This study evaluated the effects of a school-based intervention on growth trajectories of smoking, drinking, and antisocial behavior among early adolescents. Seven middle schools were randomized to intervention or comparison conditions and students in two successive cohorts (n = 1484) provided five waves of data from sixth to ninth grade. The Going Places Program, included classroom curricula, parent education, and school environment compo- nents. Latent growth curve analyses demonstrated significant treatment group effects, in- cluding reducing increases in friends who smoke, outcome expectations for smoking, and smoking progression, but had non-significant effects on drinking or antisocial behavior. The Going Places Program was effective in preventing increases in smoking progression, but its efficacy as a more cross-cutting problem behavior preventive intervention was not confirmed.

KEY WORDS: adolescents; intervention; problem behavior; schools.

INTRODUCTION

Early initiation of substance use and antisocial behavior can represent a troubling deflection from positive and healthful development (Simons-Morton & Haynie, 2002). Early adolescent problem behavior is associated with immediate and lasting problems, including abuse and dependence, school under- achievement, and involvement with the juvenile justice system (Chassin et al., 1990; Gruber et al., 1996; Weisberg et al., 2003). Moreover, substance use and antisocial behavior co-occur (Huizinga et al., 2000), suggesting that involvement in one problem behavior may lead to involvement in another and that they may share common mediators. Antisocial

1 Prevention Research Branch DESPR, NICHD, NIH, Rockville, Maryland.

2 NeuroScience, Inc, Bethesda, Maryland. 3 Prevention Research Branch National Institute of Drug Abuse, NIH, Rockville, Maryland.

4 Georgetown University, Washington, District of Columbia. 5 Correspondence should be directed to Bruce Simons-Morton, EdD, MPH Prevention Research Branch DESPR, NICHD, NIH, 6100 Executive Blvd 7B13M Rockville, Maryland 20852- 7510; e-mail: [email protected].

behavior is a particular problem for schools because much of this behavior is directed at or occurs at school (Addington et al., 2002), disrupting school routines, preoccupying teachers and administrators, and contributing to underachievement and dropout. Most schools provide programmatic attention to substance abuse (Wenter et al., 2002), but multi- ple problem prevention programs are relatively rare.

Substance use, aggression, and antisocial behav- ior increase dramatically during early adolescence (Addington et al., 2002; Johnson et al., 2002; Snyder & Sickmond, 1999). The prevalence of smoking in the past 30 days increases from 3.8% among sixth graders (Simons-Morton, 2001) to 17.7% among eighth graders and 26.7% of 12 graders (Johnston et al., 2003). Similarly, drinking increases from 6.5% among sixth graders to 19.6% of eighth graders (Simons-Morton et al., 1999) and 48.6% of 12th graders (Johnston et al., 2003). Some 20–34% of middle school students report being in a physical fight (Snyder & Sickmond, 1999), 7.2% of sixth graders and 4.1% of eighth graders reported vi- olent victimization at school (Nansel et al., 2001; Addington et al., 2002), 7.4% reported frequent

188 Simons-Morton, Haynie, Saylor, Crump, and Chen

bullying in the past year (Haynie et al., 2001), and 14% reported carrying a weapon to school (Durant et al., 1999).

Social cognitive (Bandura, 1986) and problem behavior theories (Jessor & Jessor, 1977) empha- size the role of immediate and long-term interac- tion between adolescents and their social contexts. Early adolescence is a period of intense acquisi- tion of social skills, the development of which facil- itates school adjustment and engagement and orients them toward positive and constructive enterprise (Morrison et al., 2002; Snyder & Sickmond, 1999). Delays in the development of social skills and com- petence may lead to poor decision making, anti- social attitudes, affiliation with problem behaving friends, and aggressive and other antisocial behav- ior (Aunola et al., 2000; Paetsch & Bertrand, 1997; Vazsonyi & Flannery, 1997). However, previous re- search has shown only limited effects on problem behavior from general social skills training (Caplan et al., 1992), suggesting the need to apply social skills training to specific domains, as has been done in substance abuse prevention (Botvin, 2000) and men- tal health (Weisberg et al., 2003) Some research on social competence approaches with pre-adolescents has focused on multiple problem behavior preven- tion (Lochman & Wells, 2002), but concerns re- main about how best to foster skill development across problem behaviors within the limits of a sin- gle school-based program (Johnson et al., 1996).

The effect of the Going Places Program on adolescent substance use and antisocial behavior was tested in a school-randomized trial. The program addressed sixth graders as they entered middle school and attempted to shape the attitudes and skills thought to motivate them to try hard in school and refrain from problem behavior. In addition to individual-level goals the program sought to increase parent and school support for positive youth be- havior. In previous observational analyses of the data from the study, a variety of partially overlap- ping perceptions and attitudes, including outcome expectations, social competence, peer norms, peer affiliation, and parenting behaviors, have been linked to substance use (Simons-Morton et al., 1999, 2001; Simons-Morton, 2002; Simons-Morton & Haynie, 2003), aggression (Simons-Morton et al., 2003), and school adjustment (Simons-Morton & Crump, 2003). The purpose of this paper is to describe the effect of the Going Places Program on problem behaviors among youth randomized to The Going Places Program or to a comparison group.

METHOD

Design and Study Participants

The seven middle schools in one Maryland school district were recruited and randomized, three to the treatment and four to the comparison condi- tion. Starting with the 1996 school year, two succes- sive cohorts of sixth grade students were recruited and followed through ninth grade. Students in the treatment schools were exposed to a classroom cur- riculum designed to increase social skills and pre- vent multiple problem behaviors, including smoking, drinking, and antisocial behavior. Students were sur- veyed at the beginning and end of the sixth grade and toward the end of seventh and eighth grades and be- ginning of ninth grade.

The population of interest included 2969 stu- dents eligible to participate in the Time 1 assessment. Students who attended special education classes ex- clusively were excluded based on pilot testing that found them unable to provide reliable data. Of the 2969 eligible students 2651 (87.8%) provided consent while in the 6th grade. When consent was obtained again upon entry into high school in the ninth grade, 59 parents refused consent and 242 students failed to return completed forms, leaving a sample of 2350 with consent (79.2%). Of these, 119 became ineligi- ble during the study by failing a grade or being newly classified as special education, leaving a sample of 2231, of whom 32 moved out of the school district and 715 were absent and missed an assessment, leav- ing a final sample of 1484. An additional 164 study participants were not included in the final analyses because they did not provide data on smoking on all five surveys. Compared with the final sample, study participants lost to follow up were significantly more likely to be black, live in a single parent family, and to have reported smoking, drinking, and antisocial behavior. Attrition by treatment group did not vary by sex, one or two parent family, drinking, or antisocial behavior, but was greater among Blacks in the intervention than the comparison group.

Intervention

The Going Places program includes a social skills curriculum, parent education, and school environment enhancement designed to increase academic engagement and commitment to school;

Evaluation of Going Places 189

alter perceptions, attitudes, and expectations about substance use and antisocial behavior; and reduce multiple problem behaviors. The intervention components were based on previous research on social skills training (Schinke et al., 1991), school environmental change (Gottfredson, 1986), social development (Kumpfer & Turner, 1990), and social cognitive (Bandura, 1986) theories, consistent with recommended best prevention practices (Nation et al., 2003; Botvin, 2000).

The foci of the curriculum sessions are prob- lem solving, self-control, communication, and con- flict resolution skills. Eighteen sessions were of- fered in the sixth grade, 12 in the seventh grade, and six in the eighth grade. A typical lesson in- cludes a trigger videotape (featuring local talent and locations) in which common problems are pre- sented and problem solving approaches are mod- eled by the actors, followed by a brief teacher-lead discussion of the relevant skills, various interactive group activities and role plays, and skills practice with constructive feedback. The classroom teach- ers who delivered the program received substan- tial training and continuing support from a master teacher.

The enhanced school environment component included social marketing strategies to improve school climate, establish prosocial norms, establish a positive image for the school, reinforce student achievement, and extend exposure to Going Places curriculum concepts. Activities designed to extend the classroom lessons to the larger school environ- ment included informational “roll-outs” preceding each unit; posters and short video segments pre- sented in the cafeteria and display areas; “travelers’ checks” awarded to students by teachers for applying skills learned in a Going Places lesson; and special ac- tivities, such as assemblies and a year-end field trip, which students could earn the right to attend by par- ticipating in class activities and demonstrating skills outside the classroom, which were systematically re- warded.

The parent education component was designed to increase parental monitoring, involvement, and expectations regarding academic engagement and problem behavior. Parents received in the mail a 20-min instructional video on authoritative parenting and a 20-page booklet entitled Attentive Parenting: The Going Places Guide for Parents of Middle School Students and periodic newsletters. Students were as- signed homework that required the involvement of a parent or guardian.

Measurement

Each assessment period, including special make- up sessions in each school to accommodate students who were absent for the original assessment, was completed in all seven schools within the 3 weeks just proceeding or following implementation of the intervention. Two trained proctors per classroom ad- ministered the questionnaire to intact classes of ap- proximately 30 students each. Teachers remained to manage the classroom, but were otherwise not in- volved in the conduct of the survey. To protect confi- dentiality, students completed and turned in a cover page that included name, survey identification num- ber, birth date, and homeroom teacher’s name that was kept separate from the questionnaires. The ac- tual questionnaires included only a numerical identi- fier matching the one on the cover page.

The questionnaire included items on sex, race, study-school attended, mother’s education, and the number of parents in the home. Youth were asked how many times they smoked cigarettes and drank alcohol (in the past 30 days and past 12 months) and how often they intend to smoke and drink while in high school. Smoking and drinking were treated as ordinal variables, with nonoverlapping categories of never (nonusers), future intent (intenders), 12-month (12-month user), 30 day (recent users), and three or more times in past 30 days (frequent users). Antiso- cial behavior was assessed by items that asked how often in the past year the youth had been in a phys- ical fight, been in a physical fight in which someone got hurt, bullied or picked on someone younger or weaker, lied to a parent or guardian about where they were or whom they were with, gone someplace dan- gerous or off-limits, stole something from a person or store, or carried a weapon. The antisocial behav- ior variable was treated as a count.

Indices to measure the intermediate outcomes targeted by the intervention activities are described in detail elsewhere and have been shown to have good internal consistency (Simons-Morton et al., 1999). Problem behaving friends was assessed by ask- ing the respondent to indicate how many of his or her five closest friends (0–5) smoke drink alcohol, cheat on a test, bully someone, act disrespectfully, steal, lie to parents, or damage property. Friends who smoke and friends who drink were also treated as separate, single-item variables. Outcome expectan- cies were assessed with five items on drinking and five items on smoking that asked about perceived positive and negative consequences of engaging in

190 Simons-Morton, Haynie, Saylor, Crump, and Chen

these behaviors. We developed nine items measur- ing social competence that asked respondents to rate their ability to resist dares, communicate clearly, re- solve conflicts, and retain self-control. Deviance ac- ceptance included eight items that asked if it was okay for kids to smoke, drink, use drugs, cheat on tests, bully other kids, lie to parents, disrupt class, or go dangerous places. Parental expectations asked how upset their parents would be if they found out the teen drank alcohol, smoked cigarettes, got into a fight, acted disrespectfully to a teacher, and did poorly on a test. School climate assessed students’ perceptions of teacher supportiveness and clarity and fairness of school rules.

ANALYSIS

To assess implementation, teacher reports were collected, classes were observed, student knowledge was tested, and student participation was recorded. A volunteer sample of parents provided information over the telephone about the intervention materi- als mailed home to them. To assess the outcomes, growth in targeted intermediate outcomes and in the prevalence of each problem behavior was compared across treatment groups. Treatment group effects were examined separately for each of the problem behaviors using the generic Latent Growth Curve (LGC) modeling approach (Curran, 2000; Muthen & Muthen, 2001). Subsequently, to control for baseline differences in outcomes, ANCOVA was performed comparing the follow-up measures using the baseline of the outcome as a covariate.

LGC modeling employs advantages of Random Coefficient Modeling in the study of individual differ- ences in longitudinal settings (Raudenbush & Byrk, 2002) and Structural Equation Modeling (Bollen, 1989), a flexible approach to incorporating latent and observed variables. All analyses were conducted us- ing MPlus software (Muthen & Muthen, 2001), which offers a Weighted Least Square estimator adjusting means and variance with corrected standard errors necessary for ordered categorical outcomes.

Line graphs showing developmental trajectories for treatment and control groups were first produced to evaluate the treatment effects in an exploratory fashion. Possible treatment effects were tested us- ing the Added Growth Factor (AGF; Muthen & Curran, 1997), because for ordered categorical data the chi-square difference test used in traditional multi-group comparison is not valid (Muthen &

Curran, 1997, p. 381). Accordingly, the mean and variance of the intercept (estimated baseline) and slope values were constrained to be equal for both groups, with an additional growth factor specified only for treatment group. As the only difference in the model, the added growth factor represents the ef- fect of treatment.

Growth Mixture Modeling (Muthen et al., 2002) was conducted to evaluate the treatment effects for each latent class within treatment group. In Growth Mixture modeling, different trajectories, or patterns of responses across time are identified that represent groups of individuals with similar behavior over time. It is assumed in these analyses that an individual has a certain trajectory class membership that does not change over time and that the intervention effects are captured in average slopes for each class. Accord- ingly, this analysis captures the effect of intervention on changes in the slope of the trajectories for each class.

The best number of trajectory classes was de- termined by identifying the smallest Bayesian Infor- mation Criterion (BIC) index for various numbers of classes (control group only) (Muthen & Muthen, 2001). MPlus was used to test statistical difference in the slopes of comparable classes within treatment and control groups. Line graphs of class trajectories for treatment and control groups were produced for visual interpretation of the treatment effects.

In preliminary analyses using the Random Co- efficient Model in HLM (Raudenbush & Byrk, 2002) we found the variance of smoking stage among the seven schools not significantly different from zero (chi-square = 8.64, p = 0.194). Using One-Way ANOVA we found that mean smoking stage was not significantly different among schools. Therefore, school was not included in the final analyses.

The overall fit indices for evaluating LGC models included Comparative Fit Index (CFI), the Tucker–Lewis Index (TLI) and the Root Mean Square Error of Approximation (RMSEA). Most re- searchers regard a model “acceptable” if CFI and TLI are greater than 0.9. A model with an RMSEA less than 0.05 is regarded as a “good” fit and an RM- SEA less than 0.08 is “acceptable” (McDonald & Ho, 2002).

RESULT

Teachers reported completing 95% of the lessons sixth grade and 84% in the seventh grade. During scheduled observations, teachers completed

Evaluation of Going Places 191

all core lessons. Teacher ratings of the percentage of students fully participating in the lessons, includ- ing, staying on task, were 90% in the sixth grade and 88% in the seventh grade. On average 77% of sixth grade students and 65% of seventh grade students in- dicated usually or always paying attention to the Go- ing Places classes and 70% considered the program helpful to their lives, with 54% reporting they used the skills from the program at school and 41% in- dicating they used program skills outside school. On knowledge tests, 90.2% scored 80% or better. Of 45 parents interviewed, 40 indicated that they or their spouse viewed the video and parenting booklet, of which 80% reported liking these materials.

The means and standard deviations for smok- ing stage, drinking stage, and antisocial behavior are shown by time, sex, race, and treatment group in Table 1. Increases over time and variation by sex, race, and treatment group in smoking, drinking, and antisocial behavior can be noted. As shown in Fig. 1, average smoking stage is lower for the treatment

group compared with the control group, especially at the second (sixth grade Spring semester) and the third (sevnth grade Spring Semester) measurement. However, as shown in Fig. 1, negligible treatment group differences were found for drinking or antiso- cial behavior.

The approach to analysis of the treatment effect as an Added Growth Factor (AGF) was according to Muthen and Curren (1997). Figure 2 presents the model diagram for the multigroup comparison using the AGF approach, with the treatment group having an additional growth factor that is absent (i.e., set to zero) for the control group. The common model was specified as a linear growth model. Gender and race were used as the covariates for controlling demo- graphic differences. All parameters were constrained as equal across the treatment and the control groups, except for the mean and variance of the added fac- tor. The AGF was specified as quadratic (Muthen & Curran, 1997, p. 382), indicating that when the control group provides a linear factor it may

Table 1. Prevalence of Smoking Stage, Drinking Stage, and Antisocial Behavior From Sixth (T1) to Ninth Grade (T5)

Mean (SD)

Category T1 T2 T3 T4 T5

Smoking stage (N = 1320) Range = 0–4

0.19 (0.60) 0.53 (1.01) 0.78 (1.21) 0.87 (1.28) 0.97 (1.30)

Male (n = 570) 0.23 (0.68) 0.51 (1.01) 0.75 (1.20) 0.75 (1.25) 0.87 (1.38) Female (n = 750) 0.17 (0.34) 0.56 (1.02) 0.81 (1.21) 0.96 (1.31) 1.05 (1.43) White (n = 939) 0.20 (0.60) 0.57 (1.04) 0.83 (1.24) 0.88 (1.29) 1.02 (1.44) Black (n = 273) 0.17 (0.60) 0.45 (0.92) 0.68 (1.15) 0.89 (1.10) 0.89 (1.36) Other (n = 108) 0.19 (0.63) 0.41 (0.92) 0.69 (1.03) 0.77 (1.25) 0.79 (1.32) Treatment (n = 692) 0.15 (0.52) 0.45 (0.94) 0.66 (1.08) 0.80 (1.26) 0.85 (1.32) Control (n = 628) 0.23 (0.68) 0.63 (1.08) 0.93 (1.31) 0.95 (1.32) 1.11 (1.50) Drinking stage (N = 1311)

Range = 0–4 0.35 (0.85) 0.71 (1.11) 1.06 (1.24) 1.34 (1.37) 1.49 (1.42)

Males (570) 0.44 (0.93) 0.77 (1.15) 1.11 (1.24) 1.31 (1.36) 1.39 (1.40) Females (750) 0.29 (0.77) 0.67 (1.07) 1.02 (1.25) 1.36 (1.38) 1.56 () White (939) 0.20 (0.60) 0.57 (1.04) 0.83 (1.24) 0.88 (1.29) 1.02 (1.49) Black (273) 0.17 (0.60) 0.45 (93) 0.68 (1.49) 0.88 (1.31) 0.89 (1.36) Other (108) 0.19 (0.63) 0.41 (0.92) 0.69 (1.03) 0.77 (1.25) 0.79 (1.30) Treatment (n = 692) 0.36 (0.86) 0.70 (1.12) 0.99 (1.17) 1.3 (1.38)6 1.47 (1.40) Control (n = 620) 0.35 (0.84) 0.73 (1.09) 1.14 (1.31) 1.32 (1.36) 1.51 (1.45) Anti-Antisocial behavior

(n = 1465) Range = 0–4 0.88 (1.31) 1.16 (1.45) 1.27 (1.47) 1.26 (1.31) 1.17 (1.46)

Males (n = 645) 1.29 (1.46) 1.54 (1.57) 1.64 (1.54) 1.35 (1.58) 1.40 (1.53) Females (n = 820) 0.56 (1.46) 0.86 (1.28) 0.97 (1.34) 1.03 (1.41) 0.99 (1.38) White (n = 1023) 0.85 (1.29) 1.11 (1.44) 1.23 (1.46) 1.25 (1.51) 1.17 (1.43) Black (n = 322) 1.03 (1.37) 1.36 (1.49) 1.44 (1.48) 1.34 (1.53) 1.19 (1.47) Other (n = 120) 0.73 (1.27) 0.98 (1.40) 1.12 (1.47) 1.19 (1.45) 1.06 (1.45) Treatment group (n = 773) 0.93 (1.33) 1.14 (1.44) 1.32 (1.43) 1.31 (1.52) 1.18 (1.46) Control group (n = 692) 0.82 (1.29) 1.18 (1.48) 1.21 (1.51) 1.21 (1.50) 1.15 (1.47)

192 Simons-Morton, Haynie, Saylor, Crump, and Chen

Fig. 1. Adolescents’ average smoking, drinking, and antisocial behavior sixth to ninth grade for treatment (N = 692) and control groups (N = 628)

Evaluation of Going Places 193

Fig. 2. Treatment effect as an added growth factor for the developmental trajectory of early adolescent cigarette smoking from sixth-grade fall to ninth-grade fall.

suffice to have an added quadratic factor for treat- ment group.

The model in Fig. 2 fits very well with the data, with χ2(13) = 20.28, p = 0.09, CFI = .998, TLI = .997 and RMSEA = .029. The mean for the constrained common slope factor was 0.19 (not shown in Fig. 2), indicating that both groups progressed in smoking stage across time. The variance of the slope was 0.11 (p < 0.05), indicating a significant individual variation in the smoking progression. The estimated

mean for the AGF was −0.124 (p < 0.05), showing an additional decrease in the rate of smoking progres- sion for the treatment group relative to the control group. The variance of the added group factors was not significant, indicating no increase in variability in smoking progression in the treatment group. Race was not a significant predictor of slope, but gender was, indicating that girls progressed relatively faster than boys in smoking stage. Similar gender effects were not found for alcohol use or antisocial behavior.

194 Simons-Morton, Haynie, Saylor, Crump, and Chen

As shown in Fig. 1, average smoking stage at baseline differed somewhat in the intervention and comparison groups. ANCOVA was conducted to compare average smoking stage at each time point controlling for baseline smoking. The analyses in- dicate significant differences at S6 (F1, 1317 = 6.06, p = 0.014), S7 (F1, 1317 = 12.2, p = 0.001), and F9 (F1, 1317 = 8.4, p = 0.004), with the control group hav- ing significantly higher smoking stages when the F6 baseline value was partialed out.

Growth Mixture Modeling (GMM; Muthen, et al., 2002) was applied to examine in detail the impact of the Going Places program on subgroups (unobserved classes) that characterize different smoking trajectories. Figure 3 presents the smoking

trajectories over time in each treatment group for the five-class model resulting from the GMM analysis: Class 1, those who never smoked and did not intend to start; Class 2, those who indicated intent to smoke, but did not start over the period of assessment; Class 3, those with delayed initiation followed by rapid acceleration; Class 4, those who initiated early but smoked infrequently; and Class 5, those who initiated early and accelerated rapidly to frequent smoking. As shown, most of the group differences are due to a slower increase in smoking stage for Classes 3–5.

Having found a treatment group effect on smok- ing, we examined the effect of the program on inter- mediate outcomes. Using the same AGF approach as

Fig. 3. Comparing treatment effects based on five classes based on growth mixture modeling.

Evaluation of Going Places 195

before, but in this case with interval level measures, we examined the treatment group differences on the slopes of each variable. Effects were found in favor of the intervention group through the ninth grade for friends who smoked (mean of added slope = −0.21, p < 0.001); and through the seventh grade for out- come expectations for smoking (mean of added slope = −0.30, p = 0.014). Deviance acceptance ap- proached significance (mean of added slope = −0.24, p = 0.073). Parent Involvement declined less for the treatment group than the control group, but the effect was not significant (mean of added slope = 0.12, p = 0.127). There was no treatment ef- fect on perceptions of school climate or school en- gagement, which declined steadily over time in both groups.

DISCUSSION

The potential of multiple problem behavior pro- grams for early adolescents has not been realized and many questions about their efficacy remain (Johnson et al., 1996). The results of this study indicate that the Going Places Program was successful in moderating the increase in smoking, but had no effect on other targeted problem behaviors. The strengths of the study include a longitudinal, school-randomized design, an extensive theory-based intervention, adequate implementation, well-defined outcome measures, and latent growth modeling analyses. However, the study is limited by its small, regional sample and loss to follow of problem behaving youth. Fortunately, in this study, loss to follow up of high-risk youth was not different between treatment conditions, avoiding the most serious threat to validity posed by these losses.

In general it has proven difficult to prevent adolescent problem behavior. One review (Hansen, 1992) found that only about 50% of the best- designed substance-use prevention studies reported significant effects. Smoking has been somewhat less difficult to prevent than drinking or antisocial behav- ior (Tobler & Statton, 1997). Although effect sizes have been relatively small (Tobler, 1986), even small reductions in adolescent smoking may be important given the health and social costs of early experimen- tation (Chassin et al., 1990; Gruber et al., 1996). In the current study, LGM provided an unusually powerful approach to examining the nature of the treatment effect. ANOVA confirmed that the treatment ef- fect on smoking was significant when baseline smoke stage was included as a covariate. Treatment group

effects on targeted mediators were consistent with an overall effect on smoking progression.

Surprisingly, the primary intervention effect was to moderate the rate of increase in smoking among youth who had already begun to experiment with smoking. This suggests that the intervention may have been more salient for early adolescents with some experience with smoking than for those who had never smoked. Most program evaluation stud- ies have examined effects on smoking initiation and have not examined progression (Tobler & Statton, 1997). It would be useful for other researchers to re- examine the effects of completed interventions on similar subgroups.

Despite the positive effect of Going Places on smoking progression, no effects were noted for drinking or antisocial behavior. The amount of curriculum time devoted to smoking, drinking, and antisocial behavior was similar. Analyses of interme- diate outcomes indicated significant effects on the increase in the number of friends who smoke and outcome expectations for smoking, confirming the impact of the program on smoking related variables. We believe that the combination of more negative outcome expectations for smoking and association in the program with other nonsmoking youth led to an increase in friendships with non-smoking youth and to reduced progressing of smoking behavior. This novel finding suggests the potential for school-based interventions to foster prosocial peer affiliations, an approach we are exploring in our current re- search. In contrast, deviance acceptance was the only intermediate outcome thought to be common to each of the problem behaviors that approached sig- nificance, suggesting that the program did not impact on key intermediate outcomes and this failure of the program helps explain why effects on drinking and antisocial behavior were not significant.

Many prevention studies have addressed mul- tiple substances (Botvin, 2000) and some have re- ported effects on smoking and other outcomes such as diet (Perry et al., 1986), but few prevention pro- grams directed at early adolescents have successfully addressed both substance use and antisocial behav- ior. One of the few examples of studies targeting both substance use and antisocial behavior is the All Stars Character Education Program, which reported negative findings for all problem behavior outcomes (Harrington et al., 2001). The lack of effects on multiple outcomes of Going Places and other pro- grams that have attempted to alter multiple problem behaviors suggests that each problem behavior may

196 Simons-Morton, Haynie, Saylor, Crump, and Chen

require a substantial and specific dose of interven- tion for each target behavior. In Going Places sub- stantial amounts of curriculum time were required to help students, particularly 6th graders, understand the concepts, place the lessons into personally mean- ingful contexts, and apply them to specific behav- ioral situations, suggesting the need for highly spe- cific learning activities. Multiple problem behavior prevention might have more potential with older adolescents.

Many factors can impact on program success, including the extent and quality of implementation (Hansen et al., 1991), the selection of program ob- jectives, and the creative linking of program activ- ities with objectives. Implementation of the Going Places Program appears to have been adequate and the treatment effect found for smoking, friends who smoke and outcome expectancies, and deviance ac- ceptance provide evidence that program theory and development were appropriate for this behavior. Al- though the program had programmatic effects on smoking-related variables, it did not have similar ef- fects on variables linked to drinking and antisocial behavior. Possibly, the program activities were more appropriate for smoking than drinking and antiso- cial behavior, implementation may have been better for smoking, the selection of intermediate variables was better for smoking than other behaviors, or the limited time available to provide students with re- peated skills practice in each problem behavior lim- ited the program’s broader effectiveness. In any case, it must be concluded that exposure to the Going Places Program modified smoking progression but did not have an effect on multiple problem behaviors among middle school students. The results do not support the promise of school-based prevention for multiple problem behavior, at least with early ado- lescents.

REFERENCES

Addington, L. A., Ruddy, S. A., Miller, A. K., DeVoe, J. F., & Chandler, K. A. (2002). Are America’s schools safe? Students speak out: 1999 school crime supplement: statistical analysis report. National Center for Educational Statistics, Office of Educational Research and Improvement, U.S. Department of Education, NCES Publication no. 2002–331.

Aunola, K., Stattin, H., & Nurmi, J.-E. (2000). Adolescents’ achievement strategies, school adjustment, and externalizing and internalizing problem behaviors. Journal of Youth and Adolescent, 29(3), 289–306.

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, : Prentice-Hall NJ

Bollen, K. A. (1989). Structural equations with latent variables. Wiley New York

Botvin, G. J. (2000). Preventing drug abuse in schools: Social and competence enhancement approaches targeting individual— Level etiologic factors. Addictive Behaviors, 25(6), 887–897.

Caplan, M., Weissberg, R. P., Grober, J. S., Sivo, P., Grady, K., & Jacoby, C. (1992). Social competence promotion with inner- city and suburban young adolescents: Effects on social ad- justment and alcohol use. Journal of Consulting and Clinical Psychology, 60, 56–63.

Chassin, L., Presson C. C., Sherman, S. J., & Edwards, D. A. (1990). The natural history of cigarette smoking: predict- ing young-adult smoking outcomes from adolescent smoking patterns. Health Psychology, 9(6), 701–716.

DuRant, R. H., Krowchuk, D. P., Kreiter, S., Sinal, S. H., & Woods, C. R. (1999). Weapon carrying among middle school students. Archives of Pediatric and Adolescent Medicine, 153(1), 21–26.

Gottfredson, G. D. (1986). An empirical test of school-based envi- ronmental and individual interventions to reduce the risk of delinquent behavior. Criminology, 24, 705–730.

Gruber, E., DiClemente, R. J., Anderson, M. M., & Lodico, M. (1996). Early drinking onset and its association with alco- hol use and problem behavior in late adolescence. Preventive Medicine, 25, 293–300.

Hansen, W. B. (1992). School-based substance abuse prevention: a review of the sate of the art in curriculum, 1980–1990. Health Education Research, 7(3), 403–430.

Hansen, W. B., Graham, J. H., Wolkenstein, B. H., & Rohrbach, L. A. (1991). Program integrity as a moderator of Prevention progam effectiveness: Results for Fifth-grade students in the adolescent alcohol prevention trial. Journal of Studies on Al- cohol, 52(6), 568–579.

Harrington, N. G., Giles, S. M., Hoyle, R. H., Feeney, G. J., & Yungbluth, S. C. (2001). Evaluation of the all stars charac- ter education and problem behavior prevention program: Ef- fects on outcome variables for middle school students. Health Education and Behavior, 28(5), 533–546.

Haynie, D. L., Nansel, T., Eitel, P., Crump, A. D., Saylor, K., Yu, K., & Simons-Morton, B. G. (2001). Bullies, victims, and bully—victims: Distinct groups of youth at risk. Journal of Early Adolescent, 21(1), 95–107.

Huizinga, D., Loeber, R., Thornberry, T. P., & Cothern, L. (2000). Co-occurrence of delinquency and other problem behav- iors. Juvenile Justice Bulletin, Office of Juvenile Justice and Delinquency Prevention, OJP, USDJ, November, 1–5.

Jessor, R., & Jessor, S. (1977). Problem behavior and psychosocial development. New York, NY: Academic Press.

Johnson, C. A., MacKinnon, D. P., & Pentz, M. A. (1996). Breadth of program and outcome effectiveness in drug abuse preven- tion. American Behavioral Scientist, 39(7), 884–896.

Johnston, L. D., O’Malley, P. M., & Bachman, J. G. (2003). Mon- itoring the future study national survey results on drug use, 1975–2002. Vol. I: Secondary school students. NIH Pub. NO. 03-5375. Bethesda, MD: National Institute on Drug Abuse.

Kumpfer, K. L., & Turner, C. W. (1991) The social ecology model of adolescent substance abuse: Implications for prevention. The International Journal of the Addictions, 25(4A), 435– 563.

Lockman, J. E., & Wells, K. C. (2002). The coping power program at the middle-school transition: Universal and indicated pre- vention effects. Psychology of Addictive Behaviors, 16(Suppl. 4), S40–S54.

Loeber, R., Farrington, D. P., & Petechuk, (2003). Child delin- quency: Early intervention and Prevention. Child Delin- quency Bulletin Series, Office of Jevenile Justice and Delinquency Prevention, Office of Justice Programs, USDJ, 1–18 (also available at ojjdp.ncjrs.org).

McDonald, R. P., & Ho, M. (2002). Principles and practice in re- porting structural equation analysis. Psychological Methods, 7, 64–82.

Evaluation of Going Places 197

Morrison, G. M., Robertson, L., Laurie, B., & Kelly, J. (2002). Protective factors related to antisocial behavior trajectories. Journal of Clinical Psychology, 58(3), 277–290.

Muthen, B., & Curran, P. J. (1997). General longitudinal model- ing of individual differences in experimental designs: A la- tent variable framework for analysis and power estimation. Psychological Methods, 2, 371–402.

Muthen, L. K., & Muthen, B. O. (2001). Mplus user’s guide. Muthen & Muthen: Los Angeles.

Muthen, B. (2001). Latent variable mixture modeling. In G. A. Marcoulides & R. E. Schumacker (Eds.), New developments and techniques in structural equation modeling (pp. 1–33). Mahwah, NJ Erlbaum.

Muthen, B., Brown, C. H., Masyn, K., Jo, B., Khoo, S. T., Yang, C. C., Wang, C. P., Kellam, S., Carlin, J., & Liao, J. (2002). General growth mixture modeling for randomized preven- tive interventions. Biostatistics, 3, 459–475.

Nansel, T. R., Overpeck, M., Pilla, R., Ruan, W. J., Simons- Morton, B., & Scheidt, P. (2001). Bullying behaviors among U.S. youth: Prevalence and association with psychosocial ad- justment. JAMA, 285(16), 2094–2100.

Nation, M., Crusto, C., Wandersman, A., Kumpfer, K. L., Seybolt, D., Morrissey-Kane, E., & Davino, K. (2003). What works in prevention: Principles of Effective Prevention Programs. American Psychologist, 58(6/7), 449–456.

Paetsch, J. J., & Bertrand, L. D. (1997). The relationship between peer, social, and school factors, and delinquency among youth. Journal of School Health, 67(1), 27–32.

Patterson, G. R., Reid, J. B., & Dishion, T. J. (1992). Antisocial boys. Eugene, OR: Castalia.

Perry, C. L, Klepp, K. I., Halper, A., Hawkins, K. G., & Murray, D. M. (1986). A process evaluation study of peer leaders in health education. Journal of School Health, 56, 62–67.

Raudenbush, W. R., & Byrk, A. S. (2002). Hierarchical linear mod- els: Applications and data analysis methods (2nd ed.). Thou- sand Oaks, CA: Sage.

Schinke S. P., Botvin G. J., & Orlandi, M. A. (1991). Substance abuse in children and adolescents: Evaluation and interven- tion. New York: Sage.

Simons-Morton, B. G. (in press). Prospective association of peer influence, school engagement, alcohol expectancies, and par- ent expectations with drinking initiation among sixth graders. Addictive Behaviors.

Simons-Morton, B. G. (2002). Prospective Analysis of Peer and Parent Influences on Smoking Initiation Among Early Ado- lescents. Prevention Science, 3(4), 275–283.

Simons-Morton, B. G., & Crump, A. D. (2003). The association of parental involvement and social competence with school

adjustment and engagement among sixth graders. Journal of School Health, 73(3), 121–126.

Simons-Morton, B. G., Crump, A. D., Haynie, D., Eitel, P., Saylor, K., & Yu, K. (1999). Psychosocial, school, and par- ent factors associated with smoking among early adolescent boys and girls. Preventive Medicine, 28, 138–148.

Simons-Morton, B. G., Hartos, J., & Haynie, D. L. (2004). Fac- tors related to minor aggression among sixth graders. Health Education and Behavior, 31(1), 22–33.

Simons-Morton, B. G., & Haynie, D. (2002). Growing up drug free: A developmental challenge. In M. H. Bornstein, L. Davidson, C. L. M. Keyes, K. A. Moore, & The Center for Child Well-being. Well-being: Positive development across the life course, (pp. 109–122). Mahwah, NJ: Erlbaum.

Simons-Morton, B. G., & Haynie, D. L. (2003). Proportional odds analysis of psychosocial predictors of increased smoking dur- ing sixth grade. American Journal of Health Behavior, 27(6), 592–602.

Simons-Morton, B. G., Haynie, D., Crump, A. D., Eitel, P., & Saylor, K. (2001). Peer and parent influences on smoking and drinking among early adolescents. Health Education and Be- havior, 28(1), 29–49.

Simons-Morton, B. G., Haynie, D., Crump, A. D., Eitel, P., Saylor, K., & Yu, K. (1999). Expectancies and other psychosocial fac- tors associated with alcohol use among early adolescent boys and girls. Addictive Behaviors, 24(2), 229–238.

Snyder, H. N., & Sickmond, M. (1999). Juvenile offenders and vic- tims: 1999 National Report. Washington, DC: Office of Justice Programs, Office of Juvenile Justice and Delinquency Pre- vention, National Center for Juvenile Justice (also available at www.ncjrs.org/html/ojjdp/nationalreport99/toc.html).

Tobler, N. S. (1986). Meta-analysis of 143 adolescent drug pre- vention programs: Quantitative outcome results of program participants compared to a control or comparison group. The Journal of Drug Issues, 16(4), 537–567.

Tobler, N. S., & Statton, H. H. (1997). Effectiveness of school- based drug prevention programs: A meta-analysis of the re- search. Journal of Primary Prevention, 18, 71–128.

Vazsonyi, A. T., & Flannery, D. J. (1997). Early adolescent delin- quent behaviors: Associations with family and school do- mains. Journal of Early Adolescence, 17(3), 271–293.

Weisberg, R. P., Kumpfer, K. L., & Seligman, M. E. P. (2003). Pre- vention That Works for Children and Youth: An Introduc- tion. American Psychologist, 58 (6/7), 425–432.

Wenter, D. L., Ennett S. T., & Ribisl K. M. (2002). Compre- hensiveness of substance use prevention programs in U.S. middle schools. Journal of Adolescent Heatlh, 30(6), 455– 462.

View publication statsView publication stats

https://www.researchgate.net/publication/7668230