Clinical Trial of Second Step Middle School Program: Impact on Bullying, Cyberbullying, Homophobic Teasing, and Sexual Harassment Perpetration
Espelage DL, Low S, Van Ryzin MJ, Polanin JR · 2015
grade Brctindependentfailed
Sample
36 middle schools in 18 matched pairs across Kansas and Illinois; 3,616 sixth-graders at baseline; 41 lessons over three years
Population
United States; middle-school students.
Design
A three-year cluster-randomised trial of the most widely used American SEL programme, evaluated by a university team independent of the publisher and funded by the Centers for Disease Control. Every outcome is STUDENT SELF-REPORT, which is the trial's principal limitation - no disciplinary record, attendance or attainment outcome was measured.
Key findings
The widely quoted result is from year one, where intervention students were 42% less likely to self-report physical aggression, with no significant effects on verbal or relational bullying, peer victimisation, homophobic teasing or sexual violence. Year two showed significant effects on 2 of 7 outcomes, present in Illinois but not Kansas schools. At the three-year endpoint no direct intervention effects were found for multiple forms of aggression perpetration; only indirect effects mediated by self-reported delinquency remained.
Genetic confound
Low. School-level randomisation with matched pairs.
Replication notes
Failed within itself. The 42% year-one figure that circulates in programme marketing did not survive to the three-year endpoint of the same trial.
DOI / URL
10.17105/spr-15-0052.1
Effects
| Outcome | Metric | Value | Measure | Timing | Vs | Horizon | Class |
|---|---|---|---|---|---|---|---|
| Self-reported physical aggression, year 1 | relative likelihood | 42% less likely in intervention schools | researcher-designed | 1 year | business-as-usual | end-of-treatment | behaviour |
| Aggression perpetration at the 3-year endpoint | significance | no direct intervention effects; only indirect effects via self-reported delinquency | researcher-designed | 3 years | business-as-usual | over-2yr | behaviour |
| Year-1 outcomes other than physical aggression | significance | no significant effects on verbal/relational bullying, victimisation, homophobic teasing or sexual violence | researcher-designed | 1 year | business-as-usual | end-of-treatment | behaviour |
Cited by
- Should a school buy a social and emotional learning programme?mixedconf: highgc: low