Social Desirability in Measures of Subjective Well-being: a Systematic Evaluation

Abstract

The present investigation addressed the problem of a social desirability response bias in measures of psychological well-being. Data on 150 people, between the ages of 50 and 82, yielded high correlations between three measures of well-being (the MUNSH, the LSI-Z, and the PGC) and the Edwards Social Desirability Scale, but only moderate ones between well-being scales and the Marlowe-Crowne Social Desirability Scale. Part correlations between well-being measures and an external criterion of happiness, controlling for social desirability, failed to improve on the zero-order criterion/well-being relationship. Controlling for social desirability, therefore, does not enhance the construct validity of well-being scales. These and related results suggest that the high zero-order correlations between measures of well-being and the Edwards scale are more readily attributed to content similarity between the Edwards scale and measures of well-being than to a social desirability response bias in well-being measures.

Happiness in the Face of Adversity: Reformulating the Dynamic and Modular Bases of Subjective Well-Being

“…Further analyses, however, revealed that social desirability does not necessarily distort, but rather amplifies, the meaning of psychological constructs (Furnham, 1986;McCrae & Costa, 1983). Thus, controlling for social desirability does not enhance the validity of SWB measures and may even attenuate it (Diener, Sandvik, Pavot, & Gallagher, 1991;Kozma & Stones, 1987). In line with the view that contextual artifacts for attitudes may actually be inherent in attitude formation (Tourangeau & Rasinski, 1988), it is concluded, then, that self-report measures of SWB constitute a vehicle of social declarations that are part and parcel of SWB at large.…”

Rethinking Social Desirability Scales

“…The role of IM scales as suppressors or moderators of the correlation between self and informant reports have been put to test several times over the years by using different scales and types of self-informant relationships. All these tests have failed to find statistical effects that would support the interpretation of IM scales as reliable validity scales (Borkenau & Ostendorf, 1992;Borkenau & Zaltauskas, 2009;Diener, Sandvik, Pavot, & Gallagher, 1991;Hunsley, Vito, Pinsent, James, & Lefebvre, 1996;Kozma & Stones, 1987;Pauls & Stemmler, 2003;Piedmont, McCrae, Riemann, & Angleitner, 2000).…”

Measures of self-perceived well-being

“…A second-order factor representing general well-being has also been obtained [60]. Correlations with the LSIA range from 0.55 to 0.75 [59,64] and 0.6 with the negative score on the Bradburn scale [33].…”