After 1997, the Panel Study of Income Dynamics (PSID) will move to biennial data collection, with the next scheduled interviews to be conducted in 1999. This move will require questionnaire design changes so as to gain retrospective reports of events occurring up to two and a half years prior to the interview date. This project will directly test the data quality obtained by questionnaire design methodologies that require the use of scripted questions versus a flexible event history calendar (EHC-F). The EHC-F methodology utilizes sequential and parallel autobiographical memory cues that are unavailable to scripted questionnaires, as such cues must be customized to meet the idiosyncratic needs of respondents. As a secondary objective, the study also assesses the data quality of gaining reports from a scripted questionnaire list (Q-list) design versus a scripted questionnaire that includes an event history calendar component as an interviewing aide (EHC-S). A total of 600 PSID respondents who were interviewed in 1997 to obtain economic and demographic measures for 1996 will be randomly assigned to conditions. They will be interviewed in 1998 (an off-year for the full PSID panel) about events that occurred in both 1996 and 1997. The economic and demographic outcome measures will include reports on weeks worked, hours per week worked, weeks unemployed, annual salary, receipt of food stamps and AFDC, and changes in occupation, residence, and household composition. The consistency between the 1997 and 1998 reports on 1996 events will serve as the primary indicator of data quality. By providing the first direct comparative test of the quality of responses obtained from a scripted Q-list, a scripted EHC, and a flexible EHC, this project will be able to determine the relative utility of such methodologies in promoting the quality of retrospective reports obtained on events occurring up to two and a half years prior to the interview date.
