Study members completed the MPQ at age 18. We used the MPQ because (a) it was developed and standardized with nonclinical populations, (b) it yields a comprehensive profile of human psychological differences along multiple personality dimensions, (c) its reliability and validity are well-established, and (d) previous developmental studies with the MPQ have established that the personality traits measured by this psychological test are partially heritable, predictable from childhood, and very stable from adolescence to adulthood (Caspi & Silva, 1995; McGue, Bacon, & Lykken, 1993; Tellegen et al., 1988). Thus, the traits assessed by the MPQ appear to reflect consistent and enduring patterns of behavior in the general population.The MPQ (Form NZ) yields scores on 10 different personality scales, each of which includes a minimum of 11 items. Scale names and descriptions of high scorers on each scale are presented in Table 1, The MPQ was designed to promote low correlations among its primary scales, allowing for a description of personality in terms of multiple, relatively independent dimensions. The scale intercorrelations range from -.37 to .40, with a mean absolute value of .17. The low magnitudes of these intercorrelations are similar to those obtained in American samples (Tellegen et al., 1988).The personality scales comprising the MPQ can be viewed at the higher order level as defining three distinct superfactors (Tellegen, 1985; Tellegen & Waller, in press). As shown in Table 1, Constraint is a