The following four alternative hypotheses were tested: that low self-control and psychopathy constitute a single construct, that they constitute primary and secondary psychopathy or interpersonal and intrapersonal traits, or that they constitute antisociality (the tendency to perform antisocial acts) and low self-direction (the tendency to act in one's long-term benefit). The study used self-report data from a sample of 152 male and 155 female college students (n=307), who completed surveys during classes at a southwestern university in 1996. In measuring psychopathy and self-control, an initial pool of 65 items came from several sources. Twenty-one were drawn from the data used by Hirschi (1969) in his initial presentation of social control theory (the predecessor of self-control theory); 5 were drawn from the self-report scale of self-control (the Grasmick scale); and 2 were drawn from the work of Gibbs and Giever (1995). Other items were written for this study, and some were pilot-tested with college students. Crime and delinquency were measured by asking participants whether they had ever participated in seven listed activities, including three categories of theft, car theft or joyriding, vandalism, robbery, and assault. The analysis began with a series of exploratory factor analyses of the initial pool of 65 items. With one exception, the factors seemed to correspond to self-control and psychopathy. Factor-based scales were constructed for use in subsequent analyses. The scales were entered into bivariate regression analyses with self-reported delinquency. Finally, SEM's were estimated to investigate the three alternative higher-order factor structures. The structural equation models that produced the best fit suggest that antisociality, the degree to which one's activities are antisocial or prosocial, should be considered a separate construct from self-direction, the degree to which one is able and willing to work in one's own long-term self-interest. Alternative models that propose a unitary self-control or psychopathy construct, or constructs the represent primary and secondary psychopathy, fit the data poorly and were unequivocally rejected. The results also suggest that the effects of self-direction on delinquency are indirect, working through antisociality, implying a possible developmental process. Study limitations are noted, and recommendations for future research are offered. 4 tables, 1 figure, 110 references, and appended factors, items, and loadings