THE EFFECTS OF KEEPIN´ IT R.E.A.L.´S CLASS VIDEOS AND TELEVISED PSA'S ON MIDDLE-SCHOOL CHILDREN´S ALCOHOL, CIGARETTE, AND MARIJUANA USE. Jennifer Warren1, Michael Hecht1, David Wagstaff1, Elvira Elek2, Khadidiatou Ndiaye1, Flavio Marsiglia3, Patricia Dustman3, 1Pennsylvania State University, University Park, PA United States; 2Pennsylvania State University, Waldorf, MD United States; 3Arizona State University, Tempe, AZ United States
Once a prevention program has proven successful it is important to examine the efficacy of its various components. This paper reports on two video component of the keepin' it REAL curriculum which proved successful in reducing ATOD use between 7th and 8th grades (Hecht, et al., 2003). keepin´ it R.E.A.L. is a 10-lesson, drug prevention curriculum that was developed for a culturally diverse group of students attending 35 middle schools in Phoenix, Arizona. The curriculum used culturally grounded narratives within a communication competence and performance framework to enhance students´ anti-drug norms and attitudes, and facilitate the development of their risk assessment, decision making, and resistance skills. Students´ regular classroom teachers showed the five, four-to-six-minute videos (produced by local high school students) that introduced the curriculum and used culturally relevant values, stories, and communication styles to teach students the four resistance skills (Refuse, Explain, Avoid, and Leave). The narratives were subsequently incorporated into eight, thirty-second public service announcements (PSAs) that were designed to reinforce the curriculum and aired on local television stations. The present study assessed the effects achieved by the keepin´ it R.E.A.L.´s videos and televised PSAs.
A total of 4,734 intervention and control students reported the amount and frequency of alcohol, cigarette, and marijuana use in the past 30 days during a pretest, and the follow-up posttest 14 months after the end of the intervention. The analysis models focused separately on the six substance use outcomes. For testing the impact of the videos, an analysis of covariance (ANCOVA) model regressed the change in substance use between the 14 month posttest and the pretest on the substance use reported at baseline and two dummy variables to contrast control students with (a) intervention students reporting that they had seen 4 or 5 videos and (b) intervention students reporting that they had seen 0 to 3 videos. For testing the impact of the PSAs, models regressed substance use at the 14 month posttest on the prettest substance use, the number of ads seen, condition (intervention, control) and the interaction between condition and number of ads seen. All of the models were fit with the complex survey sample routines in Stata. The results suggested that students who saw 4 or 5 videos had better outcomes than students who saw fewer videos. The number of PSAs seen did not predict substance use 14 months after the intervention.
THE E-CHUG: A RANDOMIZED, CONTROLLED STUDY OF A WEB-BASED BINGE DRINKING INTERVENTION WITH COLLEGE FRESHMAN. John Steiner1, W. Gill Woodall1, Jill Anne Yeagley1, 1University of New Mexico, Albuquerque, NM United States
“Binge” or dangerous drinking remains a serious problem on college campuses despite modest progress made in recent years through environmental, policy, and social norms approaches. The efficacy of brief motivational feedback to reduce drinking among college students has been reported by numerous researchers (Agostinelli, Brown & Miller, 1995; Marlatt, Baer, Kivlahan, et al.1998). The Check Up To Go (CHUG), designed by William Miller and Joseph Miller at UNM´s Campus Office of Substance Abuse Prevention (COSAP), incorporated the successful techniques of motivational interviewing and normative feedback into a written instrument for college students. COSAP has successfully used the paper/pencil version of the CHUG with student policy violators referred to its Alcohol and Other Drug Awareness & Education Program since 1998. In a review conducted for COSAP, Walters (1999) found that students receiving mailed personalized feedback regarding their levels of risk, usage percentile rankings as compared to U.S. and UNM norms, and alcohol expenditures, significantly reduced their drinks per week (DPW).
In response to the growing numbers of students who indicate they use a computer and access Internet sites on a daily basis, Walters and his colleagues developed an electronic, web-based version of the CHUG. To date, research has focused on the paper/pencil version of this instrument and almost solely on its efficacy with student policy violators; consequently, research staffs at COSAP and UNM´s Center on Alcohol, Substance Abuse and Addictions (CASAA) decided to examine the E-CHUG in a randomized, controlled study with UNM freshman, an identified high-risk student group. Currently, a randomized trial is underway where freshman students are recruited from large introductory classes at UNM. Once recruited and randomized, all communication with participants occurs via email. At baseline, experimental group subjects receive an email with the internet Universal Resource Locator (URL) for the E-CHUG on COSAP´s website, and control group subjects receive a 30-day quantity and frequency measure identical to the one imbedded in the E-CHUG with instructions to return it via email reply. At a 90-day follow-up assessment, participants in both groups receive another email containing a 30-day quantity and frequency measure. The trial will be completed in the spring of 2005; the target sample size is 200 subjects. Between groups differences in reduction of binge drinking and DPW will be analyzed and presented. Additionally, other preliminary data from the E-CHUG trial will be presented.
LONGITUDINAL GROWTH CURVE EFFECTS OF A MULTI-YEAR UNIVERSAL INTERVENTION: THE PATHS CURRICULUM IN THE FAST TRACK PROJECT. Chi-Ming Kam1, Mark Greenberg1, And C.P.P.R.G1, 1Pennsylvania State University, State College, PA United States
The present study evaluates the effect of a universal intervention in the Fast Track Project. The universal intervention included the Fast Track variation of the PATHS curriculum, and a series of school-wide efforts that include consultation with the principal. Fast Track is a multi-site conduct disorder trial focused on the prevention of early aggression in high-risk children. The present report only focuses on the universal aspect of Fast Track and examines the progression of children's socio-emotional outcomes across early elementary school years. The sample in this study comprises 3,027 students in 37 schools (17 experimental, 20 control) at three study sites (Rural Pennsylvania: 99% white; Nashville, TN: 42% white and 56% African Americans; Seattle, WA: 53% white and 18% African Americans). Schools were randomized to intervention status. PATHS lessons were taught by teachers and were focused on self-control, emotions and relationships, and problem-solving skills. The intervention began in grade 1 and continued through grade 3 in intervention schools. PATHS consultants visited classrooms regularly and also provided behavioral consultation. Teachers completed yearly ratings each spring on aspect of behavior including both aggressive and prosocial domains. Data from students who have all three years of data were analyzed. Individual growth curve analyses (using SAS Proc Mixed) were used to estimate trajectories of teacher-ratings of students' outcomes. Moderators (baseline level of problems, gender, and ethnicity) were added to the growth curve models to test for their moderating effects on both treatment/control differences in grade 3 outcomes and the trajectories of study outcomes. Treatment effects were observed for both boys and girls in rural Pennsylvania. By grade 3, children in PATHS classrooms showed lower rates of aggression and higher levels of prosocial behavior. In addition, treatment effects were stronger for subgroups of children who started with a higher level of aggression. Similarly, in Seattle, high-risk children in treatment classrooms showed significant reductions in aggression, but no changes in prosocial behavior. However, in Nashville, there were no significant treatment effects on aggression, but significant effects were found for prosocial behavior (with greater effects for treatment vs. control girls). Discussion will focus on (1) alternative explanations for site differences as well as similarities in outcomes, (2) the influence of universal interventions on students with high starting levels of problem behavior, (3) the need to study multi-year programming.
THE IMPACT OF CLASSROOM CHARACTERISTICS ON THE EFFECTIVENESS OF THE LIFT PREVENTIVE INTERVENTION PROGRAM. Rebecca Silver1, J. Mark Eddy2, 1University of Oregon, Eugene, OR United States; 2Oregon Social Learning Center, Eugene, OR United States
Meaningful progress has been made towards the school-based prevention of conduct problems and juvenile delinquency in children and adolescents. Numerous programs have been developed and have empirical support of their effectiveness. However, little is known about what kinds of classroom-level factors may influence the effectiveness of these school-based preventive interventions. While there is some evidence that the aggregate level of classroom aggression in elementary school increases the risk of aggressive behavior in middle school for boys, especially for highly aggressive boys, it remains unclear whether this classroom-level variable has an impact on program effectiveness (Kellam, Ling, Merisca, Brown, & Ialongo, 1998). Although Kellam and his colleagues found trends in the expected direction, not all tests of the hypothesis reached statistical significance.
The current study will examine the impact of classroom-level factors on the effectiveness of the Linking the Interests of Families and Teachers (LIFT) preventive intervention program for conduct problems. Approximately 670 children participated in LIFT when they were in 1st or 5th grades during the early 1990's. Participants are still being followed today. LIFT is a multifaceted, school-based preventive intervention based on an ecological-developmental theory of conduct problems. It combines parent training, a classroom-based social skills program, a playground behavioral program, and systematic communication between teachers and parents. In previous reports, LIFT has been found to reduce child physical aggression, decrease aversive maternal behavior, and increase child positive behavior with peers, all of which are precursors and correlates of conduct problems. However, it is not yet known if classroom-level variables influence the effectiveness of LIFT on these outcomes; that is the aim of this poster. Classroom characteristics examined will include classroom behavioral climate (an aggregate classroom score based on teacher, parent, peer, and independent observer report of child behavior) and aggregate levels of classroom poverty. Multiple and logistic regression models, including interactions between classroom-level factors and intervention condition, will be used.
Understanding the ways in which classroom characteristics impact the effectiveness of preventive interventions is critical for improving the impact of such programs. The move from efficacy models to real-world, effectiveness programs involves implementing programs in complex systems. The better we understand the role that characteristics of structures within these systems (such as classrooms within a school-based intervention) play in program implementation, the more likely we will be successful.
EFFECTS OF THE "SUNNY DAYS, HEALTHY WAYS" CURRICULUM ON STUDENTS IN GRADES 6-8.. David Buller1, Kim Reynolds2, Amy Yaroch3, Gary Cutter4, Joan Hines5, Cristy Geno6, Julie Maloy5, Melissa Brown7, W. Gill Woodall8, Joseph Grandpre9, 1Cooper Institute, Denver, CO United States; 2University of Southern California, Alhambra, CA United States; 3National Cancer Institute, Bethesda, MD United States; 4University of Alabama, Birmingham, Birmingham, AL United States; 5Cooper Institute, Golden, CO United States; 6Kaiser Permanente of Colorado, Aurora, CO United States; 7Jefferson County Public Schools, Lakewood, CO United States; 8University of New Mexico, Albuquerque, NM United States; 9Wyoming Department of Health, Cheyenne, WY United States
Sun protection of children is a national priority to combat the rising rates of skin cancer, but there are few effective education programs for use in secondary schools. The project aims were to create a sun safety curriculum for grades 6-8 and to test whether exposure to the curriculum would increase children´s sun protection behavior. The "Sunny Days, Healthy Ways" curriculum contained six 50-minute lessons with content and instructional strategies based on Social Cognitive Theory (SCT). It was intended to increase perceived personal risk for skin damage, positive outcome expectations about sun protection to reduce personal risk, self-efficacy expectations for performing sun protection, and key sun protection skills. It contained activities to help children set and monitor sun protection goals and to overcome commonly encountered barriers to sun safety. A pair-matched group-randomized pretest-posttest controlled trial was performed with 30 middle schools in Colorado, New Mexico, and Arizona. Teachers implemented the "Sunny Days, Healthy Ways" curriculum in 2001-03 and analyses were performed in 2003-04. Children reported on their sun protection behavior using frequency ratings and a diary. In the schools, 2,038 children were enrolled, and 1,788 completed the posttest. However, one school pair was eliminated, yielding a final sample of 1,769. Compared to control schools, children receiving the curriculum reported more frequent sun protection (p=.0035), and a greater proportion wore long-sleeved shirts during recess (p<.0001) and applied sunscreen (p<.0001). Colorimeter measures of skin redness and darkening due to sun exposure validated children´s self-reports. Exposure to the curriculum improved knowledge of sun protection (p<.0001), decreased perceived barriers to using sunscreen (p=.0046), and enhanced self-efficacy expectations about sun safety (p=.0577), and reduced favorable attitudes toward sun tanning (p=.0026 to <.0001). In intent-to-treat analyses, the effect of the curriculum remained significant for the frequency of sun protection (p=.0032) and wearing long-sleeved shirts (p<.05) and wearing sunscreen (p<.05), when dropouts were assigned their pretest value as the posttest response. The effect was eliminated only under the most conservative assumption, i.e., dropouts did not wear long-sleeved shirts (p>.05) or sunscreen (p>.05). The educational approaches for secondary schools in "Sunny Days, Healthy Ways" based on principles of SCT appear to be effective for improving young adolescents´ sun safety. Potential trial limitations include measuring short-term outcomes, focusing on young adolescents, using active parental consent, and testing in the American southwest.
PREDICTORS OF PARTICIPATION IN A PREVENTIVE PARENTING INTERVENTION FOR DIVORCED FAMILIES. Emily Winslow1, Darya Bonds1, Sharlene Wolchik1, Irwin Sandler1, 1Arizona State University, Tempe, AZ United States
Parenting interventions have proven to be helpful in preventing child mental health and substance use problems. However, participation rates in parenting programs are often low (20-25%), severely limiting their potential impact at the population level. Research conducted to date suggests high SES parents are more likely to participate in parenting interventions than lower SES parents. Some evidence also suggests parents who perceive higher levels of family dysfunction may be more likely to participate than those who report fewer problems. Similar to several other prevention trials, we have found that families functioning more poorly at baseline receive the most benefit from our program. We are particularly interested in determining if those most likely to benefit agree to participate. We hypothesized that mothers who perceived high family dysfunction would appropriately self-select into the intervention; however, we expected this relation to be more consistent among higher SES than lower SES families. Participants included 356 mothers recruited from court records for a randomized controlled trial of an 11-session parenting group intervention for divorced mothers of children aged 9 to 12. Of the 650 families reached by phone who met eligibility criteria, 45% (n=294) initially agreed to participate. A subset of those who initially refused the intervention were recruited for interviews--60% of eligible intervention refusers agreed to be interviewed (n=62). Measures included maternal report of sociocultural characteristics, family dysfunction characteristics (i.e., child adjustment, parenting quality, interparental conflict, maternal distress), paternal involvement in child-related decision-making, and time since divorce. A factor analysis was performed on sociocultural and family dysfunction variables to create composite measures of these constructs. Results of multivariate logistic regression analyses, χ2(4, N=355)=20.60, p<.0001, indicated that SES (OR=2.12, SE=.46, p<.01), family dysfunction (OR=1.85, SE=.46, p<.05), and paternal involvement in decision-making (OR=2.15, SE=.65, p<.05) uniquely predicted maternal agreement to participate, whereas time since divorce was not significant (OR=.97, SE=.02, p=.13). These findings suggest that mothers whose families could most benefit from intervention (i.e., more poorly functioning) tend to self-select appropriately into the intervention; however, low SES mothers and those who are the sole decision-makers are at risk for nonparticipation. Additional generative research is needed to identify processes that mediate associations between these predictors and participation to develop strategies likely to be effective in recruiting families at risk for nonparticipation who could benefit most from intervention.
EFFECTS OF A LOCALIZED ANTI-VIOLENCE MEDIA CAMPAIGN ON TARGETED OUTCOMES. Randall Swaim1, Kathleen Kelly1, Julie Chen1, 1Colorado State University, Fort Collins, CO United States
A randomized, controlled trial of a school-based anti-violence media campaign ("Resolve It, Solve It") was conducted among 1492 7th and 8th-grade rural students over a three-year period. Six rural middle schools were randomly assigned to treatment and control conditions and students were assessed on measures of violence and relevant covariates at baseline and two follow-ups. In three treatment communities, 11th grade students were trained to administer the campaign to middle school students and to serve as local media spokespersons. During a one-week training program, the 11th-grade students were featured in print, radio, and TV anti-violence PSA's. During the intervention period, each treatment school was exposed to the media products (print, radio, TV PSA's) that featured local 11th-grade students presenting anti-violence messages. Program effects were assessed using multi-group tests of latent growth models for measures of attitudes toward violence, intent for future violence, self efficacy for avoiding violence involvement, perceived school safety, verbal assault, physical assault against objects, physical assault against people, verbal victimization, and physical victimization. Comparison of slopes across treatment and control groups for the entire sample indicated significant program effects in the expected direction for intent for violence (p.<.001), perceived school safety (p.=.011), physical assualt against people (p.=.029), and verbal victimization (p.=.041). Analyses conducted within gender indicated only one significant program effect for male students, with treatment students indicating a slower rate of growth for intent for future violence compared to control students (p.=.001). Female students in the treatment condition evidenced a lower rate of growth in intent for future violence (p.=.050), and a reduction in physical assault against people compared to an increase among control student (p.=.017), Trends for significant effects (p.<.10) were observed among male students for perceived school safety (p.=.07), physical assault against objects (p.=.082), verbal victimization (p.=.072), and physical victimization (p.=.068). Past media campaigns targeting substance use have resulted in relatively weak effects compared to school-based interventions. In some cases, media components, coupled with school based interventions have demonstrated independent effects. The current results indicate promising effects for a stand-alone localized media intervention for violence prevention. Future study comparing localized and non-localized designs, as well as stand-alone media interventions vs. those coupled with school-based interventions will clarify the effectiveness of this media approach to violence prevention.
EFFECTS OF A LOCALIZED ANTI-VIOLENCE MEDIA CAMPAIGN ON NON-TARGETED SELF-ESTEEM. Randall Swaim1, Kathleen Kelly1, 1Colorado State University, Fort Collins, CO United States
A randomized, controlled trial of a school-based anti-violence media campaign ("Resolve It, Solve It") was conducted among 1492 7th and 8th-grade rural students over a three-year period. Six rural middle schools were randomly assigned to treatment and control conditions and students were assessed on measures of violence and relevant covariates at baseline and two follow-ups. In three treatment communities, 11th grade students were trained to adiminister the campaign to middle school students and to serve as local media spokespersons. During a one-week training program, the 11th-grade students participated in the development and anti-violence print, radio, and TV PSA's in which they were featured. Program effects were assessed using multi-group tests of latent growth models for overall self-esteem and three subscales of self-esteem: (1) self-confidence, (2) social acceptance, and (3) competence. Comparison of slopes across treatment and control groups for the entire sample indicated that while overall self-esteem declined for both groups, the rate of decline was slower among treatment students (p.=.008). Tests of slopes across gender indicated that the effect was present only among female students (p.<.001). Similar patterns of effects were observed for subscales for self-confidence and social acceptance. For each of these factors, significant effects were observed for the total sample, but were found only among female students following tests within gender. As with overall self-esteem, self-confidence declined for both treatment and control students over the three-year period, but the rate of decline was less among treatment students (p.=.015). This effect was due entirely to a slower rate of decline among female treatment students (p.<.001). Social acceptance increased slightly among treatment students overall and decreased among control students (p.=.011). Tests by gender confirmed previous patterns in which effects were obtained only among female students, for whom treatment students reported an increase in social acceptance compared to a decrease among control female students (p.=.001). For the subscale competence, no effects were obtained for the overall sample. Tests within gender indicated that the rate of decline in self-reported competence was slower among treatment female students compared to control female students (p.=.025). While not directly targeted in this localized anti-violence media campaign, significant effects were obtained for overall self-esteem, and its subscales. However, effects were limited to female students.