Meta-Analysis Snapshot
|
Literature Coverage Dates |
Number of Studies |
Number of Study Participants |
| Meta Analysis 1 |
1980-2015 |
37 |
37895 |
Meta Analysis 1Valdebenito and colleagues (2019) conducted a meta-analysis to determine the impact of school-based interventions on exclusion. The comprehensive search strategy included a keyword search of electronic databases (such as Criminal Justice Abstracts, ERIC, and PsycINFO) between September and December 2015. Key authors were also contacted for information on primary studies to be included in the review and lists of references from previous primary studies and reviews related to school-based interventions or exclusion were also assessed. There were no language restrictions placed on the searches, as long as the abstract was written in English. The methodological quality of each included study was evaluated using the Cochrane Effective Practice and Organization of Care risk-of-bias tool (Cochrane Effective Practice and Organization of Care 2007).
Literature included in the meta-analysis was drawn from published or unpublished studies from 1980 onward. Only studies with interventions that met the following criteria were included: 1) were randomized controlled trials; 2) targeted students from 4 to 18 years of age, irrespective of nationality, ethnicity, language, and cultural or socioeconomic background (studies involving students with serious mental disabilities or those in need of special schools were excluded); 3) were school-based interventions (delivered on school premises or supported by schools with at least one component implemented in the school setting); 4) encompassed a wide range of theoretical bases (targeting students or the whole school); and 5) addressed school exclusion as an outcome, regardless of whether the exclusion was fixed or permanent, in-school or out-of-school.
A total of 37 randomized controlled studies (reporting 38 effect sizes) of school-based interventions were included in the analysis. Included studies represented interventions carried out between 1980 and December 1, 2015. These studies were roughly equal in terms of being published (51 percent) or unpublished reports (49 percent). All were written in English and primarily represented studies from the United States (89 percent) and United Kingdom (8 percent). The remaining 3 percent represented one study in which the country of the sample was not reported.
In the meta-analysis, the mean age of the sample was 13 years old. The students attended schools with a high percentage of Black (54 percent) and Latino students (20 percent). Approximately two thirds of students were eligible for/receiving free school meals (66 percent). A wide range of school-based interventions were included in the review, organized into the following nine categories: 1) enhancement of academic skills, 2) afterschool programs, 3) mentoring/monitoring programs, 4) social skills training for students, 5) skills training for teachers, 6) schoolwide interventions, 7) violence reduction, 8) counseling and mental health interventions, and 9) other interventions. Broadly, 27 percent of the interventions were focused on changes at the school or teacher level, and 73 percent focused on changing students’ skills/behaviors to affect exclusion rates. School staff (32 percent) or school staff assisted by external facilitators (24 percent) delivered the interventions. Most of those delivering the interventions were school psychologists or counselors (32 percent), and in two studies (5 percent), the intervention was delivered by police or probation officers. Forty percent of the interventions were designed and/or delivered by the same researchers who also evaluated the impact of the intervention. Thirty-eight percent of the interventions were delivered over a period of 12 weeks or less, and an equal percentage were delivered over a period of more than 24 weeks.
In 23 studies, the control group received no treatment, 6 studies reported controls receiving treatment as usual, 4 offered a placebo to the control group, and 4 allocated controls via a waiting list. Regarding the unit of randomization, 70 percent of the studies randomized individuals, and almost 30 percent randomized clusters of students, either as entire schools or classrooms. The measures of exclusion reported in the included studies were mainly based on official records (81 percent) provided by schools or other official institutions.
Using a random effects model, standardized mean differences were calculated to measure program effects between the treatment and control groups. Using a 95-percent confidence interval, mean effect sizes were calculated for all 37 studies. Moderator and meta-regression analyses were also conducted.
