Journal of Official Statistics, Vol.26, No.1, 2010. pp. 165–191
Get It or Drop It? Cost-Benefit Analysis of Attempts to Interview in Household Surveys
Dmitri Romanov, Michal Nir
Abstract:We develop a cost-benefit model for streamlining allocation of field staff efforts invested in attempting to interview the designated sample units. By accounting for heterogeneous response propensity among various population groups, the costs of fieldwork, and the utility of the information gathered in a survey, the model provides a guideline for determining the optimal maximum number of attempts to interview. Limiting interview attempts lowers the surveys response rate, possibly creating a nonresponse bias a factor that is not directly reflected in a cost-benefit reckoning. We demonstrate the use of the model by simulating different limits on the number of attempts to conduct a face-to-face interview in Israels Household Expenditure Survey. Under the most stringent simulated limit three attempts to interview in the Arab sector and four attempts in the Jewish sector, in which the response rate falls from 88% to 76% we found that limiting the number of attempts to interview causes no significant bias in estimates of the main survey variables.
Keywords:Fieldwork efficiency, nonresponse bias, limit on attempts to interview, duration analysis model
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group