Psychological Assessment 1999, Vol. 11, No. 4, 546-555 Copyright 1999 by the American Psychological Association, Inc. 1040-3590/99/S3.00 Factor Structure and Convergent Validity of the Conflict Tactics Scale in High School Students Michele Cascardi, Sarah Avery-Leaf, K. Daniel O'Leary, and Amy M. Smith Slep State University of New York at Stony Brook This study explored the factor structure of a modified version of the Conflict Tactics Scale (CTS; M. A. Straus, 1979) in a large multiethnic high school sample. Exploratory and confirmatory factor analytic approaches were used. Results generally supported 2-factor models for males and females. A substantial proportion of residual variance remained after the 2 primary factors were extracted, and correlations among this residual variance suggested meaningful differences in the perpetration and experience of dating violence for males and females. Furthermore, the factor structure for males' self-reported victimization suggested that items representing psychological and mild physical aggression, which loaded on 1 factor, may be perceived similarly. Convergent validity analyses that examined the correlation among CTS traditional and factor scores with jealous actions, control tactics, and attitudes justifying males' and females' use of dating aggression provided initial support for the constructs identified. Results are discussed in terms of improving measurement of dating aggression. Adolescent dating violence has received increased research at- tention over the past decade. Physical aggression occurs with notable frequency in high school dating relationships. Estimates range from 9% to 41%; rates of self-reported victimization are similar, ranging from 12% to 41% (Avery-Leaf, Cascardi, O'Leary, & Cano, 1997; Bergman, 1992; Henton, Gate, Koval, Lloyd, & Christopher, 1983; O'Keefe, 1997; Roscoe & Callahan, 1983). Unfortunately, psychometric properties of the most fre- quently used measure of this phenomenon, the Conflict Tactics Scale (CTS; Straus, 1979), have not been investigated in high school samples. The purpose of this study was to conduct a preliminary examination of the factor structure and convergent validity of a modified version of the CTS in a multiethnic high school student sample. The CTS includes measures of both one's own use of physical aggression and one's experience of one's partner's physical aggression (i.e., victimization), so this article is divided into two separate analyses: one reporting on self-reported aggression, the other on self-reported victimization. The CTS's status as the most commonly used measure of partner aggression supports exploration of its factor structure in an adolescent sample before steps are taken to alter it in ways that might enhance its validity but also render comparisons to the large CTS-based literature impossible. Furthermore, our work and that Michele Cascardi, Sarah Avery-Leaf, K. Daniel O'Leary, and Amy M. Smith Slep, Department of Psychology, State University of New York at Stony Brook. Michele Cascardi and Sarah Avery-Leaf are now with the Dating Violence Prevention Project, Inc. of Glen Ridge, NJ, and Durham, NH. Michele Cascardi is also affiliated with Women Against Abuse in Phila- delphia, PA. Support for this study came from National Institute of Mental Health Grant MH 4780103 to K. Daniel O'Leary. Correspondence concerning this article should be addressed to Michele Cascardi, 169 Forest Avenue, Glen Ridge, New Jersey 07028. Electronic mail may be sent to mcascardi@aol.com. of others suggests that the aggression items included on the CTS are relevant to adolescent'samples. In our work, the CTS was reviewed by several clinicians and educators, as well as piloted in three separate studies (one conducted in a middle school and two in high schools; Avery-Leaf et al., 1997; Cano, Avery-Leaf, Cas- cardi, & O'Leary, 1998). Other research teams that have conducted extensive development generally conclude that additional behav- iors (e.g., scratching, burning) need to be added to the CTS. Nonetheless, the behaviors included in the CTS are still salient and appropriate to include when administering it to adolescents (see Foshee et al., 1996). Factor Analysis Studies of Self-Reported Aggression The factor structure of the CTS has been demonstrated in adult military, community, clinic, and college student samples (e.g., Barling, O'Leary, Jouriles, Vivian, & MacEwen, 1987; Caulfield & Riggs, 1992; Pan, Neidig, & O'Leary, 1994; Straus, 1990). In general, the CTS has received support as a reliable and valid measure, with few gender differences in the interpretation of factor solutions. A commonly accepted factor solution includes four factors: Reasoning, Verbal/Psychological Aggression, Mild Phys- ical Aggression, and Severe Physical Aggression. However, items that are face valid for physical or psychological aggression do not always load accordingly in factor analysis (e.g., "threatened to throw something at partner," which is face valid for psychological aggression, loads with physical aggression items; Caulfield & Riggs, 1992). Indeed, Caulfield and Riggs (1992) noted that many factor analytic studies of the CTS fail to demonstrate clear dis- tinctions between verbal/psychological and physical aggression items. Furthermore, a severe physical aggression factor has not been consistently identified (Barling et al., 1987; TenVergert, Kingma, & Gillespie, 1990). One potentially useful way to reconcile these item-factor dis- crepancies is with confirmatory factor analysis (CFA). With CFA, one can test whether there is a statistically significant difference in goodness of fit between a measurement model that allows contro- 546