Assessing response fatigue in phone surveys: Experimental evidence on dietary diversity in Ethiopia

The COVID-19 pandemic has spurred interest in the use of remote data collection techniques, including phone surveys, in developing country contexts. This interest has sparked new methodological work focusing on the advantages and disadvantages of different forms of remote data collection, the use of incentives to increase response rates and how to address sample representativeness. By contrast, attention given to associated response fatigue and its implications remains limited. To assess this, we designed and implemented an experiment that randomized the placement of a survey module on women’s dietary diversity in the survey instrument. We also examine potential differential vulnerabilities to fatigue across food groups and respondents. We find that delaying the timing of mothers’ food consumption module by 15 minutes leads to 8-17 percent decrease in the dietary diversity score and a 28 percent decrease in the number of mothers who consumed a minimum of four dietary groups. This is driven by underreporting of infrequently consumed foods; the experimentally induced delay in the timing of mothers’ food consumption module led to a 40 and 11 percent decrease in the reporting of consumption of animal source foods, and fruits and vegetables, respectively. Our results are robust to changes in model specification and pass falsification tests. Responses by older and less educated mothers and those from larger households are more vulnerable to measurement error due to fatigue.