Antiviolence Education in High Schools: Implementation and Evaluation.

Studies of college students typically find that at least 20% report perpetrating or sustaining physical violence in dating relationships (Barnes, Greenwood, & Sommer, 1991; Bethke & DeJoy, 1993; Cate, Henton, Koval, Christopher, & Lloyd, 1982; Kelly & DeKeseredy, 1994; Magdol et al., 1997; Makepeace, 1981; Pederson & Thomas, 1992; Riggs, 1993). High school students report similar rates of physical aggression (Day, 1990; Fitzpatrick, 1990; Fitzpatrick & Halliday, 1992; Gray & Foshee, 1997; Henton, Cate, Koval, Lloyd, & Christopher, 1983; Mercer, 1987; O'Keeffe, Brockopp, & Chew, 1986; Peterson & Olday, 1992; Roscoe & Callahan, 1985; Roscoe & Kelsey, 1986; Sudermann & Jaffe, 1993). Sexual assault is also reported by teens in dating relationships (Ageton, 1983; Davis, Peck, & Storment, 1993; Poitras & Lavoie, 1995). Marital violence often begins before marriage (e.g., Avni, 1991; O'Leary, Malone, & Tyree, 1994; Roscoe & Benaske, 1985), and dating violence can be a very serious problem in itself (e.g., Levy, 1991).

Sexual-assault prevention programs targeting college men generally consist of lectures designed to improve attitudes and empathy and typically use weak evaluations (see review by Schewe & O'Donohue, 1993). Programs for women have used stronger designs. Hanson and Gidycz (1993) provided written information about sexual assault, its prevention, and community resources for 181 college women. Participants also watched and discussed videos describing acquaintance rape and its prevention. Compared with a control group, participants reported engaging in fewer risky behaviors (e.g., drinking on a date) and suffering fewer sexual assaults during the 9-week follow-up period.

Antiviolence education for high school students, especially concerning male-to-female violence in relationships, also has become popular (e.g., Board of Education for the City of London, 1992; Canadian Teacher's Federation, 1990; Feltey, Ainslie, & Geib, 1991; Marin Abused Women's Services, 1986; Mitchell et al., 1993; Tulloch & Tulloch, 1992; Violence Prevention, 1993; see reviews by Jaffe, Sudermann, Reitzel, & Killip, 1992; Mercier & Lavoie, 1992). The prevalence of such interventions is unknown. The few published evaluations cover interventions for students ranging from Grade 7 to high school seniors, designed primarily to change attitudes and knowledge.

Attitude Change

One of the first published evaluations demonstrated that attitudes toward violence against women can change after a 1-day session involving an assembly and classroom discussions (Jaffe et al., 1992). However, some changes were in the wrong direction: The intervention was associated with an increase in boys condoning date rape. A 6-week follow-up showed a weakening of some earlier improvements and new changes in the undesired direction. With a larger sample, Sudermann and Jaffe (1993) reported positive changes and little attitude backlash on date rape items but undesirable changes on opinions about the causes of violence.

An attitude backlash is not surprising, given the psychological research on attitude change (e.g., Brehm, 1966; Eagly & Chaiken, 1993; Schauss, Chase, & Hawkins, 1997). Studies over the past 30 years show that trying to persuade people to change their attitudes in one direction can lead instead to people taking more extreme versions of their existing attitudes. This effect has been demonstrated clearly by Winkel and de Kleuver (1997), who found that a video presentation of sexual assault showing the undesirable consequences for the male perpetrator increased boys' self-reported acceptance of macho behavior, rape myths, and coerced sex.

In other evaluations, reports of attitude change are ambiguous. Lavoie, Vezina, Piche, and Boivin (1995) evaluated two high school classroom sessions about dating violence in which a selected subgroup of students with the least favorable attitudes had more desirable attitudes at posttest; however, this effect could be a statistical phenomenon resulting from the selection of a restricted range of the sample. Krajewski, Rybarik, Dosh, and Gilmore (1996) reported that a classroom curriculum for Grade 7 yielded some effects in the wrong direction, but they did not describe these results in detail. Another classroom program for middle school students, however, resulted in significant attitude changes only in the desirable direction (Macgowan, 1997).

Careful attention should be paid to attitudes, and efforts should be made to avoid undesirable changes. Winkel and de Kleuver (1997) report that attitudes were related to self-reported use of sexual intimidation, so an attitude backlash is potentially harmful. Jaffe et al. (1992) suggest that students showing an attitude backlash are already involved in violent relationships, and they recommend alternative interventions. Other evidence suggests that students in general do not have particularly bad attitudes toward violence (Hilton, in press). Can perpetrators of violence benefit from antiviolence education aimed at the general student body? In contrast, do victims show any particular benefits?

Knowledge Change

Antiviolence education also has been used to improve knowledge about dating violence (Lavoie et al., 1995), violence against women in general, warning signs of abuse, community resources (Jaffe et al., 1992), child abuse (Macgowan, 1997), conflict resolution strategies (Krajewski et al., 1996), and consequences of perpetrating violence (Winkel & de Kleuver, 1997). However, knowledge tests often consist of simplified statistics such as "at least one in ten Canadian women is assaulted by the man she lives with" (Jaffe et al., 1992), or opinions such as "respecting the other person in a dating relationship means never getting angry with him or her" (Lavoie et al., 1995, p. 520), and "when a woman is hurt or killed due to abuse, both the abuser and society are responsible" (Krajewski et al., 1996, p. 107). Can practical information (e.g., risk factors for assault, steps for controlling anger, available services, etc.) be conveyed to students? And can such information be conveyed without an attitude backlash? Can practical information be retained? Existing evaluations suggest that improvements on knowledge tests decline by 6 weeks (Jaffe et al., 1992) and disappear by 5 months (Krajewski et al., 1996).

Evaluation Methodology

Methodological improvements already are apparent since the first publications appeared, but aspects of research design still limit the possibility that the demonstrated effects can be attributed specifically to the education. Three main methodological concerns are discussed here.

Test reliability. Of the evaluations reported above, only Krajewski et al. (1996) referred to test-retest …