In cross-sectional surveys, nonresponse is categorized in two ways: unit (total) nonresponse and item nonresponse. Unit nonresponse occurs when no data are available on the sampled unit, while item nonresponse occurs when a responding unit fails to provide data for one or more items in the questionnaire. These two types of nonresponse differ in the availability of data for nonresponse compensation and usually in the approaches used for missing data compensation.

For unit nonrespondents, no data are available other than from observation of the nonresponding unit and the sampling frame itself. Compensation for unit nonresponse is usually handled through weighting adjustments, a procedure in which the weights of appropriate respondent records are increased to compensate for nonrespondents. For item nonresponse, much more data are available about the respondent, namely the responses to other items in the questionnaire. As a consequence, some form of imputation, a procedure in which values are assigned for the missing responses, is usually implemented.

In longitudinal surveys, a third type of nonresponse exists--wave nonresponse. Wave nonresponse occurs when a sample unit does not respond in one or more interviews of a longitudinal survey but has provided data for at least one interview. In this situation, considerably more data are missing compared to the item nonresponse situation; however, considerably more data are available for use in nonresponse compensation strategies. This category of nonresponse falls between unit and item nonresponse as far as the availability of data for nonresponse compensation strategies. This paper proposes two simple, inexpensive and easy to implement approaches to handle single wave nonresponse in the Survey of Income and Program Participation (SIPP) for use in longitudinal estimation.

Section II of this paper presents background on the SIPP design and Section III includes a discussion of nonresponse types and their treatment in the SIPP. A general discussion of Wave nonresponse compensation strategies is presented in Section IV. Sections V and VI present imputation and multiple weighting strategies respectively to compensate for a missing interview in the SIPP. A comparison of selected options for handling a missing wave is presented in Section VII. Section VIII presents conclusions and suggestions for implementation and further research.