Crowne and Marlowe Social Desirability Scale

Author: Dr Simon Moss

Overview

Some individuals, while completing questionnaires that assess their personality, temperament, and demeanor, like to portray themselves positively. They tend to exaggerate or inflate their strengths and achievements, and often deny or trivialize their deficiencies and failures. That is, they attempt to depict themselves as similar to the norms and standards of their society and community-called a social desirability bias. Crowne and Marlowe (1960) developed a seminal scale to identify the extent to which individuals exhibit this bias.

Development of this scale

In 1960, Crowne and Marlowe attempted to identify a set of behaviors that are perceived by society to be exemplary, but enacted only infrequently. They attempted to extract these behaviors from extant personality inventories. This process generated 50 items. Examples include "I like to gossip at times" or "I am always willing to admit when I make a mistake".

A set of 10 judges then determined whether they perceive each of these 50 behaviors as desirable or undesirable. Crowne and Marlowe revealed that 47 of the 50 behaviors generated at least 90% agreement, and only these items were retained.

In addition, 76 undergraduate students were asked whether or not they engage in the 47 behaviors. Only 33 of these items were significantly related to the aggregated total. That is, 33 of the items discriminated between individuals who do and do not tend to exhibit this bias towards social desirability. Crowne and Marlowe (1960) showed the internal consistency of the 33 items is .88, and the test-retest correlation was .89.

Validation of this scale

Other researchers have also examined the psychometric properties of this scale. Several studies have shown that internal consistency is adequate& Nordholm (1974) generated an coefficient of .73, Crino, Svoboda, Rubenfeld, and White (1983) generated coefficients that ranged from .70 - .77, Tanaka-Matsumi and Kameoka (1986) generated a coefficient of .79, and Holden and Fekken (1989) generated a coefficient of .78;; all these studies subjected the scale to students, however.

O'Grady (1988) examined whether or not these psychometric properties differ between males and females. Internal consistency was similar in males and females, approaching .71 and .72 respectively. Indeed, as O'Grady (1988) showed, mean levels of social desirability bias also does not differ significantly between the sexes-a pattern that has been confirmed in other studies (Loo & Loewen, 2004;; Loo & Thorpe, 2000)

Test-retest reliability has also been established. According to Crino, Svoboda, Rubenfeld, and White (1983), for example, test-retest reliability over a period of a month was .86.

Several studies indicate the scale comprises only one dimension. Some studies, for example, have investigated whether or not the negative items, emphasizing undesirable acts such as "I like to gossip at times", correspond to the same factor as the positive items, emphasizing desirable acts such as "I always try to practice what I preach". Greenwald and Clausen (1970) showed the correlation between these two sets of items was very high, approximately .84. Crino, Svoboda, Rubenfeld, and White (1983) generated a correlation of approximately .87.

Ramanaiah, Schill, and Leung (1977) initially challenged this conclusion, maintaining the possibility of two distinct factors: attribution of success and denial of failure. However, several of the items do seem to correspond to both factors. Indeed, Ramanaiah and Martin (1980) ultimately dismissed the distinction between these two subscales. They highlighted this distinction between attribution of success and denial of failure diminishes if both scales are adjusted to ensure an equal number of reverse-scored items.

Paulhus (1984, 1991) proposed that social desirability biases comprise two factors: self deception and impression management. Self-deception usually reflects a genuine oblivion towards subtle and subjective personal limitations. In contrast, impression management reflects the deliberate attempt of individuals to portray themselves positively to someone else. Because the scale developed by Crowne and Marlowe (1960) primarily encompasses overt and tangible acts, elevated scores probably represent impression management not self deception.

The scale is not appreciably related to maladjustment or psychopathology-one of the explicit objectives of Crowne and Marlowe (1960). Tanaka-Matsumi and Kameoka (1986), for example, showed that correlations between this index of social desirability bias and various measures of depression or anxiety ranged from -.19 to -.32.

To examine whether the 33 items are sensitive to social desirability bias, Crino, Rubenfeld, and Willoughby (1985) applied an illuminating approach, called the randomized response technique. To illustrate, some of the participants are asked to roll a dice before answering each question. If they roll a 5, they must claim they engage in the behavior and thus answer true. If they roll a 6, they must claim they do not engage in the behavior and thus answer false. If they roll 1 to 4, they answer as they would otherwise. The researcher is not apprised of the numbers they roll.

This technique is intended to curb intentional biases in responses. That is, because the researcher is not certain of whether or not the answer reflects the person, participants might be less inclined to distort their responses and portray themselves positively.

In a study conducted by Crino et al. (1985), only 144 of the 378 participants were instructed to apply this technique when they completed the questionnaire. Five of the items were more likely to be endorsed by individuals who rolled the dice, indicating these questions were considered undesirable and usually denied. Nine of the items were less likely to be endorsed by individuals who rolled the dice, indicating these questions were considered desirable and usually exaggerated. Only of these conclusions, however, departed from the evaluations of desirability conducted by Crowne and Marlowe (1960).

The extent to which some items correlate with the aggregate or total does depend on gender, as shown by Goldfried (1964). In one study, in which the sample was 100 students, only 17 of the items correlated significantly with the aggregate in males and only 15 of the items correlated significantly with the aggregate in females. Only eight of the 33 items correlated with the aggregate in both males and females.

The perceived desirability of these items is also contentious. Goldfried (1964), instructed another sample of 68 students to evaluate the extent to which they perceive the behaviors as desirable or undesirable. Twelve of these items generated perceptions that did align closely with the original study by Crowne and Marlowe (1960).

This issue has been further explored by Ballard, Crino, and Rubenfeld (1988). In their studies, participants were asked to evaluate the perceived desirability of the items. Only nine of the items were regularly evaluated as either desirable or undesirable-and also were consistent with the ratings reported by Crowne and Marlowe (1960).

Shortened versions of this scale

Strahan and Gerbasi (1972) subjected the original scale to a sample of 361 students. A principle components analysis was then conducted, and two factors were extracted: called X1 and X2, each with 10 items. The combination of these 20 items is often called XX.

Four studies were undertaken to establish the psychometric properties of this scale. Across the four studies, the internal consistency of these three sets ranges from .28 to .54.

Other studies have generated mixed estimates of internal consistency (e.g., Thompson & Phua, 2005), ranging from .73, .64 and .82 for X1, X2, and XX respectively (Fraboni & Cooper, 1989) to 16, .21 and .36 (O'Gorman, 1974).

Thompson and Phua (2005) highlighted that such scales, and indeed any derivations of these social desirability scales, might not be as applicable to managerial samples. Most psychometric have derived their samples from the student population. Nevertheless, Fraboni and Cooper (1989) did recruit a sample from the general population and uncovered promising indicators of reliability.

Reynolds (1982) subjected the original scale to 608 students and also conducted a principle components analysis as well as examined the correlations between each item and the aggregate. These analysis identify three factors, called A, B, and C, which comprise 11, 12 and 13 items respectively.

The internal consistency of these three factors was shown to be .74, .75 and .76 respectively (Reynolds, 1982). Other studies have demonstrated favorable, but variable, levels of internal consistency, ranging from .64, .66 and .68 respectively (Barger, 2002) to .86, .88 and .89 respectively (Fischer & Fick, 1993).

Ballard (1992) also developed three short scales from a principal components analysis that was applied to the responses of 399 students. The first scale comprises 11 items-derived from questions in which factor loadings exceeded .40. The second scale was identical, except another item, in which the factor loading was .399, was added. Finally, the third scale was identical to the second scale, except another item, in which the factor loading was .39, was included.

The internal consistency of these three subscales is modest, between .60 and .65 for all three scales (see Loo & Loewen, 2004;; Loo & Thorpe, 2000).

Nine of the items were also included in both of the previous shortened versions. In addition, four items are included in two of the three sets of shortened versions. The combination of these 13 items is sometimes called the composite subscale, as defined by Ballard (1992).

Fischer and Fick (1993) undertook confirmatory factor analysis to ascertain whether the X1, X2, XX, A, B, and C versions correspond to a single dimension. They also deleted items to improve measures of fit. Overall, this article implies that X1 or X2 might be preferable. Nevertheless, Fischer and Fick (1993) did generate indices of internal consistency on the various subscales that exceed the estimate of other authors-and, hence, their study might not be applicable to all contexts. Loo and Thorpe (2000) showed that indices of fit are more compelling for most of the shortened versions, relative to the full version, perhaps as an artifact of scale length (for a discussion, see Barger, 2002).

Generally, fit is adequate for most of the subscales (for more details, see Leite & Beretvas, 2005;; Loo & Loewen 2004). No obvious patterns have been discerned to indicate which of these subscales is most appropriate.

References

Ballard, R. (1992). Short forms of the Marlowe-Crowne Social Desirability Scale. Psychological Reports, 71, 1155-1160.

Ballard, R., Crino, M. D., & Rubenfeld, S. (1988). Social desirability response bias and the Marlowe-Crowne Social Desirability Scale. Psychological Reports, 63, 227-237.

Barger, S. D. (2002). The Marlowe-Crowne affair: Short forms, psychometric structure and social desirability. Journal of Personality Assessment, 79, 286-305.

Beretvas, S. N., Meyers, J. L., & Leite, W. L. (2002). A reliability generalization study of the Marlowe-Crowne Social Desirability Scale. Educational and Psychological Measurement, 62, 570-589.

Crino, M. D., Rubenfeld, S., & Willoughby, F. W. (1985). The random response technique as an indicator of questionnaire item social desirability/personal sensitivity. Educational and Psychological Measurement, 45, 453-468.

Crino, M. D., Svoboda, M., Rubenfeld, S., & White, M. C. (1983). Data on the Marlowe-Crowne and Edwards social desirability scales. Psychological Reports, 53, 963-968.

Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354.

Fischer, D. G., & Fick, C. (1993). Measuring social desirability: Short forms of the Marlowe-Crowne Social Desirability Scale. Educational and Psychological Measurement, 53, 417-424.

Fraboni, M., & Cooper, D. (1989). Further validation of three short forms of the Marlowe-Crowne Scale of Social Desirability. Psychological Reports, 65, 595-600.

Goldfried, M. R. (1964). A cross-validation of the Marlowe-Crowne Social Desirability Scale items. Journal of Social Psychology, 64, 137-145.

Greenwald, H. J., & Clausen, J. D. (1970). Test of relationship between yeasaying and social desirability. Psychological Reports, 27, 139-141.

Holden, R. R., & Fekken, G. C. (1989). Three common social desirability scales: Friends, acquaintances, or strangers? Journal of Research in Personality, 23, 180-191.

Leite, W. L., & Beretvas, S. N. (2005). Validation of scores on the Marlowe-Crowne

Social Desirability Scale and the Balanced Inventory of Desirable Responding. Educational and Psychological Measurement, 65, 140-154.

Loo, R., & Loewen, P. (2004). Confirmatory factor analyses of scores from full and short versions of the Marlowe-Crowne Social Desirability Scale. Journal of Applied Social Psychology, 34, 2343-2352.

Loo, R., & Thorpe, K. (2000). Confirmatory factor analyses of the full and short versions of the Marlowe-Crowne Social Desirability Scale. Journal of Social Psychology, 140, 628-635.

Nordholm, L. A. (1974). A note on the reliability and validity of the Marlowe-Crowne scale of social desirability. Journal of Social Psychology, 93, 139-140.

O'Gorman, J. G. (1974). Limits to the generality of the Marlowe-Crowne measure of social desirability. Journal of Clinical Psychology, 30, 81.

O'Grady, K. E. (1988). The Marlowe-Crowne and Edwards Social Desirability Scales: A psychometric perspective. Multivariate Behavioral Research, 23, 87-101.

Paulhus, D. L. (1984). Two-component models of socially desirable responding. Journal of Personality and Social Psychology, 46, 598-609.

Paulhus, D. L. (1991). Measurement and control of response bias. In J. P. Robinson, P. R. Shaver & L. S. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 17-59). San Diego, CA: Academic Press, Inc.

Ramanaiah, N. V., & Martin, H. J. (1980). On the two-dimensional nature of the Marlowe-Crowne Social Desirability Scale. Journal of Personality Assessment, 44, 507-514.

Ramanaiah, N. V., Schill, T., & Leung, L. S. (1977). A test of the hypothesis about the two-dimensional nature of the Marlowe-Crowne Social Desirability scale. Journal of Research in Personality, 11, 251-259.

Ray, J. J. (1984). The reliability of short social desirability scales. Journal of Social Psychology, 123, 133-134.

Reynolds, W. M. (1982). Development of reliable and valid short forms of the Marlowe-Crowne Social Desirability Scale. Journal of Clinical Psychology, 38, 119-125.

Strahan, R., & Gerbasi, K. C. (1972). Short, homogeneous versions of the Marlowe-Crowne Social Desirability Scale. Journal of Clinical Psychology, 28, 191-193.

Tanaka-Matsumi, J., & Kameoka, V. A. (1986). Reliabilities and concurrent validities of popular self-report measures of depression, anxiety, and social desirability. Journal of Consulting and Clinical Psychology, 54, 328-333.

Thompson, E. R., & Phua, F. T. T. (2005). Reliability among senior managers of the Marlowe-Crowne Short-Form Social Desirability Scale. Journal of Business and Psychology, 19, 541-554.