Contact time optimization in panel surveys

In the face of declining contact and response rates and with increasing costs of conducting surveys, many survey agencies resort to implementing contact strategies. As for obtaining contact in panel surveys as early as possible, without annoying people by calling at undesired times and ultimately causing a refusal, this means to implement call time strategies other than randomization of calls to individual households. In this research, we use call data from the Swiss Household panel, a centralized CATI panel survey with a randomized (experimental) call-household assignment. Using random effects models, we analyze efficiencies of obtaining initial contact by assigning optimal times to first calls, and times and spacing to second and later calls, depending on household socio-demography and prior call patterns. We conclude by giving some recommendations that may help to making early and successful contact during fieldwork.