Do prayers crowd out donations?

How to cite

Thunstrom, Linda. 2017. "Do prayers crowd out donations?." AEA RCT Registry. AEARCTR-0002653. https://doi.org/10.1257/rct.2653-1.0

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

Linda Thunstrom*

* Primary investigator

Registry ID
AEARCTR-0002653
DOI
https://doi.org/10.1257/rct.2653-1.0
First registered
Last updated
Status
Completed
Trial dates
2017-11-24 – 2017-11-27
Countries
United States of America (Wyoming)
JEL codes
D12; D64
Registry record
http://www.socialscienceregistry.org/trials/2653

Abstract

Are prayers substitutes or complements to monetary donations, in the face of a natural disaster? We design an economic experiment to empirically answer this question. In the experiment, subjects are offered to donate money to hurricane Harvey victims. Our results suggest praying crowds out donations.

Intervention

Subjects are asked to donate to hurricane Harvey victims. There are three treatments: baseline (subjects are asked to donate directly), treatment think (subjects are asked to take a moment to think about the victims before donating), treatment pray (subjects are asked to pray for the victims before donating).

Experimental design

The following is the description of the experimental design found in the working paper:

The experiment is designed to examine the impact of the act of praying on online donations to hurricane Harvey victims, via the Red Cross. It entails three treatments. In the first treatment (treatment baseline), subjects are informed about the current situation of the hurricane Harvey victims, and thereafter offered to donate. In the second treatment (treatment pray), subjects receive the same information about the current situation of the hurricane Harvey victims, and are asked to pray for the victims, before they are offered to donate. We further want to be able to tease out if any impact of prayers on donations is caused by prayers increasing salience of the victims (by causing a person to spend extra time thinking about the victims), or if the act of praying has an impact that is uniquely different from that of spending time thinking about the victims. We therefore have a third treatment (treatment think), where subjects get the same initial information, and are asked to take a moment to think about the victims, before they are offered to donate. For obvious reasons, treatment pray entails religious subjects only. The other treatments entail both religious and non-religious subjects, to enable comparisons of treatment responses of both religious and non-religious subjects.
Specifically, the experiment is conducted in the following steps:
1. All subjects are asked questions on common demographics (gender, age, income, religion). To reduce the focus on religion, we also ask a couple of questions on behavior irrelevant to the study (frequency of buying organic food, spendthriftiness, political affiliation).
2. All subjects are asked to read the following short text about the hardships caused by hurricane Harvey, and how the Red Cross helps.
Hurricane Harvey was a category 4 storm that hit Texas in late August 2017, and affected people from Texas through Louisiana, Mississippi, Tennessee and Kentucky. Please read the below text from the Red Cross about the impact of hurricane Harvey, the worst flooding disaster in U.S. history (American Red Cross, "Hurricane Harvey, One-month-Update," October 2017):

"When survivors returned to their homes, they faced unimaginable destruction. Entire homes were torn to pieces or waterlogged and coated with mud. Vehicles, appliances and furniture were damaged, and ruined belongings stacked in piles on the street. While residents labored to salvage what they could, Red Cross workers visited flood-ravaged neighborhoods to provide food, water and essential relief and cleanup supplies.

Recovery from a disaster of this magnitude will take months and even years. Many people are still unable to return home, and thousands more are just beginning the long process of putting their lives back together."
3. Treatment baseline: subjects are informed that they have $5, which they may use for donations to the Red Cross, to help hurricane Harvey victims. Anything they do not donate, they get to keep.
Treatment prayer: subjects are asked to pray for the hurricane Harvey victims. Subjects are asked a follow-up question on if they did pray or not. They are thereafter subjected to the same donation information as subjects in treatment baseline.
Treatment think: subjects are asked to take a moment to think about the hurricane Harvey victims. Subjects are thereafter asked if they did take a moment to think about the hurricane victims or not. They are thereafter subjected to the same donation information as subjects in treatment baseline.
4. Subjects decide on their donation to the hurricane Harvey victims, via the Red Cross.
5. Subjects answer questions on if they had previously donated to hurricane Harvey victims, and religious subjects were asked how frequently they pray.
The experiment was conducted online, since, for our purposes, the context of an incentivized online experiment presented several advantages over the economic laboratory or “lab-in-the-field” context.

Design details

Randomization method
Randomization into the different treatments was done by a computer, as part of the survey tool Qualtrics.
Randomization unit
Individual. However, I needed both religious and non-religious subjects in the study, and religious subjects only can participate in "treatment pray".
Clusters
Three treatments.
Observations
375 subjects.
Arms
Subjects would be distributed over treatments as follows: treatment baseline: 75 religious subjects, 75 non-religious subjects; treatment pray: 75 religious subjects; treatment think: 75 religious subjects, 75 non-religious subjects.
Primary outcomes
Subjects donate substantially less when they have prayed for the hurricane Harvey victims, than they do if they have taken a moment to think of the victims.

Cited by

No citation data available for this registration yet.

Keywords