Response Bias in Surveys of Mental Health: An Empirical Investigation

Abstract

A substantial body of literature indicates that (1) the tendency to yeasay or naysay, (2) the perception of the desirability or undesirability of a given trai, and (3) the need for social approval have a significant and independent impact on the statements made by respondents in surveys. Most investigatiors concerned with response bias appear to believe that these response bias variables act as a form of systematic bias which significantly distorts the relationship observed between an independent and a dependent variable. However, while the literature indicates such a distortion may occur, this has not been demonstrated. Drawing on the data from a national survey, we look at the effect of these three forms of response bias on the pattern of relationships among seven demographic variables (sex, race, education, income, age, marital status, and occupation) and three very different indicators of mental health (psychiatric symptons, self-esteem, and feelings of positive affect). The results lead to an almost unequivocal conclusion: the response bias variables have very little impact on the relationship, and it seems safe to conclude that these sources of response bias do not act in a systematic way to invalidate the pattern of observed relationships between mental health and common demographic variables.