Within a multivariate statistical design, the relationship of various interpersonal, emotional, and cognitive variables to suicidal ideation and behavior in college students was examined. A total of 158 subjects (58 males, 100 females) completed self-report measures of life stress, faulty cognitions, loneliness, depression, hopelessness, family cohesiveness, adaptive reasons for living, and suicidal ideation and behavior. The results of a multiple-regression analysis, forward-inclusion algorithm, indicated that a predictive equation consisting of loneliness, irrational beliefs, and low adaptive reasons for living best accounted for suicidal behavior scores. To determine the basic structures and power of the predictor variables under study, a factor analysis and composite regression were employed, resulting in a predictive equation consisting of three factors: Social/Emotional Alienation, Cognitive Distortions, and Deficient Adaptive Resources. Results are discussed in terms of an interactional model of suicidal behavior.

The results suggest that the effectiveness, and ultimately the cost, of homeless services can be improved by matching the type of service to the consumer's level of psychiatric impairment and substance use rather than by treating mentally ill homeless persons as a homogeneous group.

This study examined the reliability and construct validity of a modified version of the Colorado Symptom Index (MCSI), a brief, self-report measure of psychological symptomatology, in a study of interventions to prevent homelessness. Eight projects in a national, cooperative study collected new data at baseline, 6, and 12 months using a set of common measures as well as site-specific instruments. The pooled sample consisted of 1,381 persons in treatment for mental illness or substance abuse (or both), of which 84% had a history of homelessness. The analyses employed classical and Rasch methods to examine the MCSI's content validity, internal consistency and item quality, test/retest reliability, dimensionality, appropriateness for the sample, construct validity, and responsiveness to change. This 14-item scale was found to be a reliable and valid measure of psychological symptoms in this sample. Its content was consistent with other symptom measures, its high internal consistency and test-retest coefficients supported its reliability, its relationships to other measures indicated that it had good construct validity, and it was responsive to change. We conclude that the MC

No abstract

In order to test the predictive validity of a stress--vulnerability model of suicide ideation and behavior, a longitudinal study was conducted with college undergraduates. Based on previous research (Bonner & Rich, 1987; Rich & Bonner, 1987a), measures of social/emotional alienation, adaptive resources for living, and cognitive rigidity were taken at the beginning of the semester and combined to develop a vulnerability score. Then, at midterm, measures of midterm stress and cumulative negative life stress were obtained from the same subjects. It was hypothesized that the vulnerability factor would interact with midterm and negative life stress to predict suicide ideation. Instead of an interactive model, a linear, compensatory model of suicide ideation was found. The combination of alienation, deficient adaptive resources, and life stress best predicted ideation scores. Implications for research and theory are noted.