individuals' beliefs and behavior. I study impacts on prosociality towards neighbors and attitudes about women's economic and social participation to further our understanding of the role of religion, religious leaders, and religious messaging in shaping social beliefs and behavior.

Experimental design

The study is a randomized control trial implemented in collaboration with the Punjab Auqaf Department. To evaluate the role of religious leaders and religious messaging in shaping beliefs and behavior, Auqaf imams, who are state employees, will be given sermon scripts and expected to deliver them verbatim. Enumerators will record the sermon delivered to evaluate compliance.

300 Auqaf imams will be randomized to either a business-as-usual control condition, treatment 1 (sermon scripts on women's rights), or treatment 2 (sermon scripts on community rights). Randomization will be stratified by whether the mosque is commercial or residential and congregation size (categorized into three bins).

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
Mosque
Clusters
300 Mosques
Observations
4500 male mosque-goers (and wives for those who consent); 300 imams
Arms
-Control (100 mosques): Business-as-usual sermons -Treatment 1a (25 mosques): Fully religious sermon on women's rights with information sharing script -Treatment 1b (25 mosques): Fully religious sermon on women's rights with no information sharing script -Treatment 1c (25 mosques): Partially secular sermon on women's rights with information sharing script -Treatment 1d (25 mosques): Partially secular sermon on women's rights with no information sharing script -Treatment 2a (25 mosques): Fully religious sermon on community rights with information sharing script -Treatment 2b (25 mosques): Fully religious sermon on community rights with no information sharing script -Treatment 2c (25 mosques): Partially secular sermon on community rights with information sharing script -Treatment 2d (25 mosques): Partially secular sermon on community rights with no information sharing script
Primary outcomes
- Community cleanliness - Allocations to a poor neighbor and women's employment fund in dictator games - Signing of a pro-women's rights petition - Men's willingness to allow wives to be surveyed
Secondary outcomes
My secondary outcomes consist of self-reports about: - Interaction with neighbors - Importance of neighbors - Trust in neighbors - Likelihood of helping neighbors - Interaction with wives and daughters - Perceptions of women's competence, education, and employment - Psychological well-being of women - Aspirations for daughters - Beliefs about social norms - Relative importance of women's vs neighbors' rights
Partners
Name: Punjab Auqaf Department Partner_type: government Url: https://auqaf.punjab.gov.pk

Cited by

No citation data available for this registration yet.

Keywords