Cognitive Testing of the Census 2010 Experimental Overcount Questions for the Census 2010 Alternative Questionnaire Experiment: Final Project Report

摘要

This report documents the development and cognitive testing in 2008-9 of an alternative mailout census booklet with special coverage questions to compare to the standard census form in terms of coverage in the Census 2010 Alternative Questionnaire Experiment split-panel test. Both forms include a question asking whether each person in the household sometimes lives or stays somewhere else, and for what reason. On the standard census form, this question functions as a flag for later phone followup to get more complete coverage data. The alternative mailout booklet converts this question into a screener for a new set of questions on the mailout form itself to identify a persons alternative address, usual residence, and Census Day location, making it possible to determine residence from answers on the census form itself during processing with reduced need for telephone followup. If it works, the alternative approach has the potential to improve coverage as well as cut the costs and time involved in conducting followup operations. The report describes the objectives and methods of the testing and presents findings in three main sections: cognitive testing results, alternative wording for the Census Day location question, and how answers to the screener question affect answers to the experimental address, usual residence, and Census Day location questions. The testing indicated that the questionnaire works well and should go forward into the split-panel test. Recommendations are made for making the skip instructions more noticeable, for rewording some questions, and for emphasizing key words in address question 8 and coverage question 10. Final changes to key questions in the questionnaire are shown in the appendix.