Enumerator Education and Survey Responses: An Experimental Analysis

How to cite

Doces, John. 2022. "Enumerator Education and Survey Responses: An Experimental Analysis." AEA RCT Registry. AEARCTR-0008798. https://doi.org/10.1257/rct.8798-1.0

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

John Doces*

* Primary investigator

Registry ID
AEARCTR-0008798
DOI
https://doi.org/10.1257/rct.8798-1.0
First registered
Last updated
Status
In development
Trial dates
2022-01-12 – 2022-03-01
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/8798

Abstract

Based on my prior research in West Africa, I have found variation in the data by enumerator education. The prior data shows a difference by enumerator education. Three of my male enumerators with various levels of education (advanced degree, university degree, elementary education) exhibit different responses to questions about wellbeing/happiness and identity. This paper will test several hypotheses based on the preexisting data to see if there is an effect. Each respondent will collect data as they have done on prior projects asking the same questions. They will reveal nothing explicitly about their education. Based on prior data the hypotheses are as follows: 1. Respondents will report lower personal level of wellbeing to highly educated enumerator compared to enumerator with lower level of education; 2. Respondents will rate enumerator’s level of wellbeing as higher when enumerator has high level of education compared to low level of education; 3. Respondent is more likely to report good things about the government (i.e., less corruption and more support for the current president) to the highly educated enumerator compared to the less educated enumerator. Results of this analysis carry important implications as major surveys of life in Africa and beyond use only highly educated enumerators. Doing so could be producing serious measurement error.

Experimental design

All enumerators will sample in the same neighborhood of Abidjan. Each will follow a random walk and will use a coin flip to determine if they should ask a person to participate or not. Treatment will be randomly assigned by the random selection process.

Design details

Randomization method
Coin flip in the field.
Randomization unit
Unit of randomization will be the individual within the same cluster.
Clusters
There will be one cluster.
Observations
There will be a total of 150 observations, 50 per enumerator.
Arms
50 observations by enumerator. The sample is not pre-selected.
Primary outcomes
There will be four outcomes: respondent wellbeing; respondent evaluation of enumerator's wellbeing; degree of corruption in the police force; support for current president.

Cited by

No citation data available for this registration yet.

Keywords