Measuring Norms and Enumerator Effects: Survey Method Matters
How to cite
Álvarez-Aragón, Pablo; Champeux, Hugues. 2023. "Measuring Norms and Enumerator Effects: Survey Method Matters." AEA RCT Registry. AEARCTR-0011696. https://doi.org/10.1257/rct.11696-1.2
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
While the process of data collection can lead to bias, little empirical evidence investigates the role of the survey method. In this paper, we compare two survey methods: the standard face-to-face interview and an alternative method we call Human-Assisted Self-Administered survey (HASA). In the latter, respondents are guided by an enumerator reading questions, but they answer privately on an electronic device. Taking advantage of an RCT in Benin, we randomize the survey method across respondents. We show that the survey method leads to different results depending on enumerator influence. Identifying this influence, we document that variables that are likely to be influenced by enumerators differ systematically across survey methods. Interestingly, these variables are mainly related to gender norms and women agency. We find that respondents who answer directly on a tablet report less gender-equal values. Investigating the mechanisms, we show that social desirability bias affects responses in face-to-face interviews.
Intervention
We take advantage of an RCT in Benin in which respondents are randomly assigned to receive a business training. Participants are grouped together to receive the training. In this context, we administer a survey at the beginning of the session and compare the results with those obtained in face-to-face interviews.
Experimental design
Women invited to the training program are randomly selected from a sample of pineapple producing households who applied for a subsidy to start or expand a pineapple production (these households typically include a man producing pineapple and in half of the cases a woman producing as well). A control group is selected among the initial pool of applicants.
Data collection occured in two waves: baseline in 2020 and endline in 2022.
Design details
Cited by
No citation data available for this registration yet.