Does the switch to a mixed-mode design increase panel attrition?

Presenter: Alessandra Gaia, Università degli studi di Milano-Bicocca, Italy

Author: Alessandra Gaia

I evaluate the effect of a switch to a mixed-modes design in a longitudinal survey on panel attrition. I use experimental data from the Innovation Panel (IP) of Understanding Society. At IP5 the sample was randomly allocated to two experimental groups: one assigned to a unimode face-to-face design and another to a mixed-mode design (web with a face-to-face follow up). I use a logistic regression framework to model the effect of the experimental allocation on attrition at IP6. The treatment (mixed-modes allocation) is interacted with sample members’ characteristics to target the effect of the mixed-modes allocation on different types of sample members. Finally, the samples obtained with the two different designs are compared both across them and with the sample at the fourth wave.

I do not find evidence that a mixed-modes design increases attrition. Adults from the original sample that did not participate in the 4th Wave are less likely to attrite if they are approached with a mixed-mode design at the 5th Wave. The finding that a mixed-mode design decreases attrition for previous wave’s non respondents is particularly promising since this group is at higher risk of attrition. No effect is found for the entire sample, or for the original sample respondents or for the refreshment sample. Results are discussed within the theoretical framework of the leverage-saliency theory. This study constitutes a new contribution as the evidence on the effect of mixed-modes on panel attrition is scarce. Moreover, the research may contribute to the decisions on the mode of data collection to be adopted in other panel surveys.