This paper is concerned with social desirability as a source of bias and invalidity in survey studies. Two components of social desirability-"trait desirability" and "need for siscal approval"-are examined to assess the extent to which they effect people's responses to several sociological measures (e.g., happiness, religiosity). This analysis reveals that both components of a social-desirability response set, while generally unrelated to one another, are independently related to our dependent variables, people's responses on the various measures. The more important of the two components, trait desirability, is then examined with regard to its effect on the relationship between people's sexual status and the dependent variables. This introduction to the analysis of trait desirability as a test factor tends to specify both the magnitude and the direction of the original relationships involving sexual status. Finally, the results are discussed with regard to their implications for the validity of various sociological measures and the validity of the sociological investigator's findings.
Phillips et al. (1972) studied this question.