Table of Contents

Variable

Edwards ( 1957) described the Social Desirability SD as measuring “the tendency to give socially desirable responses in self-description” ( p. 35), more specifically, an individual’s characteristic level of self-presentation without special instructions or motivation to do so ( p. 230).

Description

Edwards ( 1957) asked ten judges to rate whether “True” or “False” was the most desirable response to 79 items assembled from the K, F, and Lie scales of the MMPI. The 39 items on which the judges unanimously agreed formed the SD. Most of the items (30 of 39) are keyed negatively. As on the MMPI, respondents must answer “True” or “False,” with one point added for each response that matches the key. Hence, possible scores range from 0 to 39, higher scores indicating more socially desirable responding. Given their source, it is not surprising that item content is heavily laden with references to psychologi cal distress.

Samples

Edwards (1957) reported means of 28.6 (s.d. = 6.5) and 27. I (s.d. = 6.5) for males and females in a sample of 192 college students. Edwards and Walsh (1964) found a mean of 28.8 in a sample of I 30 paid students. More recently, Paulhus (1984) reported means of 28.4 (s.d. = 5.5) and 30.3 (s.d. = 4.9) for students in anonymous (n = 60) and public disclosure conditions (n = 40), respectively. In a sample of 503 students, Tanaka-Matsumi and Kameoka (1986) found means of 26.7 (s.d. = 6.3) and 20.9 (s.d. = 5.3) in normal and depressed samples, respectively.

Reliability

Internal Consistency

Alpha coefficients range from .83 to .87 in the samples reported above.

Test-Retest

Test-retest reliabilities of .66 (males) and .68 (females) after 2 weeks were reported by Rorer and Goldberg (1965).

Validity

Convergent

The ESD is robust in that it correlates highly with scales from a variety of content areas if they too were assembled from items with extreme desirability ratings (Edwards, 1970; Jackson & Messick, 1962). For example, the SD correlates .71 with the Desirability scale of Jackson’s (1967) PRF, which comprises items of extreme desirability chosen from diverse personality domains (Holden & Fekken, 1989).

Discriminant

Early critics alleged that, because 22 of the 39 items overlapped with Taylor’s ( 1953) anxiety scale, the SD was simply another anxiety measure. Similarly, Crowne and Mar lowe (1960) complained that the scale was intrinsically confounded with psychopathology because many of the items referred to psychological distress. Edwards and Walsh (1964) responded by showing that the pattern of correlations with other measures was unchanged when the psychopathology items were replaced.

Edwards distinguished the construct measured by the SD from tendencies to lie deliberately as measured by impression management scales (Edwards, 1957, 1970). This conceptual distinction has been clearly sustained by the data (Edwards & Walsh, 1964; Paulhus, 1984; Wiggins, 1964).

A major controversy was stirred by reports of a very high correlation between SD and the first factor of the MMPI (e.g., Jackson & Messick, 1962). To some observers this correlation suggested that the MMPI assesses SDR instead of psychopathology. Block (1965), however, argued forcefully that the SD reflects a substantive trait, namely, ego resiliency. Although most commentators agree that the SD reflects a more general disposi tion, the precise nature of the disposition remains moot (Paulhus, 1986).

Location

Because the items are taken directly from the MMPI (a copyrighted instrument), only a few sample items can be provided. However, the 39 MMPI booklet numbers are listed so that the full scale may be assembled by the reader. The MMPI is available from University of Minnesota Press, Minneapolis, Minnesota 55455.