Alternative Approaches To Multilevel Modelling Of Survey Noncontact And Refusal

We review three alternative approaches to modelling survey noncontact and refusal: multinomial, sequential and sample selection (bivariate probit) models. We then propose a multilevel extension of the sample selection model to allow for both interviewer effects and dependency between noncontact and refusal rates at the household and interviewer level. All methods are applied and compared in an analysis of household nonresponse in the UK, using a dataset with unusually rich information on both respondents and nonrespondents from six major surveys. After controlling for household characteristics, there is little evidence of residual correlation between the unobserved characteristics affecting noncontact and refusal propensities at either the household or the interviewer level. We also find that the estimated coefficients of the multinomial and sequential models are surprisingly similar, which further investigation via a simulation study suggests is due to there being little overlap between the predictors of noncontact and refusal. Alternative Approaches to Multilevel Modelling of Survey Noncontact and Refusal Fiona Steele and Gabriele B. Durrant Centre for Multilevel Modelling, University of Bristol Southampton Statistical Sciences Research Institute, University of Southampton Summary. We review three alternative approaches to modelling survey noncontact and refusal: multinomial, sequential and sample selection (bivariate probit) models. We then propose a multilevel extension of the sample selection model to allow for both interviewer effects and dependency between noncontact and refusal rates at the household and interviewer level. All methods are applied and compared in an analysis of household nonresponse in the UK, using a dataset with unusually rich information on both respondents and nonrespondents from six major surveys. After controlling for household characteristics, there is little evidence of residual correlation between the unobserved characteristics affecting noncontact and refusal propensities at either the household or the interviewer level. We also find that the estimated coefficients of the multinomial and sequential models are surprisingly similar, which further investigation via a simulation study suggests is due to there being little overlap between the predictors of noncontact and refusal.