Table of Contents
Variable
The L scale was designed to identify respondents who are deliberately trying to appear socially desirable while completing the MMPI.
Description
The scale comprises 15 statements about attitudes and practices that are socially undesirable but common. Topic areas include minor dishonesties, aggression, bad thoughts, and weaknesses of character. As on the MMPI as a whole, the response format is True-False. For all items, “False” is the response scored as a lie. In current usage, scores of8 or above are considered suggestive of purposeful self-presentation (Greene, l 980).
Samples
Hathaway and McKinley (1951) reported means of 4.2 (s.d. = 2.6) and 4.5 (s.d. = 2.6) for males and females, respectively. In a more recent sample of 765 college students, Goldberg (1972) reported means of 2.5 (s.d. = 1.9) and 2.7 (s.d. = 1.8) for males and females, respectively. In a massive sample of 50,000 medical outpatients, Swenson, Pearson, and Osbourne (1973) reported means of 4.2 (s.d. = 2.3) and 4.8 (s.d. = 2.3) for males and females, respectively.
Reliability
Internal Consistency
Gocka (1965) reported an alpha coefficient of .72 on a patient sample. Paulhus (l 984) reported an alpha value of .60 on a student sample.
Test-Retest
Test-retest correlations for intervals of up to l week range from .70 to .85. For intervals of I year or more, correlations range from .35 to .60 (Greene, l980; Rorer & Goldberg, 1965).
Validity
Convergent
Evidence for concurrent validity is available from studies showing high correlations with similar constructs, for example, Eysenck’s Lie scale (Paulhus, 1986) and the Marlowe Crowne scale (e.g., Edwards & Walsh, 1964). As noted in the introduction, the L scale loads on both factors of SDR but primarily on the second factor, impression management. No claim has been made that high scores should predict lying outside of self-reports.
In the only known laboratory study, high scorers performed better than lows in a stressful situation (Burish & Houston, 1976).
Discriminant
The measure shows low correlations with measures loading on the first factor of SDR. For example, correlations with the K scale, the alternative SDR measure from the MMPI, range from low to moderate (Dahlstrom, Welsh, & Dahlstrom, 1972).
Location
Hathaway, S. R., & McKinley, J. C. (1951). MMPI manual. New York: Psychological Corporation.
Dahlstrom, W. G., Welsh, G. S., & Dahlstrom, L. E. (1972). An MMPI handbook (Vol. 1). Minneapolis: University of Minnesota Press.
Comments
It was soon apparent to the constructors of the L scale that the desirability demand of the items was less than subtle: “The L score was a trap for the naive subject but easily avoided by more sophisticated subjects” (Meehl & Hathaway, 1946). It is obvious to a sophisti cated test-taker that, even if one is trying to appear desirable, it is unrealistic to deny such ubiquitous attributes.
Given its negative correlation with intelligence, some have construed the L scale as a measure of psychological sophistication. College-educated persons and those of higher socioeconomic status rarely score above 4 (Dahlstrom et al., 1972). When they do, it may suggest a gross lack of insight into their own behavior (Greene, I 980).
MMPI Lie Scale
Sample Items
-
At times I feel like swearing. (F)
TRUE FALSE
-
I get angry sometimes. (F)
-
Sometimes when I am not feeling well I am cross. (F)
Complete Scale
The MMPI booklet numbers for the 15-item Lie scale are l 5, 30, 45, 60, 75, 90, 105, 120, 135, 150, 165, 195, 225, 255, 285. All items are keyed negatively, that is, one point is assigned for each “False.”