Optimal Timing for Random Digit Dialing

Abstract

We examine optimal time of day and day of week for conducting random digit dial (RDD) surveys in low- and middle-income countries (LMICs). Different types of survey respondents have competing time demands that influence when they are likely to be able and willing to answer the phone and complete an interview. In this brief, we consider whether there is a best time of day or day of week for improving survey response rates and sample representativeness based on RDD surveys in nine countries. We restrict our analysis to first attempt calls, which function like a randomized experiment. We find that midday calls produce a slightly higher survey completion rate on average than morning calls across the set of nine countries we studied. Evening calls have the lowest survey completion rate. For days of week, there is no evidence of a statistically significant difference in completion rates. We find some evidence that calls earlier in the week have higher contact rates than those made later in the week and that calls made in the evening have lower contact rates than those made earlier in the day. We do not find evidence that the time of day or day of week of the first attempt affects the composition of the sample. It may be meaningfully cost-effective to increase effort during certain time periods when productivity is highest.

Keywords: research methods, phone surveys, response rates