Bystanders, noise, and distractions in face-to-face surveys in Africa and Latin America

Abstract

Survey researchers are increasingly concerned that the presence of other people (bystanders) may affect data quality in structured, face-to-face survey interviews. In this article, we study bystanders using data from 15,309 face-to-face surveys about technology from Ghana, Nigeria, Uganda, Brazil and Guatemala. Our analysis (1) describes the frequency of bystander presence (spouses, parents, other family and non-family members), (2) investigates the correlates of bystander presence, and (3) studies whether bystanders affect response distributions, ‘don’t know’ responses, and satisficing behavior. We also consider the influence of other environmental factors – noise and distractions – on interviewer–respondent interaction and satisficing. The results are encouraging for the field of survey research. Bystanders did not affect response distributions, the number of ‘don’t know’ responses, or satisficing behavior. Although noise negatively affected interviewer–respondent interactions, noise did not affect data quality. Our research suggests that for less sensitive topics, bystanders have limited influence, and also that respondents and interviewers are resilient to challenges such as noisy environments.

Keywords:

• Bystander
• third party
• survey
• Africa
• Latin America
• error
• social desirability

Notes

1. A supplementary analysis (available on request) showed that non-family members were more likely to be present when interviews were done outside (vs. inside) the household.