As part of the Current Population Survey's (CPS) redesign efforts, two alternative questionnaires were tested with the current version of CPS in a CATI/RDD test between July 1990 and January 1991. The primary objective of the test was to identify problem areas in question wording/sequences in order to finalize development of a revised CPS questionnaire that will reduce sources of response error. Various methodological techniques were utilized to evaluate the alternative question wordings. These included coding of interviewer/respondent behavior, respondent debriefing studies, vignette analysis and interviewer debriefing studies. Additionally, item nonresponse rates by question version and differences in response distributions were evaluated. This paper discusses the results of the CATI/RDD test leading to the development of the proposed revised questionnaire, the usefulness of each of the techniques mentioned above in contributing to the alternative wording of the proposed questionnaire, and the plans for future questionnaire testing.
