Testing for Fatalism in Responses to HIV Risk Information

How to cite

Kerwin, Jason. 2020. "Testing for Fatalism in Responses to HIV Risk Information." AEA RCT Registry. AEARCTR-0006666. https://doi.org/10.1257/rct.6666-1.1

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

Jason Kerwin*

* Primary investigator

Registry ID
AEARCTR-0006666
DOI
https://doi.org/10.1257/rct.6666-1.1
First registered
Last updated
Status
Completed
Trial dates
2012-08-14 – 2012-12-14
Countries
Malawi (Zomba District)
JEL codes
D81, D83, I12
Registry record
http://www.socialscienceregistry.org/trials/6666

Abstract

I study whether Malawians' responses to HIV risks are "fatalistic", meaning that they respond to increased risks by having more risky sex rather than less. I randomly provide information about the true risk of HIV transmission, which is lower than people think it is, to a sample of sexually-active adults in the Zomba district of southern Malawi. This is an area where previous ethnographic and correlational research has documented behavior that appears fatalistic. I then measure how people change their sexual behavior in response to the information treatment, looking for evidence of a U-shaped relationship between risk beliefs and risky sex.

Experimental design

I draw a random sample of 70 villages from one sub-district of Zomba district, stratified by distance to the closest trading center, and assign half to the treatment group and half to the control group. Within each village I sample 30 sexually-active adults, stratified by gender. I conduct a baseline survey to measure beliefs about HIV risks. At the end of the baseline survey, the treatment group gets an additional survey module presenting information about the true risk of HIV transmission from unprotected sex with an infected partner. 1-4 months after the baseline survey, I return for a followup survey of all respondents to measure changes in risk beliefs and sexual behavior.

Design details

Randomization method
Randomization done in office by a computer
Randomization unit
Village
Clusters
70 villages
Observations
2100
Arms
35 villages treatment, 35 villages control
Primary outcomes
Total sex acts in the past week, as measured on a retrospective sex "diary" in which interviewers walk subjects back through the past 7 days to record all sexual activity.

Cited by

No citation data available for this registration yet.

Keywords