The Effect of Interview Length on Data Quality in the Consumer Expenditure Interview Survey

Brian T. Nix


The Consumer Expenditure Interview Survey is an ongoing panel survey of U.S. households in which sample households receive the same survey in a series of four interviews, each typically lasting about an hour. Survey length is one element thought to negatively affect survey results, by increasing respondents’ perception of burden. Because of the high burden associated with the survey, and the steady decrease in the survey’s response rates, measures are being taken to shorten the interviews. This paper examines the effect of the interview’s length on response rates, reported expenditure values, and other measures of data quality.