In 1993, the Census Bureau undertook a qualitative research project to investigate respondents' comprehension and interpretation of a subset of questions in the American Housing Survey (AHS). These items form a "Moderate Physical Problems" scale, used by housing analysts as a key indicator of the quality of housing stock.
The research on the Moderate Physical Problems (MPP) series used several different approaches to identify potential questionnaire problems: an "expert panel," telephone and field interview observations, analysis of AHS item nonresponse data, and a critical review of the questionnaire.
Both complementing these initial investigations, and following up on the hypotheses they generated, was the primary research activity -- a series of cognitive laboratory interviews to explore respondents' comprehension and interpretation of the MPP questions. Specific techniques used included think-aloud protocols, paraphrasings, and respondent debriefings.
This paper first briefly describes the impetus for the research and the methods used, focusing in particular on cognitive interviewing techniques. It then presents some of the important substantive findings highlighting problems with the AHS instrument and our proposed solutions. Finally, the paper discusses the methodological advantages and disadvantages of using cognitive laboratory interviews in questionnaire redesign.