Programs that Work
Second Step Violence Prevention
Healthy and Safe Children
Children and youth not engaging in violent behavior or displaying serious conduct problems
Age of Child
Early Childhood (0-8)
Middle Childhood (9-12)
Type of Setting
Child Care / Preschool
Type of Service
Type of Outcome Addressed
Evidence Level (What does this mean?)
Second Step: A Violence Prevention Program is a classroom-based social skills curriculum for students from preschool through middle school. The curriculum aims to reduce impulsive and aggressive behaviors and increase protective factors and social-emotional competence. Organized by grade level, the program teaches children empathy, problem-solving skills, risk assessment, decision-making, and goal-setting skills. The Second Step program is classified as a universal intervention, meaning that it is appropriate for whole classrooms of children and not just those at risk.
Second Step lessons are organized into three skill-building units that focus on empathy, impulse control and problem solving, and anger management. Lessons are sequential, developmentally appropriate, and provide opportunities for modeling, practice, and skills reinforcement. The curriculum includes discussion, teacher modeling, coaching skills, and role-plays. Stories are used to demonstrate important peer-relations skills and to teach affective (emotional), cognitive, and behavioral social skills. Lessons can be incorporated into health, science, math, social studies, and language arts.
The Second Step Family Guide is available as a supplement to the program for preschool-grade 5. The six-session, facilitator-led program helps parents learn about the curriculum and assists them in reinforcing childrenís skills in communicating feelings, solving problems, controlling anger, and dealing with conflict.
Segundo Paso, a Spanish translation supplement of the Second Step curriculum, is available for preschool through grade five classroom kits. A Spanish translation of the Second Step Family Guide is also available.
Children in preschool through middle school (4Ė14 years of age)
Orpinas et al. (1995) evaluated the effects of Second Step in a sample of 223 sixth graders from ten classes in four middle schools in a large urban school district in Texas. The school principals were asked to choose "good and interested" teachers to teach the curriculum, as well as a control class that was similar to the intervention classes. Four of the classrooms implemented the Second Step program two to three times a week, and six of the classrooms served as the control group. In two of the four intervention schools, three classes were assigned nonrandomly to one of the following conditions: (1) Second Step administered by the teacher, (2) Second Step administered by the teacher with assistance from trained peer leaders (leaders were from the same classroom and were nominated by peers), or (3) control group. The other two intervention schools each nonrandomly assigned one class to teacher-implemented Second Step and one class to the control group. Outcomes were measured one week before and one week after implementation of the curriculum, as well as three months after its completion. Of the 223 respondents who completed all three surveys, 64 percent were Hispanic, 17 percent were African American, and 18 percent were Caucasian. Fifty percent of the students were eligible for free or reduced-price lunch. At baseline, no significant differences were found between intervention and control groups on the researcher-developed Aggression Scale, a measure of common aggressive behaviors, during the week prior to the survey.
Grossman and colleagues (1997) conducted a randomized controlled trial involving 1,100 students in six pairs of matched schools. The study was conducted in a total of 49 second- and third-grade classrooms in 12 elementary schools from four school districts in King County, Washington. Participating schools were matched based on their school district, the percentage of students receiving free or reduced-cost school lunches, and the proportion of minority students. After matching, schools in each pair were randomly assigned to control or treatment status. A total of 418 intervention and 372 control students (66 percent of the 1,100 eligible students) obtained parental consent and completed all parent, teacher, and observer evaluation measures. These students were mostly similar at baseline; however, a somewhat larger proportion of control group students were identified as receiving special education services and were African American, while a higher proportion of intervention group students were Asian American. For the purposes of the evaluation, the Second Step program was implemented by classroom teachers over a 16- to 20-week period. Twice during the intervention period, classrooms were observed to determine the quality and fidelity of program implementation. Outcome data were collected before the start of the curriculum, two weeks following the conclusion of the program, and at follow-up six months after completion of the program. Outcomes included parent reports of child behavior problems and positive and negative social adjustment; teacher reports of child behavior problems and social competence and aggressive/antisocial behavior; and classroom and playground observations of social interactions between the child and other students or the teacher. In addition, 12 children from each study classroom (for a total of 588 students) were randomly selected to be part of an "intense observation" subsample. Children included in the observation subsample were observed in classroom, cafeteria, and playground settings for approximately one hour each at baseline, at two weeks after program completion, and at six months following program completion.
A dissertation by Lillenstein (2001) assessed Second Step in a sample of 285 kindergarten, first, and second grade students from 15 classrooms from four elementary schools in York County, Pennsylvania. The majority of students in the participating schools were white, Catholic, middle to upper class, and from two-parent households. A total of 184 students were nonrandomly assigned by the school principal to the Second Step program (nine classrooms) and 101 students to the waiting list control group (six classrooms). Second Step lessons lasted 15-30 minutes and were implemented once per week for six months. Of the 391 students in the participating classrooms, 367 obtained initial parental consent, and 285 parents (83 percent) completed both pretest and posttest outcome measures. Outcomes were assessed one week prior to and one week following implementation of the Second Step curriculum. Child outcomes were assessed via parent and teacher ratings of social skills and problem behavior, as well as classroom and lunchroom observations of aggressive behaviors and pro-social behaviors.
A second dissertation assessed the effects of Second Step in a sample of 109 third grade students from six classes in two public schools in Corpus Christi, Texas (Nicolet, 2004). The study evaluated the effectiveness of a brief form of the Second Step curriculum administered by the school counselor, which included three of the five lessons from each of the standard topics. The two schools were randomly assigned to participate, with a total of 54 students in the Second Step group and 55 students in the control group. Attrition throughout the study was very minimal. The schools included low- to upper-middle-class students, with slightly less than half of the students eligible for free or reduced-price lunch. The groups were pretested one week prior to implementation of the curriculum, and again the week following program implementation. Outcome measures included teacher rating of student aggressive behavior and pro-social behavior, and student self-reported anger management.
Frey et al. (2005) studied Second Step in a sample that involved a total of 1,989 children in grades two and four from 15 elementary schools in western Washington. The schools were located in urban districts of two moderate-sized cities, two proximal suburban districts, and one small city adjacent to a naval base. In the first year of the study, 11 schools were randomly assigned to the intervention (two-thirds of the sample) or to the control group (one-third of the sample). The other four schools were recruited the following year and were all assigned to the control group. None of the schools differed with respect to ethnic makeup or the proportion of students receiving free or reduced-price lunch. All participating schools, regardless of group assignment, received program materials, teacher training, and substitute teachers to fill in for teachers during regular teacher training; control schools received these benefits for classrooms that did not participate in the study. Teachers in the treatment group taught one or two program lessons per week. Of the 1,253 children with parental consent (63 percent of the initial sample), 462 (74 percent) in the treatment group and 436 (71 percent) in the control group completed all outcome measures over two years. No significant group differences in attrition were found. Outcomes were assessed via teacher ratings of studentsí behavior at pretest and in April of each school year, and at baseline the intervention groups scored significantly higher on antisocial behavior and marginally lower on social competence than did control students.
Key Evaluation Findings
The study by Orpinas et al. (1995) found:
- When the full intervention and control groups were evaluated across the three measurement periods, no significant intervention effect was found on the Aggression Scale scores.
- Race was significantly associated with aggression scores among boys of only one school. Since almost 90 percent of the students at the school were Hispanic, the analysis was repeated only for Hispanic boys in this school.
- At posttest, results showed that boys from the six intervention classes had reduced their aggressive behavior in rates ranging from 4 to 51 percent more than the control group. This reduction was marginally significant in only one class with a sufficient sample size to meet Promising Practices Network criteria, with boys from the teacher-only group reducing their aggressive behavior by 23 percent.
- No statistically significant differences between the intervention and control groups were found for girls.
- No significant differences were found between intervention and control groups for any of the teacher or parent reports.
- Observer ratings found:
- At posttest, there was a marginally significant difference between the intervention and control groups in physical negative and overall negative behaviors in the classroom. Rates of negative behavior decreased from baseline to posttest in the intervention group but increased in the control group.
- At posttest, there was a significant difference between groups in observed instances of negative physical behavior in the cafeteria or playground, with intervention group rates decreasing from 2.2 to 1.6 episodes per child-observation hour and control group rates increasing from 1.8 to 2.6 episodes per child-observation hour. At posttest, there were no significant differences between the intervention and control groups in instances of negative verbal behavior in the cafeteria or playground setting.
- At posttest, the rate of observed neutral or prosocial behavior in the cafeteria or playground setting increased by 17.1 more episodes per child-observation hour in the intervention group than in the control group.
- At the six-month follow-up, most of the significant differences between the intervention and control groups had dissipated because of a decline in negative behavior in the control group. The exception was physical aggression in the classroom setting, which was significantly lower in the intervention schools than in the control schools.
- A marginally significant difference was found between groups for the teacher-rated total problem behavior score, with the control group exhibiting significantly more problem behaviors than the intervention group.
- No significant differences were found between groups for either the parent-rated or teacher-rated social skills scores, parent ratings of problem behaviors, or classroom observations of pro-social behaviors or aggressive behaviors.
- No significant differences between groups for aggressive behavior or anger management scores.
- No significant differences between groups for pro-social scores; however, results indicated that the intervention groupís scores increased significantly more from pretest to posttest than did the control groupís scores.
- In the first year, significant group differences were found in teacher-reported social behavior.
- Among those children with high baseline ratings in antisocial behavior, the intervention group showed greater declines in antisocial behavior than the control group.
- Significant but smaller differences also favored the intervention group among students with low baseline scores. Intervention students with low baseline scores showed no change, whereas control students increased their rates of antisocial behavior.
- Positive effects faded by the second year, with no remaining group differences in antisocial behavior.
Public and private preschools, elementary schools, middle schools, after-school programs, and community-based child and youth organizations.
In the past, sites have funded the program through the Federal Safe and Drug Free Schools Program and through state violence prevention initiatives. Committee for Children maintains a list of possible funding sources on their website at www.cfchildren.org.
The Second Step program is designed for use by classroom teachers and requires minimal teacher-preparation time. A program kit contains all needed materials including an Administratorís Guide, a Teacherís Guide, scripted lessons, and a classroom video. The preschool level includes puppets and a sing-along audio CD.
The multiple levels of lessons allow students in Second Step schools to receive multiyear training in pro-social skills designed at a grade-appropriate level.
Preschool and Elementary school program:
- Program materials include photo-lesson cards and videos depicting children expressing emotions in real-life situations. (Video is not used in the preschool kit.)
- Teachers follow the scripted lessons written on the back of each photo card to help students connect their own emotional experiences to those depicted on the cards and in the videos.
- Lesson cards include discussion, role-playing, and other activities, as well as suggestions for how teachers can model the skills taught in the lesson throughout the week.
- Younger students are encouraged to interact with the Impulsive Puppy and Slow-Down Snail puppets and the Be-Calm Bunny plush toy.
- Lessons place a stronger emphasis on student attitudes and beliefs about aggression and provide skills for conflict resolution.
- Helps students understand their own social and emotional experiences; students are encouraged to relate them to those of a book character.
- The curriculum includes videos that stimulate discussion and role-playing lessons that are relevant to studentsí lives.
- The video-based Family Guide encourages family support of the Second Step program.
- Skills in empathy, impulse control, problem solving, and anger management are taught to help families practice and reinforce these skills with their children at home.
Lessons vary in length from about 20 minutes at the preschool level to 50 minutes in middle school/junior high. There are between 15-25 lessons in preschool-grade 5. For middle school/junior high, there are 15 lessons in the first year and 8 lessons each in the second and third years.
Each lesson uses a social scenario that forms the basis for discussion, role-playing, and other applied activities. The same topics are covered in each grade level, and content and activities are tailored and targeted to each grade level. Lessons build sequentially and are designed to be taught in order.
Second Step lessons are divided into the following areas of focus: empathy, problem solving, impulse control and anger management.
- Empathy: Teaches empathy as a series of skill steps, which include recognizing feelings in oneself and others, learning to understand othersí perspectives during discussions and role-playing, reducing labeling and stereotyping, and practicing respectful methods of communication and listening.
- Problem Solving: Provides children with problem-solving procedures. Children are taught to (1) identify problems, (2) brainstorm for solutions, (3) evaluate the possible effects and outcomes of solutions, (4) appropriately implement their solution, and (5) evaluate the outcome of their solution. In addition, this unit focuses on behavioral-skills training, providing students with the opportunity to practice interpersonal skills.
- Impulse Control/ Anger Management: Combines behavioral-skills training with problem-solving steps to manage anger. Lessons focus on recognizing anger cues and triggers and using positive self-talk and relaxation strategies to prevent escalation of anger. The unit focuses particularly on situations that typically generate stress for children, including feeling left out in social situations or activities, being criticized, and conflicts with siblings, peers, or parents.
The classroom program is implemented by regular classroom teachers. The Family Guide sessions are led by trained facilitators who are comfortable with leading parent groups.
Issues to Consider
This program received a "promising" rating. Two large studies using random assignment and large sample sizes, Grossman et al. (1997) and Frey et al. (2005), found that Second Step participants produced more positive scores in observer ratings of negative behaviors in the classroom, lunchroom, and playground, as well as reductions in teacher-reported antisocial behavior. Findings across other outcomes were not as positive; Grossman et al. found that reports of student behavior by teachers and parents were not significantly affected by program participation. Furthermore, quasi-experimental studies by Orpinas et al. (1995), Lillenstein (2001), and Nicolet (2004) found few to no significant program effects.
While the Grossman et al. (1997) study involved randomized assignment of treatment status at the school level, a schoolís selection for inclusion in the study as a whole was not random. Schools were selected for inclusion based on their perceived willingness and ability to deliver the curriculum. In addition, participation was contingent on the teachersí willingness to attend the program training sessions. Therefore, it is possible that these selection criteria for participation may have resulted in an atypical set of schools and teachers. Further, only 66 percent of the total number of eligible students obtained consent, had all evaluation measures completed, and were included in the evaluation. This nonrandom participation in the evaluation may have influenced outcomes.
When considering the fade-out of results that emerged from the Grossman et al. (1997) study, it is important to note that for the purposes of the Grossman teamís study and evaluation, the implementation of the program was altered from the regular design and procedure. Under normal program conditions, the program is implemented evenly and consistently throughout the school year and throughout the entire school. In contrast, the Grossman et al. implementation took place over a 16- to 20-week period. Further, not all second- and third-grade classrooms at the participating schools took part in the program and evaluation. Under regular program design and implementation, students at six months would have progressed into another Second Step unit for continued program exposure, as opposed to the study design in which participants were exposed to one single Second Step unit for the length of the study and then exposure ceased. In addition, attenuation may have been further compounded by the fact that at the six-month follow-up students had progressed into a new grade level. The experimental and control groups did not remain within intact cohorts, but were mixed. It may be that the experiences and behaviors of each group affected each other at the six-month follow-up. These data do, however, call into question whether, even under full program design and implementation, positive outcomes would persist after students were no longer exposed to the program and were actively interacting with individuals that had not received the program.
There are several limitations of the study by Frey et al. (2005), including the possibility that the manner in which the program was implemented may have led to spillover effects in the control schools, thus contaminating the control student population. The authors note that schools in the study were reluctant to agree to a waiting-list control design, and an agreement was reached so that those schools assigned to the control group were trained and allowed to implement the Second Step program in grades other than those being studied. There is a possibility that students and teachers who were not in the intervention Second Step classrooms may have shared information with students and teachers from the control classrooms, thus obscuring the true outcomes of the control groups in the absence of any Second Step implementation at the schools. Another study design limitation is that random assignment occurred only with the 11 schools recruited in the first year; in the second year, the remaining four schools were assigned nonrandomly to the control group. Thus, the overall design cannot be considered a true randomized controlled trial. In addition, similar to Grossman et al. (1997), positive treatment effect findings for reductions in antisocial behavior that were present in the first year of the study faded by the second year, suggesting that positive effects of the program may be limited in duration.
Finally, although the program has curricula available for preschool through middle school, the program has been evaluated only among children of elementary-school age, i.e., kindergarten through sixth grade. To date, there has been no rigorous research conducted and published on the programís effectiveness and outcomes with preschoolers or middle school children.
Prince Georges County, Maryland
Public Schools, Chicago
Catholic Charities Samaritan House Homeless Shelter, Denver
York County, Pennsylvania
Corpus Christi, Texas
San Antonio, Texas
King County, Washington
Client Support Services
Committee for Children
568 First Avenue South, Suite 600
Seattle, WA 98104-2804.
phone: (800) 634-4449 ext. 200
fax: (206) 343-1445
Program training, support, materials, and information are available from Committee for Children, a nonprofit organization that develops classroom programs focusing on youth violence, bullying, personal safety/child abuse prevention, and emergent literacy. Committee for Children offers both regional and on-site training for Second Step staff, trainers, and Family Guide facilitators; the schedule is available at http://www.cfchildren.org.
Program materials for staff, students, and families can be ordered online, and are available English. A Spanish language supplement is also available for an additional purchase.
Frey, Karin S., Susan Bobbitt Nolen, Leihua Van Schoiack Edstrom, and Miriam K. Hirschstein, "Effects of a School-Based Social-Emotional Competence Program: Linking Childrenís Goals, Attributions, and Behavior,"
Journal of Applied Developmental Psychology,
Vol. 26, No. 2, 2005, pp. 171-200.
Grossman, David C., Holly J. Neckerman, Thomas D. Koepsell, Ping-Yu Liu, Kenneth N. Asher, Kathy Beland, Karin Frey, and Frederick P. Rivara, "Effectiveness of a Violence Prevention Curriculum Among Children in Elementary School: A Randomized Controlled Trial," Journal of the American Medical Association, Vol. 277, No. 20, May 28, 1997, pp. 1605-1611.
Lillenstein, Jennifer A., "Efficacy of a Social Skills Training Curriculum with Early Elementary Students in Four Parochial Schools," dissertation, Indiana, Pa.: Indiana University of Pennsylvania, 2001.
Nicolet, Irene A., "The Second Step Violence Prevention Program: Effectiveness of a Brief Social Skills Curriculum with Elementary-Age Children," dissertation, Minneapolis, Minn.: Capella University, 2004.
Orpinas, Pamela, Guy S. Parcel, Alfred McAlister, and Ralph Frankowski, "Violence Prevention in Middle Schools: A Pilot Evaluation," Journal of Adolescent Health, Vol. 17, No. 6, December 1995, pp. 360-371.