Household Income Data Reports in Mail Surveys Robert A. Peterson, University ofTexas Roger A. Kerin, Southern Methodist University zyxwvutsrqponmlkjihgfedcbaZYX An empirical study was conducted to assess the quality of reported household income data in the context of a mail survey, where quality was measured in terms of the difference between actual household income and reported house- hold income. In the study, nonresponse error w as found to be greater than re- sponse error (distortions in reporting household income), although both were statistically significant. Under- and overreporting of household income w ere found to be associated with absolute income levels, and several possible rea- sons for under- and overreporting were hypothesized. The quality of data derived from mail surveys is determined, in general, by three factors-response rate, response completeness, and response distortion. Response rate-the number of mail questionnaires returned to the researcher(sthas long been a subject of research interest. A high response rate improves data quality in that it simultaneously increases reliability and promotes respondent representativeness. Conversely, a low response rate typically decreases data quality in that it culminates in “nonresponse error.” While a plethora of literature exists on mail survey response rates, unfortunately most of it concentrates only on the reliability issue-how various pre- and postcompletion response in- ducement techniques influence the number of respondents (e.g. [7, 11, 14, 17, 181). Completeness of response refers to the tendency of respondents to not answer all survey questions; in a lay sense this may be considered an “error of omission”; response distortion, on the other hand, may be termed an “error of commission”; it represents a bias or distortion in responses-a discrepancy between the actual and the reported values of a survey item. Failure to obtain completed questionnaires or unbiased responses has frequently been shown to reduce the reliability of estimates drawn from data collected by means of mail surveys [e.g., 5, 12, 131. Address correspondence to: Robert A. Peterson, Institute for Constructive Cap- italism, University of Texas, Austin, TX 78712. JOURNAL OF BUSINESS RESEARCH 301 0 Elsevier North Holland, Inc., 1980 52 Vanderbilt Ave., New York, NY 10017 0148-2963/80/03301-13$01.75