Programs that Work
Seattle Social Development Project
Healthy and Safe Children
Children Succeeding in School
Youths not using alcohol, tobacco, or illegal drugs
Students graduating from high school
Youths abstaining from sexual activity or not engaging in risky sexual behavior
Children and youth not engaging in violent behavior or displaying serious conduct problems
Children not experiencing anxiety or mood disorders, such as depression
Age of Child
Early Childhood (0-8)
Middle Childhood (9-12)
Type of Setting
Type of Service
Type of Outcome Addressed
Cognitive Development / School Performance
Substance Use and Dependence
Teen Sex / Pregnancy
Evidence Level (What does this mean?)
The Seattle Social Development Project (SSDP) was a multi-year, school-based intervention that used a risk-reduction and skill-development strategy to improve outcomes for participating children and youths. The program was guided theoretically by the social development model, which hypothesizes that youths who are provided with opportunities for greater involvement with their schools and families, who develop the competency or skills they need for fuller participation with their schools and families, and for whom skillful participation is constantly reinforced, ultimately develop strong bonds with their families and schools. Further, the model proposes that these strong bonds set children on a positive developmental trajectory, resulting in more positive outcomes and fewer health-risk behaviors later in life.
The SSDP was first implemented in 1981. It combined teacher, child, and parent components with the goal of enhancing children’s bonding with their families and schools. Teachers were trained in proactive classroom management, interactive teaching, and cooperative learning, while the students themselves were provided with direct instruction in interpersonal problem-solving skills and refusal skills to avoid problem behaviors. Parents were offered courses in child behavior management skills, academic support skills, and skills to reduce their children’s risk of drug use.
Program participants initially included students in grades one through six in eight Seattle, Washington, public schools. Students were from mixed socioeconomic and ethnic backgrounds, and males and females were equally represented. The study was subsequently expanded to include fifth- and sixth-grade students in ten additional Seattle public schools.
The SSDP has been implemented and evaluated in Seattle. Several evaluations of the program have been conducted that involve an ongoing longitudinal follow-up study of the original participants. The program was first implemented with a group of multiethnic urban students who entered the first grade in eight Seattle public schools in fall 1981. At that time, one whole school was assigned to be a control site, and one whole school was assigned to be an experimental site. In the remaining six schools, students entering the first grade and their teachers were randomly assigned to intervention or to control classrooms. Over the next three school years, in grades one through four, newly entering students were randomly assigned to intervention or to control classrooms and were added to the study panel. Of the 520 students remaining in the eight schools at the end of second grade, 458 completed the post-test assessments (285 from the treatment group and 173 from the control group).
In 1985, when the original subjects entered the fifth grade, the evaluation study was expanded to include all fifth-grade students in ten additional Seattle elementary schools. Schools added in the expansion group were matched to the original intervention schools with respect to the grades served and the percentage of students drawn from high-crime neighborhoods. Schools were then assigned to a treatment status nonrandomly. Beginning in fifth grade, the treatment groups participated in SSDP for two years. Therefore, the additions to the program resulted in three possible groups for analysis:
- students participating in SSDP from the first to sixth grades (full intervention),
- students participating in SSDP in the fifth and sixth grades only (late intervention),
- students in the control group, which was a combination of the control group participants from 1981 and those that were added to the study in 1985.
Both treatment-group and control-group SSDP teachers were observed periodically throughout the year to evaluate classroom performance and to determine the level and integrity of SSDP implementation.
An early study of SSDP by Hawkins, von Cleve, and Catalano (1991) assessed students in the original eight participating schools in the spring of the second grade. The sample included 285 treatment and 173 control students in 21 classes, representing an 88 percent participation rate. Parents of 122 (43 percent) of the 285 treatment-group subjects attended one or more parent-training classes. Comparisons between treatment and control groups indicated that there were no significant differences between the groups on demographic variables. However, pretest data indicated that control-group subjects were less at risk than were treatment-group subjects at baseline on eight of 28 self-reported items. Significant baseline differences favoring the control group were found on measures of students’ attachment to school, family communication, family supervision, and antisocial orientations. Treatment-group and control-group outcomes were derived from a teacher-reported assessment of the Achenbach Child Behavior Checklist (CBCL). The checklist included measures on behavior problems, ten measures for boys and ten for girls, including external antisocial measures (e.g., levels of aggression and inattentiveness) and internal antisocial measures (e.g., anxiety levels and extent of social withdrawal).
Hawkins et al. (1992) reported on data collected from 919 students during fall 1985 when normally progressing students entered the fifth grade. The intervention group consisted of students exposed to at least one semester of the full intervention in grades one through four (199 students). The control group consisted of students enrolled in control classrooms during grades one through four plus unexposed students who were added to the project in the fall of fifth grade when the study was expanded (709 students). Fifty-two percent of the student sample was male, 47 percent of the sample was Caucasian, 24 percent was African-American, 21 percent was Asian American, and 8 percent was of another racial/ethnic background. Thirty-eight percent of the students qualified for the National School Lunch Program in the fall of fifth grade. No significant differences were found between intervention and control students on these demographic variables. In addition, an analysis compared the initial control group with the additional available control students on each of the descriptive and outcome variables and found no significant overall differences. The outcomes assessed included family involvement and interaction, academic outcomes, alcohol use, and delinquency.
Another early study of SSDP was conducted by O’Donnell et al. (1995). It examined the impact of the full intervention (grades one through six) on a high-risk, low-income subsample of SSDP students. Students were identified as high risk based on their participation in the National School Lunch/School Breakfast Program during the fall of fifth grade. Intervention and control groups were matched on ethnicity, resulting in samples of 102 students in the intervention group and 75 students in the control group. Data were collected from students during fall 1985, when normally progressing students entered the fifth grade, and in spring 1987, when they completed the sixth grade. Students were administered a self-report survey, which measured peer interactions and risk-related behaviors. In addition, teachers completed the CBCL for each student, and student grades and outcomes on the California Achievement Test (CAT), a standardized test in math, reading, and language arts, were collected. At the completion of sixth grade, 43 percent of the intervention group (44 students) and 83 percent of the control group (62 students) completed the evaluation measures. Fifty-four percent of the sample was female, 42 percent was African-American, 24 percent was Caucasian, 25 percent was Asian American, and 9 percent was of another race/ethnicity. The intervention group had a higher percentage of female students, but there were no significant differences between groups in terms of race/ethnicity.
A study by Abbott et al. (1998) involved all consenting fifth-grade students in the 18 schools (808 students total). Fifty-one percent of the sample was male, 46 percent was Caucasian, 24 percent was African-American, 21 percent was Asian-American, and 9 percent was of another ethnicity. Fifty-two percent of the students were classified as low income on the basis of eligibility for the National School Lunch/School Breakfast Program in grades five, six, or seven, and 42 percent of the students lived in single-parent households. Analyses investigated the effects of the combined intervention groups (both full and late) on academic achievement and school bonding at the end of sixth grade.
The next evaluation of the effects of SSDP was published in 1999 by Hawkins and colleagues. Ninety-three percent of the original 643 students who consented to participate in the longitudinal study were interviewed at age 18. The study assessed the impact of the full intervention (156 students) and the late intervention (267 students) compared with the no-treatment control group (220 students). Rates of dropping out of the study were approximately the same across all three treatment conditions. The sample included approximately equal proportions of males and females, and more than 56 percent of the students had participated in their schools’ free-lunch program in grades five, six, and/or seven. Forty-five percent of the sample was Caucasian, 26 percent was African-American, 22 percent was Asian American, and 8 percent was of another race/ethnicity. There were no significant racial/ethnic, gender, or free-lunch eligibility differences among the full-intervention, late-intervention, and control populations. In 1993, participants were interviewed to gather information on a variety of school-related variables and health-risk behaviors. Standardized achievement test scores, grade-point averages, and school disciplinary records through age 17 were collected, as were delinquency charges from juvenile court.
An age-21 follow-up study of sexual behavior was conducted by Lonczak et al. (2002). Comparisons between the full-intervention group (144 students) and the control group (205 students), representing 93 percent of the available sample, were assessed. Outcomes for the late-intervention group and the parent-training-only group were not assessed in this study. Fifty-one percent of the sample was male, 47 percent was Caucasian, 26 percent was African-American, 21 percent was Asian-American, and 7 percent was of another race/ethnicity. Fifty-five percent of the sample had been eligible for the free lunch program between grades five and seven. In regard to marital status, 75 percent of respondents were single, 9 percent were married, 15 percent were living with a partner, and 2 percent were separated or divorced. Analyses of possible differences between students who dropped out of the program and those who remained in the program found no significant differences with regard to any of the demographic variables. Outcome measures included sexual activity, condom use, age of sexual onset, number of sexual partners, history of sexually transmitted diseases (STDs) other than HIV, history of pregnancy, and history of having a baby.
Finally, Hawkins et al. (2005) analyzed additional outcomes at age 21 for a sample of 605 respondents (94 percent of the longitudinal study population of 643). The authors compared the full-intervention (144 students), the late-intervention (256 students), and the no-treatment (205 students) groups. Fifty percent of the sample was female, 45 percent was Caucasian, 25 percent was African-American, 22 percent was Asian American, 6 percent was Native-American, and 3 percent was of another race/ethnicity. As children, 56 percent of participants were eligible for the school free lunch program at some point in grade five, six, or seven. Outcomes assessed included high school graduation rates, generalized anxiety, depression, suicidal thoughts, self-reported crime, arrests, drug selling, court charges, and substance use.
Key Evaluation Findings
At the end of the second grade, the study by Hawkins, von Cleve, and Catalano (1991) found the following for male students:
- The control-group students exhibited more externalizing antisocial behaviors, particularly aggressiveness, than did the SSDP group.
- No significant differences were found between groups for anxious, social withdrawal, unpopular, self-destructive, obsessive-compulsive, inattentive, nervous-overactive, or internalizing antisocial behavior measures.
- When outcomes for African-American boys were examined separately, no significant differences were found for them on any of the behavior measures. However, when outcomes for Caucasian boys (including 59 students in the treatment group and 40 in the control group) were examined separately, control subjects scored significantly higher (more poorly) on the aggressive and externalizing antisocial behavior measures than did the treatment-group students.
- Control-group students were rated as significantly more self-destructive than treatment-group students.
- No significant differences were found between groups on the anxious, social withdrawal, unpopular, depressive, inattentive, nervous-overactive, aggressive, externalizing, or internalizing behavior measures.
- Similar to the findings for males, when outcomes for African-American girls were examined separately, no significant differences were found between the treatment and control groups. For Caucasian girls (including 66 treatment and 41 control students), control students scored significantly higher on the self-destructive behavior measure than did treatment subjects. In addition, marginally significant outcomes were found for the depressive and nervous-overactive behavior measures.
After four years of intervention, Hawkins et al. (1992) found the following:
- Intervention students reported significantly lower rates of alcohol initiation than did control students (21 percent versus 27 percent).
- Intervention students reported significantly less delinquency initiation than did control-group students (46 percent versus 52 percent).
- Female control students scored significantly higher than intervention students on the CAT, but this difference faded by sixth grade.
- Compared with control-group boys, treatment-group boys
- were rated more socially competent by teachers.
- had significantly higher grades and scored significantly higher on the CAT.
- were not rated any differently on peer reports of antisocial behavior, were no less likely to run away from home, and were no less likely to have tried alcohol, cigarettes, or marijuana.
- Compared with their control counterparts, treatment-group girls
- were significantly less likely to have smoked cigarettes (7 percent versus 36 percent).
- were marginally significantly less likely to have tried alcohol (19 percent versus 39 percent) or marijuana (4 percent versus 17 percent).
- did not have significantly different scores on the CAT, did not show differences in their academic grades, were not rated any differently on peer or teacher reports of antisocial behavior, and were no less likely to run away from home.
- Students in the intervention classrooms had significantly higher CAT scores than did students in control classrooms.
- Compared with control-group participants, full-intervention (grades one through six) participants
- scored marginally higher in overall grade point average (2.42 versus 2.18).
- were significantly less likely to have repeated a grade (14 percent versus 23 percent).
- were marginally less likely to have an official high school disciplinary action report (46 percent versus 58 percent).
- were significantly less involved in school misbehavior, such as skipping class, cheating, or being removed from the classroom (3 percent versus 5 percent).
- were significantly less likely to have committed violent delinquent acts (48 percent versus 60 percent).
- were significantly less likely to have engaged in drinking ten or more times in the year prior to assessment (15 percent versus 25 percent).
- were significantly less likely to have engaged in sexual intercourse (72 percent versus 83 percent) or to have had multiple sexual partners (50 percent versus 62 percent).
- were significantly less likely to have been pregnant or to have caused a pregnancy (17 percent versus 26 percent).
- There were no significant differences between groups in their scores on the California Achievement Test, school dropout rates, or levels of suspension or expulsion.
- There were no significant differences between groups for lifetime rates of nonviolent delinquency, arrests, court charges, use of cigarettes, use of alcohol, use of marijuana, use of other drugs, or drinking and driving.
- Exposure to the late intervention (fifth and sixth grades only) resulted in only two impacts on long-term outcomes. Compared with control-group participants, late-intervention participants were
- significantly less likely to have engaged in school misbehavior (i.e., cheating, truancy, being removed from class for misbehavior).
- less likely to report sexual activity, although this was only marginally significant.
- A small but statistically significant effect was found for the average age at which those in the intervention group reported having had their first sexual experience versus the average age of those in the control group (16.3 years versus 15.8 years). A statistical analysis found that the program had a marginally significant effect in reducing the overall relative risk of engaging in sexual intercourse for the first time before age 21.
- No significant effects were found for the full sample for frequency of past-year condom use among single individuals.
- On average, young adults who had participated in the intervention group reported significantly fewer sexual partners in their lifetimes than did those in the control group. The difference between the intervention and control groups was especially pronounced for those reporting six or more partners (32 percent versus 43 percent).
- Participants in the full-intervention group were significantly more likely to report condom use during the most recent sexual intercourse than were those in the control group (60 percent versus 44 percent). There was no significant effect from the intervention on condom use during the first sexual intercourse.
- There was no significant overall treatment effect for diagnosis of STDs.
- Females in the intervention group were significantly less likely to become pregnant (38 percent versus 56 percent) and were significantly less likely to have had a baby (23 percent versus 40 percent) by age 21 than were females in the control group. The proportion of males in each group who reported causing a pregnancy or birth did not differ significantly.
- After controlling for the effects of poverty, statistically significant program effects were found for African-Americans.
- The difference in condom-use frequency between the intervention group and the control group was significantly greater for single African-Americans. Fifty percent of those in the intervention group reported always using a condom, compared with 12 percent of those in the control group.
- The difference in most-recent condom use between the intervention and control group was significantly greater for single African-Americans, with 79 percent of intervention-group participants and 36 percent of the control-group participants reporting condom use during the last intercourse.
- Among African-Americans, 7 percent of intervention-group participants and 34 percent of control-group participants reported being diagnosed with an STD over their lifetimes.
- Full-intervention participants were significantly more likely to have graduated from high school than were control-group participants (91 percent versus 81 percent). No significant differences were found between the late-intervention group and the control group on high school graduation rates.
- Both the full-intervention and late-intervention participants reported significantly fewer thoughts of suicide than did the control-group participants.
- Effects of the full intervention on reducing symptoms of depression were marginally significant when comparing full-intervention with control-group participants. No significant differences in symptoms of depression were found between the late-intervention group and the control group.
- Full-intervention participants reported significantly fewer symptoms of social phobia in the past year, but no significant differences were found between groups for anxiety.
- Both full-intervention participants and late-intervention participants were significantly less likely than those in the control group to have sold illegal drugs in the past year.
- Full-intervention participants were significantly less likely than those in the control group to have had a court charge in their lifetimes. No significant differences were found between the late-intervention group and the control group in the likelihood of having had a court charge.
- Full-intervention participants were marginally less likely than were control participants to have used alcohol, tobacco, or illicit drugs in the past month or year. No significant differences were found between the late-intervention group and the control group in the likelihood of using these substances in the past month or year.
- No significant effects were found among any of the groups for past-year crime, past-year arrests, or past-year court charges.
- Women who had participated in the full intervention had significantly fewer generalized anxiety symptoms than did women in the control group, while women in the late-intervention group reported marginally fewer generalized anxiety symptoms than women in the control group.
- Although the prevalence of having been arrested was low in both groups, women in the full-intervention group were significantly more likely to have been arrested in the past year than were women in the control group (7 percent versus 1 percent).
Public or private elementary schools
The SSDP was funded by the National Institute on Drug Abuse (Prevention Research Branch), the Office of Juvenile Justice and Delinquency Prevention, the Robert Wood Johnson Foundation, and the Burlington Northern Foundation.
Classroom teachers were trained in instructional methods with three major components: proactive classroom management, interactive teaching, and cooperative learning.
- Proactive classroom management is a technique through which teachers engage in positive management practices before problem behaviors, such as academic failure and conduct disorders, begin to develop. Teachers were trained to establish classroom routines at the beginning of the year to create a consistent pattern of expectations for students.
- SSDP teachers also used interactive teaching methods, which require mastery of specified learning objectives before proceeding on to new material, grading based on improvement over past performance, and frequent monitoring of students to assess progress.
- Finally, teachers used cooperative learning methods, which involve having small groups of students with differing abilities and backgrounds work together on curriculum material.
- During the first grade, students participated in the "Interpersonal Problem-Solving" curriculum, a program teaching communication skills, decision-making, negotiation, and conflict resolution.
- Sixth-grade students received a four-hour refusal-skills training program. The program’s goals were to help students recognize and resist social influences that may encourage problem behaviors (e.g., drug use or sex) and to develop positive alternatives to engaging in those negative behaviors.
- Parents of first- and second-grade students were able to participate in the seven-session "Catch ‘Em Being Good" curriculum. The curriculum teaches parents to effectively convey expectations for behavior, to identify and reinforce desirable behavior, and to address what would be appropriate responses to negative behavior.
- During the students’ third-grade year, parents could enroll in "How to Help Your Child Succeed in School," a four-session curriculum designed to improve parent-child communication and increase parental involvement in a child’s academic activities.
- The final component of parent education was offered during the fifth and sixth grades. "Preparing for the Drug (Free) Years" is a five-session, risk-focused, skills-training curriculum to help parents establish family policies, enhance family communication, and teach their children resistance skills.
The classroom program was implemented by regular classroom teachers. Five days of teacher training were sequenced throughout the school year, and teachers were regularly observed by SSDP staff and the school principal, who provided feedback on the use of the program’s techniques. In addition, teachers of first- and sixth-grade students received training in the specific curricula provided to students in those grades.
Issues to Consider
This program received a "promising" rating. The initial phase of the research involved random assignment of teachers and students in six schools to treatment or control groups and non-random assignment of two schools (at the whole-school level) to treatment or control conditions. In the second phase of the study, a quasi-experimental design was used, wherein ten additional schools were assigned to treatment or control groups non-randomly. The impact of this program was evaluated at multiple time periods, including evaluations while the intervention was ongoing (at ages 7, 8, and 10), soon after the end of the intervention (age 13), and at six- and nine-year follow-ups (ages 18 and 21). Each evaluation period yielded significantly positive results, although the findings were not always consistent among time periods, nor were they significant for all measures. Findings sometimes differed by gender or race. There is suggestive evidence that the "full intervention," which treated children from first through sixth grade, may be more effective than the "late intervention," which takes place only in fifth and sixth grade.
The majority of research on this program is based on the quasi-experimental portion of the study. Several evaluation reports have examined whether there is evidence of bias due to the non-random group assignment of schools. These found similar observable characteristics of the treatment and control groups (gender, ethnicity, childhood poverty, residential stability, etc.). For a discussion of potential threats to validity due to the quasi-experimental design, see Hawkins and Catalano (1995).
Some consideration should be given to the impact of attrition (dropouts from the study). There was a high rate of attrition in the O’Donnell et al. (1995) study, with only 60 percent of the original sample included in the evaluation at the completion of the sixth grade. In addition, in the O’Donnell analysis of sixth-grade data, the intervention group had nearly twice the attrition level of the control population. Attrition analysis indicated that students for whom sixth-grade data were unavailable were less attached to their peers, had more suspensions from school, and were less committed to school. However, with one exception (girls’ opportunities for antisocial behaviors, such as delinquency and drug use), there was no evidence based on observable characteristics of differential attrition among the study groups. In other words, those students who dropped out of the intervention and control groups differed little on baseline characteristics.
Another issue with regard to attrition is the response rate for the age-21 follow-up studies, which analyzed outcomes for 144 full-intervention youths. There has been debate in the literature about the appropriate baseline denominators to use for attrition analysis. A 93 percent follow-up rate was calculated using as a baseline the 156 fifth-graders who had parental consent to participate in the follow-up study. However, Gorman (2002) suggests that because involvement of these subjects in the study began four years earlier, in 1981, the correct follow-up rate for the full-intervention group at age 21 years is no more than 50 percent (using a baseline of 285 second grade respondents). Hawkins and Catalano (2005), in reply to Gorman, argue that the sample for the longitudinal quasi-experimental study was constituted in grade five and takes advantage of including those students from the earlier experimental study who remained in project schools.
Two studies assessed the effects of the intervention on "school bonding," that is, the development of a positive emotional link and personal commitment to school (Abbott et al., 1998; Hawkins et al., 2001). Previous research has shown that school bonding can be a protective factor against such negative outcomes as school dropout, delinquency, violence, and drug abuse and can be a mediating factor in improving such outcomes as academic achievement (e.g., see Hirschi, 1969; Catalano and Hawkins, 1996). While Abbott and colleagues did not find any significant differences among the groups, Hawkins et al. (2001) found that by age 18, the level of bonding to school in the full-intervention group was significantly higher than in the control group. This result remained significant after controlling for gender, ethnicity, poverty, and earlier academic achievement. No significant differences were found between the late-intervention group and the control group, suggesting that positive impacts on school bonding are attained only by program implementation prior to fifth grade.
Although the parental component was considered to be an integral and vital part of the program and one of the foundations on which the theoretical model was built, parent participation in the program was not a mandatory component. Research by Hawkins et al. (1999) showed that a relatively low percentage (43 percent) of parents attended the parenting workshops. No measure was in place to determine how well skills and information were disseminated in the parental workshops, or if those skills and information were used in the home. However, Hawkins et al. (1992) found that children in the full-intervention condition reported significantly better family management and family communication, as well as more family involvement and attachment to family, than did control-group children. While this finding suggests that the parenting component may have a positive impact, the evaluations do not indicate the extent to which the parent component may have played a part in generating the impact that was observed.
Finally, when weighing the program evaluation outcomes, it is important to keep in mind that the program has been studied in only one urban metropolitan area, thus limiting the applicability of the findings to other populations. In addition, all evaluations of the program to date have been conducted by the program’s designers rather than outside evaluators.
Seattle and Edmonds, Washington
For information on the Seattle Social Development Project, contact:
Karl G. Hill, PhD
Project Director, Seattle Social Development Project
Social Development Research Group
University of Washington
9725 3rd Avenue NE, Suite 401
Seattle, WA 98115-2024
Phone: (206) 685-3859
Fax: (206) 543-4506
Web site: www.depts.washington.edu/ssdp
The Social Development Research Group (SDRG) provides guides for the teacher training components of the SSDP intervention as well as trainers for these components. Two of the parent programs included in SSDP, now called Preparing for School Success (grades one through three) and Guiding Good Choices (grades 5 and 6), are available from the Channing Bete Company, www.channing-bete.com. Contacts for trainers for these components are provided by SDRG.
The Social Development Research Group summarizes the Seattle Social Development Project at http://depts.washington.edu/ssdp/.
Abbott, Robert D., Julie O'Donnell, J. David Hawkins, Karl G. Hill, Richard Kosterman, and Richard F. Catalano, "Changing Teaching Practices to Promote Achievement and Bonding to School,"
American Journal of Orthopsychiatry,
Vol. 68, No. 4, 1998, pp. 542-552.
Catalano, Richard F., and J. David Hawkins, "The Social Development Model: A Theory of Antisocial Behavior," in J. David Hawkins, ed., Delinquency and Crime: Current Theories, Cambridge, England: University Press, 1996.
Gorman, Dennis M., "Overstating the Behavioral Effects of the Seattle Social Development Project," Archives of Pediatric and Adolescent Medicine, Vol. 156, 2002, pp. 155-156.
Hawkins, J. David, and Richard F. Catalano, "Doing Prevention Science: A Response to Dennis M. Gorman and a Brief History of the Quasi-Experimental Study Nested Within the Seattle Social Development Project," Journal of Experimental Criminology, Vol. 1, 2005, pp. 79-86.
Hawkins, J. David, Elizabeth von Cleve, and Richard F. Catalano, "Reducing Early Childhood Aggression: Results of a Primary Prevention Program," Journal of the American Academy of Child & Adolescent Psychiatry, Vol. 30, No. 2, 1991, pp. 208-217.
Hawkins, J. David, Jie Guo, Karl G. Hill, Sara Battin-Pearson, and Robert D. Abbott, "Long-Term Effects of the Seattle Social Development Intervention on School Bonding Trajectories," Applied Developmental Science: Special Issue: Prevention as Altering the Course of Development, Vol. 5, No. 4, 2001, pp. 225-236.
Hawkins, J. David, Richard F. Catalano, Dianne M. Morrison, Julie O'Donnell, Robert D. Abbott, and L. Edward Day, "The Seattle Social Development Project: Effects of the First Four Years on Protective Factors and Problem Behaviors," in Joan McCord and Richard E. Tremblay, eds., Preventing Antisocial Behavior: Interventions from Birth Through Adolescence, New York, N.Y.: Guilford Press, 1992, pp. 139-161.
Hawkins, J. David, Richard F. Catalano, Rick Kosterman, Robert Abbott, and Karl G. Hill, "Preventing Adolescent Health-Risk Behaviors by Strengthening Protection During Childhood," Archives of Pediatric Medicine, Vol. 153, 1999, pp. 226-234.
Hawkins, J. David, Rick Kosterman, Richard F. Catalano, Karl G. Hill, and Robert D. Abbott, "Promoting Positive Adult Functioning Through Social Development Intervention in Childhood: Long-Term Effects from the Seattle Social Development Project," Archives of Pediatrics & Adolescent Medicine, Vol. 159, No. 1, 2005, pp. 25-31.
Hirschi, Travis, Causes of Delinquency, Berkeley, Calif.: University of California Press, 1969.
Lonczak, Heather S., Robert D. Abbott, J. David Hawkins, Rick Kosterman, and Richard F. Catalano, "Effects of the Seattle Social Development Project on Sexual Behavior, Pregnancy, Birth, and Sexually Transmitted Disease Outcomes by Age 21 Years," Archives of Pediatrics & Adolescent Medicine, Vol. 156, No. 5, 2002, pp. 438-447.
O'Donnell, Julie, J. David Hawkins, Richard F. Catalano, Robert D. Abbott, and Edward Day, "Preventing School Failure, Drug Use, and Delinquency Among Low-Income Children: Long-Term Intervention in Elementary Schools," American Journal of Orthopsychiatry, Vol. 65, No. 1, 1995, pp. 87-100.