The confounding of selection and measurement effects between different modes is a disadvantage of mixed-mode surveys. Solutions to this problem have been suggested in several studies. Most use the back-door method, which includes covariates explaining selection effects. Unfortunately, these covariates must meet strong assumptions, which are generally ignored. In this presentation, I will discuss these assumptions in greater detail and also provide alternative methods for solving the problem. One alternative method is the front-door method, which includes covariates explaining measurement effects instead of selection effects. An other alternative method is the instrumental variable method, which requires a covariate that randomly divides the sample in two groups where data of one group is entirely collected by one single mode. The application of the back-door, front-door, and instrumental variable methods is illustrated by using data from a survey on opinions about surveys, which yields mode effects in line with expectations for the front-door and instrumental variable methods, and mode effects contrary to expectations for the back-door method. However, the validity of these results depends entirely on the (ad hoc) covariates chosen. Research into better back-door covariates, front-door covariates, and instrumental variables might thus be a topic for future studies.
Presented by:
Jorre Vannieuwenhuyze (K. U. Leuven)
Date & time:
April 15, 2013 3:00 pm
External seminars home