ABSTRACT
This chapter discusses two type of nonresponse: unit nonresponse, in which the entire observation unit is missing, and item nonresponse, in which some measurements are present for the observation unit but at least one item is missing. The difference in voting rate between the nonrespondents and the selected sample was largest in the younger age groups. Among the nonrespondents, the voting rate varied with the type of nonresponse. A. D. Biderman and D. Cantor find lower victimization rates for persons who respond in three consecutive interviews than for persons who are nonrespondents in at least one of those interviews or who move before they complete the panel. Analysis of callback data can provide some information about the biases that can be expected from the remaining nonrespondents. If nonrespondents resemble late respondents, one might speculate that nonre-spondents were more likely to favor Reagan. Most surveys have some residual nonresponse even after careful design and follow-up of nonrespondents.
