Study 1
Axford and colleagues (2021) conducted a randomized controlled trial to examine the implementation and effectiveness of the Chance UK program in five sites in London, England: 1) Enfield, 2) Hackney, 3) Islington, 4) Lambeth, and 5) Waltham Forest. Children (who were ages 5–10 at referral, thus 5–11 during program participation) were referred by school staff based on behavior or social difficulties and were eligible based on initial assessment of behavior problems by teacher and/or parent report. A total of 246 children were randomly assigned to receive the mentoring intervention or standard services (123 children assigned to each group). Youths assigned to the intervention group were offered mentoring services provided by Chance UK, while youths assigned to the control group received services typically available to children within each borough area.
The great majority of the study sample (87.4 percent) were boys, with a mean age at baseline assessment of 8.4 years. Most children in both the intervention and control groups were of minority ethnic status (66.7 percent and 56.6 percent, respectively). About one participant in four came from households that reported struggling financially. Most children in both intervention and control groups lived in single-parent households (79.3 percent and 78.5 percent, respectively). Intervention and control groups were similar in terms of child age, gender, parent marital status, family socioeconomic status, and baseline assessment of outcome measures. Children in the intervention group were slightly more likely to identify minority group membership and represent the lowest income category. Attrition was somewhat higher (40 percent) for the control group, compared with intervention (27 percent).
The outcome measures included the parent and teacher Strengths and Difficulties Questionnaire, in which items are scored on a three-point scale (0 = “not true,” 1 = “somewhat true,” or 2 = “certainly true”). The Total Difficulties scale reflects subscales of conduct problems, emotional problems, hyperactivity, and peer problems, with scores ranging from 0 to 40. The Prosocial scale has five items, with scores ranging from 0 to 10. The parent-reported Eyberg Child Behavior Inventory Problem Scale (26 items, with a score range of 0 to 36) was also an outcome measure. Child self-report outcome measures included the Global Self-Worth subscale of the Harter Self-Perception Profile for Children, which has six items rated on a four-point scale in which higher scores reflect greater self-perceived competence, and the overall score on the Children’s Hope Scale, which reflects the summed score of all six scale items. These scales were administered at baseline (before randomization) and at 16 months post-randomization. The Self-Perception Profile for Children and Children’s Hope Scale were administered only to participating children who were 8 years and older at baseline.
Intervention effects were tested using an intent-to-treat analysis that compared children assigned to the mentoring intervention group with those assigned to the control group on endpoint scores (16 months post-assignment) on outcome measures, adjusting for child’s age, gender, borough, ethnicity, socioeconomic status, special educational needs status, parent marital status, and baseline depression, and the baseline value of the outcome measure.
The study conducted subgroup analyses through exploratory moderator analyses for age, gender, parental marital status, socioeconomic status, ethnicity, and parent-reported Strengths and Difficulties Questionnaire Total Difficulties score at baseline (borderline versus abnormal). A planned complier average causal effect analysis was conducted to compare intervention youths who participated in 11 or more months of mentoring (program recommended amount) with a comparable group in the control condition.
