PROBLEMThe Marlowe-Crowne Social Desirability Scale (M-C SDS) has had considerable use in the decade since its pub1ication.l Some data show, however, that several of its items contribute relatively little to the overall measure. This finding, together with a desire for shorter scales more useful in some research contexts, has led to construction of three homogeneous, short-form measures whose psychometric adequacy, in part, is supported by cross-validation. METHODApproximately 500 university students in two classes of an introductory psychology course were given packets of questionnaires to complete outside of class. Return of these anonymously filled-in materials provided credit toward a research participation requirement. One of the questionnaires was composed of the 33 M-C SDS items intermixed with 40 others to form extraversion-introversion and neuroticism scales. Three hundred and sixty-one students, 176 males and 185 females, returned usable questionnaires (a dozen or so others were discarded, for incompleteness or other reasons).A principal components analysis was performed on the M-C SDS items, with sex and class as additional dichotomous variables; correlations were phi coefficients. Size of loading on the first principal component was the primary criterion for selection of items to form two 10-item social desirability scales [M-C l(10) and M-C 2(10)] and, through their combination, a 20-item measure [M-C (20)l. (This criterion yielded results very like those based on item-total scale correlations.) A second requirement was that each scale have equal numbers of positively-and negatively-keyed items (in the M-C SDS, 18 items are keyed true, 15 false) as a control for possible acquiescence set. Two other criteria proved superfluous: items with badly imbalanced true-false splits and items with content judged somewhat inappropriate for college student Ss (e.g., "I never make a long trip without checking the safety of my car") were excluded on grounds of low component loadings.Separate consideration of sexes appeared unnecessary. The sex variable loaded very little (-.04, females higher) on the first component, and parallel analyses for the sexes showed quite similar patterns. The correlation between corresponding (absolute) item loadings for males and females was 3 5 , which speaks to the reliability of the overall analysis as well as to the congruence of the sexes. The class variable, as expected, exhibited a trivial first component loading (-.11). RESULTSThe first component accounted for about 13% of the total variance, not a large amount, but somewhat more than double that of the next largest, second component. All the M-C SDS items loaded in the appropriate directions on this first, general dimension, though 12 had absolute values less than .30. The range of absolute loadings for all items was .ll to .54, with a mean of .35; the range of loadings for the 20 selected items was .28 to .54, with a mean of .42.Items that form the new scales are given in Table 1, with direction of keying noted and with item n...

The drawbacks of a t ratio measure of psychological androgyny are discussed. Other measurement approaches are indicated, and a substantive issue is addressed. Bern (1974) recently described a sex-role inventory that provides separate Femininity and Masculinity scale scores and uses their f-ratio difference as a measure of psychological androgyny. There are, however, both descriptive and inferential problems associated with the application of Student's t. This article describes these problems, suggests their resolution through use of the simple difference score rather than the t ratio, and points out additional measurement procedures that might prove worthwhile. It also reports a further analysis of the Bern data. The Bern Sex-Role Inventory (BSRI) and Its ScoringThe BSRI consists of a number of personality characteristics to be self-rated by subjects from 1 ("Never or almost never true") to 7 ("Always or almost always true"). The 20-item Femininity scale contains traits judged to be more desirable in American society for a woman than for a man, and the mean of ratings for these traits is called the Femininity score (here denoted F). Similarly, the 20-item Masculinity scale contains traits judged to be more desirable for a man than for a woman, and the mean of these ratings is called the Masculinity score (M). A subject's Androgyny score (also called the Androgyny t ratio) is the difference between the Femininity and Masculinity scores expressed as Student's t, that is, the difference between means, F -M (the Androgyny difference score), divided by a term reflectingThe author thanks Randi Hagen for bringing this subject to his attention and Leroy Wolins for his helpful comments.Requests for reprints should be sent to Robert

Eighty subjects estimated the correlation coefficient, r, for each of 13 computer-printed scatterplots. Making judgments were 46 students in a graduate-level statistics course and 34 faculty and graduate students in a department of psychology.The actual correlation values ranged from .010 to .995, with 200 observations in each scatterplot and with the order of scatterplot presentation randomized. As predicted, subjects underestimated the degree of actual correlation. Also as predicted, but with substantial moderation by a method-of-presentation factor, this underestimation was most pronounced in the middle of the correlational range—between the 0 and 1 extremes. Though perception of correlation was shown not to be veridical (i.e., in terms of r), little support was given one alternative view—its being in terms of r 2 .

No abstract

No abstract