Mixed-mode Surveys: Design, Estimation, and Adjustment Methods

Mixed-mode survey inference literature has proposed a variety of methods to control and adjust for possibly substantial and differential survey errors due to survey mode. This chapter focuses on estimation and adjustment methods, these phases of the survey lifecycle cannot be isolated from earlier design decisions. The International Social Survey Programme (ISSP) 2011 surveys include modules on evaluation of the health-care system in the country, personal health, and health insurance. The chapter illustrates how the statistical methods that adjust for measurement error work on the 2011 ISSP data. In an attempt to unconfound the mode selection effects and the mode measurement effects, it considers propensity score matching. The demographic and socioeconomic characteristic comparisons suggest that the differences in proportions of persons reporting excellent or very good health status could be due to mode selection effects. Imputation methods can potentially account for selection effects in mode effect adjustments in statistical inference methods.