The influence of social desirability as a response set on the measurement of values/preferences related to work

Abstract

The deleterious effects of social desirability response sets on the interpretation of psychological test scores has been a field of investigation by psychologists for some decades. However, in the specific context of measuring values/preferences related to work, most test constructors and researchers appear willing to accept Super's (1973) view that social desirability is not a major problem. Tbe empirical basis for such a statement is principally a study by Super and Mowry (1962) which is open to a number of methodological criticism. This paper reports research intended to provide more information about the relationship between social desirability and value/preference test scores. In Study 1, previous data gathered for the Work Aspect Preference Scale (Pryor, 1981a) on senior high school students was reanalysed by comparing the respective “social desirability” scales of tbe Sixteen Personality Factor Questionnaire, Eysenck Personality Inventory and the Vocational Preference Inventory with the subscales of the WAPS. The results indicated that only five correlations out of a possible 39 were significant and that none of these was very high. In Study 2, a sample of 279 senior high school students were given both the WAPS and the shortened Marlowe-Crowne Social Desirability Scale. It was found that the Shortened Marlowe-Croww SD scale had adequate split-half reliability for research purposes, although the coefficient for females was marginal. The results indicated that the subscales of the WAPS have only two significant correlations with the social desirability total score for the whole sample. When analysed by sex, 3 significant correlations were found between WAPS subscale scores and the social desirability total score for marks. For females, no significant correlations were found. Overall the results suggested that Altruism scores may be higher due to social desirability and that Independence scores may be lower. It was concluded that both studies confirmed that the measurement of values/preferences related to work is not substantially affected by social desirability response sets.