Programs that Work
Midwestern Prevention Project/Project STAR
Healthy and Safe Children
Youths not using alcohol, tobacco, or illegal drugs
Age of Child
Middle Childhood (9-12)
Type of Setting
Type of Service
Type of Outcome Addressed
Substance Use and Dependence
Evidence Level (What does this mean?)
The Midwestern Prevention Project (MPP) is a comprehensive, community-based prevention program that is intended to reduce the use of cigarettes, alcohol, and marijuana among adolescents. MPP began in 1984 in Kansas City, Kansas and Missouri, where it is known as Project STAR (Students Taught Awareness and Resistance). The program was later implemented in Indianapolis and Marion County, Indiana.
The primary component of MPP is a school program offered to sixth and seventh graders, supported by four other components: a parent program, a community organization program, a program aimed at changing local health policy, and mass-media events. About 30 mass-media events, ranging from news articles to conferences, occurred every year. Further explanation of the program components is available in the Implementation Detail section below.
The program's components are designed to address the multiple influences that adolescents face related to drug use. On the demand side, the program tries to change behavior through teaching adolescents resistance skills in the school program. On the supply side, the program tries to change the environment by involving the entire community in drug-prevention activities. The different components of Project STAR are introduced sequentially over the course of several years, beginning with the school program.
Currently, program replication materials are not available for purchase.
The program targets all students within middle schools, with the intention of preventing drug abuse during the early adolescent risk period.
Several evaluations draw from the sample of schools participating in Project STAR when it was first implemented with a cohort of sixth and seventh graders in 1984 and 1985. These students attended a total of 50 middle and junior high public schools in the Kansas City area. Of the 50 schools, 16 had scheduling flexibility that allowed them to be randomly assigned to serve as a control school or to receive treatment. Half of these 16 schools were randomly assigned to receive the intervention. Thirty-four schools were not able to accommodate random assignment. Of these, 15 schools implemented the intervention and 19 schools served as controls. Between 1985 and 1986, six schools closed and two missed data collection times, leaving 42 schools for which data was collected at baseline. This left 24 schools that implemented the MPP intervention and 18 schools that served as a comparison control group. Students in all schools were potentially exposed to the multimedia events and to community leader organization efforts, because they were implemented throughout the Kansas City area.
All of the Kansas City evaluations relied on students' self-reported drug use through 133-item questionnaires, which asked about alcohol, cigarette, and marijuana use; demographic characteristics; and psychosocial variables. Questionnaires were administered in Kansas City at baseline—either sixth or seventh grade in 1984—and annually thereafter until 1987. A carbon monoxide breath kit was used at each questionnaire administration to verify accuracy of self-reported drug use.
In eight of the 42 Kansas City schools, questionnaires were collected from all students in the relevant grade cohort each year. In the remaining 34 schools, a random sample of 25 percent of the relevant grade cohort was collected each year, but this sample did not consist of the same students each year. No significant differences were found between program and control schools for lifetime drug use or demographic characteristics at baseline. Based on the eight schools in which all of the students within the relevant grade level were tracked over time, no significant differences in the number of students who failed to complete questionnaires at one year follow-up were found between control and program schools.
Using schools as the unit of analysis, Pentz et al. (1989a) examined overall program effects in all 42 Kansas City schools after one year, controlling for urbanicity, socioeconomic status, and race. By 1985, one year after baseline, mass media and the 10-week classroom and homework components of MPP had been implemented. Program effects were measured using the difference between changes in drug use observed in the program schools and in the control schools, one year after baseline measurements. Because all students across experimental conditions had been exposed to media events, the measured differences capture the effect of only the classroom and homework components of MPP.
Pentz et al. (1989c) studied the effects of MPP on cigarette smoking two years after Kansas City students began the program in middle school or junior high. Using district feeder patterns, 47 high schools were merged with baseline schools to test program effects after two years. By this time, students in the intervention had received 10 sessions of the classroom program plus homework, a five-session booster classroom program plus homework, parent training program for parents, and 70 media items and events. Students who were not in the MPP schools were also exposed to the MPP mass media component. Program effects on the proportion of students at each school who smoked cigarettes monthly, weekly, and daily were measured.
Pentz et al. (1989d) examined the effects of MPP on cigarette smoking in the eight panel schools at six months, one year, and two years after baseline. A total of 1,122 students in the panel schools were surveyed at all time points. By two years after baseline, the MPP intervention had delivered the 10 classroom sessions, five booster sessions, parent training and parent organization, and media campaigns. In the eight panel schools, the control group had a significantly higher number of students who had ever smoked than the intervention group.
Pentz et al. (1989b) tested whether MPP had an effect on cigarette smoking, alcohol use, or marijuana use in students attending the 42 study schools at six months, one year and two years after baseline. Substance use was measured as the change in the percentage of students who reported using each substance in the past week and past month.
MacKinnon et al. (1991) used questionnaire data collected at baseline (1984) and at one year follow-up (1985) from the original 42 Kansas City schools to identify program impacts on factors mediating drug use, such as intentions to use, beliefs about positive and negative consequences of drug use, communication skills, and friends' reactions to drug use. Tests of baseline characteristics showed that students in program schools had greater resistance skills and stronger beliefs about negative consequences of marijuana use than students in control schools, indicating that the two groups of students were not entirely comparable before they received treatment.
Two more recent studies have looked at specific aspects of the program in the Indiana replication. Fifty-seven schools located within 12 districts in Marion County, Indiana were in the study. In the fall of 1986, 32 schools were randomly assigned to use the MPP program for sixth and seventh graders and 25 schools were randomly assigned to serve as a comparison group. Treatment and comparison group sizes are unequal because some districts did not include an even number of schools. In these districts, the odd school was included in the treatment group. Classroom and media components of MPP began in the treatment group schools during fall 1987, while the remaining three components—parent education, policy change, and community organization—began in the 1988-1989 school year. All schools received the parent component in 1988-1989, regardless of whether the schools were in the treatment or control group. In addition, control schools began receiving the school-based program in the 1988-1989 school year. Thus, any student entering sixth or seventh grade in 1986-1987 would not go through the MPP curriculum, a student entering sixth or seventh grade in 1987-1988 within a treatment group school would go through the MPP curriculum, and any student entering sixth or seventh grade in any of the 57 schools in 1988-1989 would go through the MPP curriculum.
In a study of the Indiana replication, Chou et al. (1998) tested whether the Prevention Project reduced use of cigarettes, alcohol, and marijuana among students who were already using one or more of these substances. Data were collected from a random sample of classrooms representing one-third of all students across the 57 schools, using a 100-item questionnaire that measured substance use, demographics, attitudes, and social influences. Students were tracked for four follow-up periods: six months after baseline, one and a half years after baseline, two and a half years after baseline, and three and a half years after baseline. This study examined whether students in the program showed greater decreases than control students in their substance use over time. Effects were measured by comparing an individual's reported substance use at each baseline to reported use at each follow-up. The researchers compared changes in self-reported substance use over time among the control and program students. It should be noted that control students who reported alcohol use at baseline failed to complete questionnaires at higher rates than program students at the second follow-up.
Chou et al. (2004) examined the effects of MPP on cigarette use among students in the Indianapolis study schools. This study compared the growth in cigarette use of students who smoked cigarettes in the past month between the MPP intervention and control groups. Measurements were taken at baseline, six months later, and then annually for five years.
Key Evaluation Findings
Pentz et al. (1989a) found that:
- While prevalence of substance use had increased for both groups at the one-year follow-up, the increase in proportion reporting past month substance use among students exposed to the MPP program was lower than for students not exposed. The reported increases in past month use was as follows: cigarettes (15% vs. 22%), alcohol (9% vs. 12%) and marijuana (4% vs. 7%).
- Preliminary results from students who were tracked over time suggest program effects were maintained at the two-year follow-up questionnaire.
The Pentz et al. (1989c) study examined the proportion of monthly, weekly, and daily cigarette smokers and found that:
- The rate of increase in smoking at schools not in the MPP program was about 1.5 times the rate of increase in program schools. For example, after two years, students in control schools had increased their monthly smoking by 15.4 percent, compared with a 9.1 percent increase among students in program schools.
Pentz et al. (1989d) study of the eight panel schools found that:
- Significantly fewer students in the MPP intervention group reported smoking cigarettes in the past month than control students at six months, one year, and two years after baseline.
- Significantly fewer students in the MPP intervention group reported smoking cigarettes in the past week than control students at six months and one year after baseline; only marginally significantly fewer MPP intervention students reported smoking in the past week at two years.
- At two years after the MPP intervention, students in the control group were 1.4 times more likely to report having ever smoked in their lifetime, 1.7 times more likely to report smoking in the past month, and 1.6 times more likely to report smoking in the past week than intervention students.
Pentz et al. (1989b) found that:
- Students in the MPP intervention had a significantly lower increase in reported cigarette use in the past month and in the past week than students in the control group.
- Students in the MPP intervention had a significantly lower increase in reported alcohol use in the past month and in the past week.
- Students in the MPP intervention had a significantly lower increase in reported marijuana use in the past month and in the past week.
The MacKinnon et al. (1991) study reported that:
- Significantly fewer program students reported an intention to use cigarettes, alcohol, and marijuana within the next two months.
- Program students were significantly less likely to believe that there were positive consequences from drug use.
- Program students were more likely to report that their friends had negative views toward drug use.
In their study of substance use behavior among Indiana students already using cigarettes, alcohol, and/or marijuana, Chou et al. (1998) found that:
- Program students who reported smoking at baseline showed small but significant decreases in their use after six months. However, no significant effects were found at the next three follow-up tests.
- Program students who reported drinking alcohol at baseline showed small but significant decreases in their use after six months, small decreases again at 1.5 years, and no significant changes at the last two follow-up tests.
- Program students who reported marijuana use at baseline did not significantly differ in their use from non-program students at any follow-up.
Chou et al.'s (2004) study of past-month cigarette use among students in Indianapolis over six follow-up time periods found that:
- Students in the MPP intervention group showed a lower rate of growth in past-month cigarette use during junior high than control group students. This lower rate was only marginally statistically significant.
Middle and junior high schools
National Institute on Drug Abuse, E. M. Kauffman Foundation
- School Program. This is a two-year program presented to students in the transition year to middle or junior high school (typically sixth or seventh grade). The first year of the school program consists of about 10 classroom sessions that focus on increasing skills to resist pressures to use drugs. A five-session booster program follows in the second year. Peer leaders help teachers facilitate the program. Methods include group discussion, role-playing, and homework to be completed with the family.
- Parent Program. This is implemented in the second year of the program and includes workshops on parenting skills and neighborhood activities, with some parents becoming involved in parent committees to organize and plan prevention activities.
- Community Organization. In the third year, the program brings together and trains community and government leaders to plan and implement drug-prevention services for youth.
- Policy Program. With the relevant leaders from the third component, the goal of this component is to change local ordinances to further restrict students' access to cigarettes, alcohol, and marijuana and to support prevention efforts.
- Media Program. An ongoing component of the multiyear effort, this component uses television, radio, and newspapers to introduce the program to the community, inform the community about its activities, and reinforce the program components.
Regular school teachers are trained in a two-day workshop for the school program, followed by annual half-day training sessions. Principals and volunteer parents receive training during a one-day workshop to become leaders for the parent program component. The training for the community leaders takes place over several sessions.
Issues to Consider
This program received a "proven" rating. The program was studied in two locations, with thousands of sixth and seventh graders who were assigned to program and control groups. Several separate studies have shown significant effects on several positive outcomes over differing lengths of time.
One potential concern regarding the methodology used in the Kansas City area is that only eight of the 42 schools were randomly assigned to either the program or control group. The others were assigned based on the flexibility of the school to incorporate the program. Although a random-assignment study is ideal, control and program groups were observably similar in student demographics and in grade level-adjusted student drug use rates at baseline. This limitation does not apply to the schools in Indiana. In Indiana, all 57 schools in the study participated in either the treatment or control group because they were randomly assigned to that group.
It should also be noted that the program developer has been involved in every MPP evaluation which met PPN criteria.
Kansas City, Missouri; Kansas City, Kansas; and Indianapolis, Indiana
For more information about MPP/Project STAR, contact:
Mary Ann Pentz, Ph.D.
USC Norris Comprehensive Cancer Center
University of Southern California
1441 Eastlake Avenue, MS-44
Los Angeles, CA 90089-9175
Phone: (626) 457-6691
Fax: (626) 457-4012
USC Department of Preventive Medicine
Division of Health Behavior Research
1000 South Fremont Ave., HAS 5235
Alhambra, CA 91803
Phone: (626) 457-6600
Fax: (626) 457-4012
A Blueprints for Violence Prevention monograph on MPP with information on implementing the program is available from the Center for the Study and Prevention of Violence at the University of Colorado, Boulder. An overview of the project is available at http://www.colorado.edu/cspv/blueprints/modelprograms/MPP.html, and the full book may be purchased for $12 (in 2011 dollars).
Chou, Chih-Ping, Dongyun Yang, Mary Ann Pentz, and Yih-Ing Hser, "Piecewise Growth Curve Modeling Approach for Longitudinal Prevention Study,"
Computational Statistics & Data Analysis,
Vol. 46, 2004, pp. 213-225.
Chou, Chih-Ping, Susan Montgomery, Mary Ann Pentz, Louise Rohrbach, C. Anderson Johnson, Brian R. Flay, and David P. MacKinnon, "Effects of a Community-Based Prevention Program on Decreasing Drug Use in High-Risk Adolescents," American Journal of Public Health, Vol. 88, No. 6, 1998, pp. 944-948.
MacKinnon, David P., C. Anaderson Johnson, Mary Ann Pentz, James H. Dwyer, William B. Hansen, Brian R. Flay, and Eric Yu-I Wang, "Mediating Mechanisms in a School-Based Drug Prevention Program: First-Year Effects of the Midwestern Prevention Project," Health Psychology, Vol. 10, No. 3, 1991, pp. 164-172.
Pentz, Mary Ann, C. Anderson Johnson, James H. Dwyer, David M. MacKinnon, William B. Hansen, and Brian R. Flay, "A Comprehensive Community Approach to Adolescent Drug Abuse Prevention: Effects on Cardiovascular Disease Risk Behaviors," Annals of Medicine, Vol. 21, 1989b, pp. 219-222.
Pentz, Mary Ann, David P. MacKinnon, Brian R. Flay, William B. Hansen, C. Anderson Johnson, and James R. Dwyer, "Primary Prevention of Chronic Diseases in Adolescents: Effects of the Midwestern Prevention Project on Tobacco Use," American Journal of Epidemiology, Vol. 130, No. 4, 1989d, pp. 713-724.
Pentz, Mary Ann, David. P. MacKinnon, James H. Dwyer, Eric Y. I. Wang, William B. Hansen, Brian R. Flay, and C. Anderson Johnson, "Longitudinal Effects of the Midwestern Prevention Project on Regular and Experimental Smoking in Adolescents," Preventive Medicine, Vol. 18, 1989c, pp. 304-321.
Pentz, Mary Ann, James H. Dwyer, David P. MacKinnon, Brian R. Flay, William B. Hansen, Eric Yu I. Wang, C. Anderson Johnson "A Multi-Community Trial for Primary Prevention of Adolescent Drug Abuse: Effects on Drug Use Prevalence," Journal of the American Medical Association, Vol. 261, 1989a, pp. 3259-3266.