Non-Sampling Error and Data Quality: What Can We Learn from Surveys to Collect Data for Vulnerability Measurements?

This paper investigates the causes for non-response and measurement errors in household panel surveys designed for assessing vulnerability to poverty in Thailand and Vietnam. Using data from surveys conducted in 2007 and 2008 we show that interview environment, timing, interviewer, and some respondent characteristics significantly affect non-sampling errors. Investigating interviewer bias for household consumption we find no significant effect of interviewer variables, which suggests validity of the data collected. The paper maps out possibilities to reduce non-sampling errors such as observing suitable interview duration and timing and matching interviewer characteristics with those of respondents.