Sep 14, 2010


All Stars Character Education and Problem Behavior Prevention Program is a school-based intervention program designed to reduce adolescents’ engagement in risk behaviors such as substance use, violence and sexual activity. The program is intended for middle school students and consists of a curriculum that uses class debates, games, small group activities and individual meetings. In this evaluation, schools were randomly assigned to one of three conditions: control schools, schools where All Stars is implemented by regular classroom teachers or schools were All Stars is implemented by outside specialists. An evaluation of the program has found that, ultimately, when taught by regular teachers, All Stars has a modest but significant impact reducing levels of alcohol, cigarette, and inhalant use, but program impacts disappeared by one year.

In another evaluation, the All Stars program had short-term impacts on bonding, commitment, ideals, and normative beliefs among white students. These impacts were delayed for Hispanic and African-American students. These impacts disappeared after one year.


Target population: Middle school students

All Stars is based upon the notion that certain mediating variables affect the likelihood of a teen engaging in risky behaviors. The goal of All Stars is to reinforce positive qualities, specifically: positive ideals and future aspirations, positive norms, personal commitments, school and community organization bonds, and parental attentiveness. Many All Stars program elements have been used previously in other programs such as the Adolescent Alcohol Prevention Trial and Project STAR.

The All Stars curriculum is comprised of 9 to 13 lessons during its first year and 7 to 8 booster lessons in its second year, each of which is designed to affect one of the mediators mentioned above. These sessions include a variety of interactive activities such as class debates and games. Students are also assigned homework to reinforce lessons learned as well as involve parents at home. Four of the sessions are spent in small groups outside of class time and four of them are spent in one-on-one meetings with a teacher and student. These meetings are intended to help students become more integrated into the school and increase the level of student-school attachment. All Stars is available for delivery as part of regular classroom instruction or in after-school and community-based organizations and programs.

Two versions of All Stars are currently being implemented. In one version, specialists are recruited to implement the program. These specialists are selected based on previous experience with teaching, prevention programming and personal characteristics. All specialists take part in two days of training and receive weekly technical assistance from program staff. Many of the specialists teach more than one All Stars class and work in multiple schools. The other version of All Stars is taught by regularly employed teachers. These teachers are selected if their school offered to participate in the All Stars program. Unlike the outside specialists, teachers work in teams and only teach one section of students. Teachers receive one day of training and receive little technical assistance throughout the program. The content of the curriculum, however, is identical in both versions of the program.

The cost of teacher materials range from $125 to $540. The cost of student materials range from $45 to $145. Two day training costs $250 per person or $3,000 per group. Other levels of training cost $125 per person or $1,5000 per group. Information on materials and training are available here:


Harrington, N. G., Giles, S. M., Hoyle, R. H., Feeney, G. J., & Yungbluth, S. C. (2001). Evaluation of the All Stars Character Education and Problem Behavior Prevention Program: Effects on mediator and outcome variables for middle school students. Health Education & Behavior, 28(5), 533-546.

Evaluated population: 1,655 students from 14 middle schools in large cities in a Midwestern state. The range of ages of students at pre-test was 11 to 13 years and there were 55% female, 69% white, 25% African American, and 6% Hispanic subjects.

Approach: The schools were randomly assigned to one of three conditions. Five schools were assigned to receive the All Stars curriculum taught by seven specialists hired from the community. These specialists had prior teaching experience and received 30 hours of training on the All Stars program. Three schools (23 teachers) were later assigned to receive the All Stars curriculum taught by their regular classroom teachers. These teachers also received All Stars training; however, it was condensed into a half day. Six schools were assigned to the control condition. One treatment school out of an original six schools was dropped due to implementation problems.

Students completed an evaluation survey including questions on demographics, substance use, sexual behavior, violence, bonding, commitment, ideals, and perceived norms. The All Stars curriculum attempts to help students recognize that drug use, sex, and violence can interfere with a desired lifestyle, increase beliefs about peer norms regarding drugs, sex, and violence, have students make a commitment to avoid drugs, sex, and violence, and have students increase their attachment and acceptance at school. The program operated through the school year, in and out of the classroom, and in group and one-on-one sessions. Debates, games, discussion and homework are used to increase interaction between students and allow parents to have a role in the students’ lives.

Results: Among white students in teacher-delivered groups, the All Stars program was shown to have immediate but short-term impacts on bonding (a small effect size of 0.07), commitment (a small effect size of 0.07), ideals (a small effect size of 0.09), and normative beliefs. Similar results occurred among Hispanic students in specialist-delivered groups, but these impacts were delayed. African American students showed delayed impacts in both teacher- and specialist-delivered groups. African American and Hispanic students showed delayed developmental increases in violence in teacher-delivered groups. Program impacts disappeared at the 1-year follow-up. School-level HLM analyses indicated that school effects on the outcome variables were minimal.

McNeal, R., Hansen, W., Harrington, N. & Giles, S. (2004). How All Stars works: An examination of program effects on mediating variables. Health Education and Behavior, 31(2), 165-178.

Evaluated population: 1,822 middle school students in Lexington and Louisville, Kentucky participated in the evaluation. This number represents approximately 80 percent of the original sample of students who provided baseline data.

Approach: The 1,822 students came from fourteen different middle schools. The schools were randomly assigned to one of three conditions. Five schools received the version taught by specialists while three schools were taught by regular teachers. Six schools were assigned to the control condition and received their regular health education classes.

Of those students that participated in the evaluation, 54 percent were female, 69 percent were white, 23 percent were African American, and 98 percent were between the ages of eleven and thirteen.

Students were surveyed before and after the school year in which All Stars was implemented. Surveys measured demographic characteristics such as gender, age, and ethnicity as well as frequency and prevalence of risk behaviors. Risk behaviors included lifetime, past-year and past 30-day use of cigarettes, marijuana, alcohol, and inhalants. It also included measures of heterosexual intimate behaviors and sexual activity. Finally, the surveys included the mediating variables specifically targeted (included in the curriculum) by the program and believed to affect risk behaviors. These mediators were measured by the following scales: Normative Beliefs About Risky Behaviors, Lifestyle Incongruence (which measured the perceived disconnect between drug use and desired lifestyle), and Manifest Commitment to Avoid Risky Behaviors and Bonding to School.

Results: When evaluators looked at pretest-posttest changes among all three conditions, the results were as follows. Among students in the control group, 30-day prevalence rates revealed that: alcohol use increased from 9.1-15.0 percent; cigarette use increased from 15.1-17.9 percent; marijuana use increased from 5.0-8.7 percent; and sexual activity increased from 9.7-11.3 percent.

In comparison, schools in the specialist and teacher conditions reported the following rates: alcohol increased from 4.4-9.2 percent among specialist groups, 6.7-11.8 percent among teacher groups; cigarette use increased from 11.0-13.8 percent among specialist groups, 7.4-7.8 percent among teacher groups; marijuana use increased from 3.2-4.1 percent among specialist groups, but remained steady at 3.2 percent among teacher groups; sexual activity increased from 5.5-6.9 percent among specialist groups, 4.0-5.5 percent among teacher groups; while inhalant use decreased from 2.2-1.9 percent among specialist groups, 3.1-2.9 percent among teacher groups.

The largest impacts were found when All Stars was delivered by regular classroom teachers. Overall, students in the teacher classrooms reported significantly lower levels of alcohol, cigarette and inhalant use than students in control schools (a small effect size of 0.37). Results were not significant, however, for marijuana use or sexual activity among this group. Students who participated in the specialists’ version did not report any significant impacts when compared to control students.

When assessing the program impact on mediating variables, the teacher version had a greater impact than the specialist version. In particular, the teacher version had a significant impact altering lifestyle incongruence, manifest commitment and sensation-seeking. The specialist version had a significant impact on only manifest commitment. Overall, All Stars was moderately successful at changing mediators that affect risky behaviors.

Non-experimental analyses indicate that those students committed to avoid risky behaviors were less likely to engage in marijuana, cigarette and alcohol use posttest. They also were less likely to engage in sexual activity at posttest. Similarly, those students who believed that drug use was not associated with attaining their desired lifestyle were less likely to use alcohol, marijuana and cigarettes, or engage in sexual activity at posttest. This mediator, identified as lifestyle incongruence, was a slightly stronger predictor than normative beliefs. On the whole, targeted mediators, like those directly incorporated into the All Stars curriculum, were found to be more effective than nontargeted mediators in altering substance use behaviors.

Although All Stars experienced some degree of attrition, this caused no statistically significant differences between conditions throughout the course of the study. In cases when students did leave the program, they tended to be older, minority males. There were some statistically significant differences, however, between the control and treatment groups at baseline. On measures of age, minority status, alcohol use, sexual activity, normative beliefs and manifest commitment, the treatment groups ranked higher on every risk variable. Evaluators point out, however, that differences such as these only make the significant impacts reported for All Stars all the more impressive.


Link to program curriculum:


McNeal, R., Hansen, W., Harrington, N. & Giles, S. (2004). How All Stars works: An examination of program effects on mediating variables. Health Education and Behavior, 31(2), 165-178.

School-based, Adolescence (12-17), Substance Use, Alcohol Use, Tobacco Use, Illicit Drugs, Reproductive Health, Sexual Initiation, Education, Middle School, School Engagement, Parent or Family Component, Social and Emotional Health and Development, Physical Health.

Program information last updated 9/14/10