Coping Power Universal for Preschoolers

Study 3

Muratori et al. (2019) found that, relative to the control group, students in the intervention group showed significantly greater improvements in:

• Emotional problems (internalizing) (parent-report)
• Hyperactivity (teacher-report).

Risk and Protective Factors

• Overall stress (parent-report)

Study 1

Summary

Muratori et al. (2017) conducted a cluster randomized trial by randomly assigning 10 preschool classrooms (n=164 children) from two nursery schools in Lucca (Italy) to condition. Parents and teachers completed surveys at baseline and post-intervention on the frequency of child behaviors associated with conduct problems, such as emotional symptoms, behavioral difficulties, hyperactivity, peer relation problems, and prosocial behavior.

Muratori et al. (2017) found that one month after the intervention, compared to participants in the control condition, participants in the intervention condition showed significantly improved:

• Behavioral difficulties (teacher- and parent-report)
• Prosocial behavior (teacher-report).

Risk and Protective Factors

• Overall stress (teacher-report)

Evaluation Methodology

Design:

Recruitment: The program was proposed to five nursery schools in Lucca, a town in Tuscany, Italy. Two of the five schools agreed to participate, and all 4-year-old children from these two schools were recruited for the study. Child inclusion criteria were: 1) aged 4 years, and 2) attended a nursery school in Lucca. The final sample included 164 children (90 boys, 74 girls) from 10 classrooms.

Assignment:

Classrooms were randomly assigned to the intervention or control conditions (n=84 children in the intervention classrooms; n=80 children in the control). Randomization was computer-generated and was completed by an independent researcher.

Assessments/Attrition:

Assessments occurred at baseline and one month after completion of the six-month intervention. Attrition was not reported, though initial and analysis Ns matched in all tables (suggesting no attrition).

Sample:

The mean age of participants in the intervention condition was 4.45 years (SD = 0.10) and in the control condition 4.55 years (SD = 0.15). The intervention sample was 57% male, 75% Caucasian, and 25% African, and the control condition was 53% male, 82% Caucasian, and 18% African. Thirty percent of the intervention sample had household incomes less than 8,000 euros, 47% between 8,000-26,000 euros, and 23% greater than 26,000 euros, whereas 29% of the control condition had household incomes less than 8,000 euros, 50% between 8,000-26,000 euros, and 21% greater than 26,000 euros.

Measures:

Teachers (who delivered the intervention) and parents (who were blind to conditions) reported on child behaviors using a well-established questionnaire. Six outcomes from each informant were computed for analysis: overall stress, emotional symptoms, behavioral difficulties, hyperactivity, peer relation problems, and prosocial behavior. Subscale reliabilities in the current sample ranged from α = 0.79 (prosocial behavior) to 0.86 (hyperactivity).

Analysis:

The researchers used a repeated measures analysis of variance (ANOVA) to test for intervention effects. Time (pretest vs. posttest) was considered as a within-subjects factor and condition (treatment vs. control) was considered as a between-subjects factor. Classroom was included as a covariate to adjust for clustering by unit of assignment. All analyses controlled for age and gender, though the basis for inclusion of these covariates was not reported.

Intent-to-Treat: All available data were used in analyses.

Outcomes

Implementation Fidelity:

Intervention adherence was measured using a questionnaire completed by teachers after each session.

Baseline Equivalence:

No significant differences were detected in baseline demographics for the assigned sample, and no attrition occurred over the course of the study. However, tests for baseline equivalence on the outcome measures were not reported.

Differential Attrition:

No attrition.

Posttest:

One month after the intervention, relative to control participants, children in the intervention condition showed significantly greater improvements in four of twelve outcomes tested: teacher-rated overall stress, teacher- and parent-rated behavioral difficulties, and teacher-rated prosocial behavior.

Long-Term:

Not examined.

Study 2

Summary

Muratori et al. (2021) conducted a cluster randomized trial that randomly assigned 16 classrooms (n = 250 children) within two nursery schools located in southern Tuscany (Italy) to condition. Assessments before the intervention (November 2018) and after the intervention (May 2019) measured pre-academic skills, as well as teacher-reported youth externalizing problems, internalizing problems, and prosocial behavior.

Muratori et al. (2021) found that, relative to the control group, students in the intervention group showed significantly greater improvements from baseline to posttest in:

• Early mathematics and reading development
• Hyperactivity (teacher-report)
• Emotional symptoms (teacher-report)
• Prosocial behavior (teacher-report).

Evaluation Methodology

Design:

Recruitment: The study was conducted from November 2018 to May 2019. Sixteen classrooms in two nursery schools located in the south of Tuscany, Italy participated in the study. Child inclusion criteria were: (1) must attend the last year of nursery school; and (2) must obtain written consent from parents. The study did not report student eligibility or consent rates. The initial sample was comprised of 250 youth from 16 classrooms. No other recruitment details were provided.

Assignment: Classrooms were randomly assigned to treatment (n = 8 classrooms, 151 students) or control (n = 8 classrooms, 99 students) conditions. An inquiry to the authors confirmed that random assignment occurred within schools. Randomization was achieved using a computer-generated random allocation sequence. Intervention students received the program during school hours as part of the daily school routine, whereas control students received the standard curriculum activities provided in the Italian school context.

Assessments/Attrition: Assessments occurred at baseline and posttest (i.e., in the month following program end). It appears that all 16 classrooms remained in the study throughout the follow-up period. Overall student attrition rates were not reported, likely because student-level attrition was low. Attrition rates per outcome measure ranged from <1% to 3.2%.

Sample: The sample was evenly divided in terms of gender (50% male, 50% female). In addition, most (80%) were Caucasian and 20% were African, with a mean age of 4.61 years. No information at the cluster (i.e., classroom) level was provided. 

Measures: Teachers reported on measures of child externalizing problems, internalizing problems, and prosocial behavior at baseline and posttest. Because teachers delivered the intervention, the teacher-reported measures are non-independent. In addition, a school psychologist, who was blind to condition, individually administered measures of child pre-academic skills at baseline and posttest. Thus, nine outcomes were computed for analysis, five of which were teacher-reported (child conduct problems, hyperactivity, emotional symptoms, peer problems, and prosocial behaviors), and four of which were collected by the school psychologist (two measures of early mathematical development: ordering quantities and Arabic numbers comparison; and two measures of meta-phonological skills, prerequisites for learning reading: word fusion and word segmentation). The hyperactivity and peer problems outcomes represent risk and protective factors, whereas the outcome of emotional symptoms can be viewed as a measure of internalizing. In the current sample, scale derived alphas ranged from .73 (conduct problems) to .82 (prosocial behaviors).

Analysis: Analyses used generalized linear mixed models with fixed effects for time (baseline, posttest) and condition (intervention, control). Random effects for individual and classroom were used to account for the multiple responses per participant as well as clustering of participants within classrooms. The authors noted that they did not include school variables in the models because they collected data from only two schools. 

Intent-to-Treat: The analysis sample (per outcome measure) for the primary analyses was limited to students with complete data (i.e., missing data were handled using listwise deletion), but the sensitivity analysis used all available data. It appears that the researchers examined students according to the condition to which they were assigned, which follows an intent-to-treat protocol, but this was not explicitly stated. 

Outcomes

Implementation Fidelity: Checklists completed by teachers revealed that 86% of intervention elements were delivered as intended. In addition, a school psychologist observed 20% of the previously recorded sessions and scored teachers' behavior during two essential phases (i.e., goal sheet and daily activity) of each intervention session using a systematic observation grid. Eighty-eight percent of the reported levels of adherence were good or excellent.

Baseline Equivalence: No significant differences were detected in two tests of baseline demographic variables (gender, race) for the randomized sample (n = 250). In addition, the authors noted that "no differences emerged between the two groups on the study's variables at pre-test" (p. 620), but no significance tests were reported. However, Table 3 (p. 619) presents mean values for all baseline outcomes, which appear similar across the two groups. Baseline equivalence at the cluster level was not reported. 

Differential Attrition: Not examined, though student-level attrition per outcome measure was <5% for all outcomes.

Posttest: At posttest, relative to children in the control group, those in the intervention group showed significantly greater improvements in three of five teacher-reported outcomes (hyperactivity, emotional symptoms, and prosocial behaviors), as well as three of four measures of pre-academic skills (ordering quantities, Arabic numbers comparison, and word fusion). These significant posttest effects all held in sensitivity analyses that included all available data.

Long-Term: Not examined.

Study 3

Summary

Muratori et al. (2019) randomly assigned six classrooms (n=102 children) from two preschools located in Rovigo (Italy) to condition in July 2016. Parent- and teacher-reported assessments were conducted in September 2016 (baseline) and May 2017 (after the end of the intervention).

Muratori et al. (2019) found that, relative to the control group, students in the intervention group showed significantly greater improvements in:

• Emotional problems (internalizing) (parent-report)
• Hyperactivity (teacher-report).

Risk and Protective Factors

• Overall stress (parent-report)

Evaluation Methodology

Design:

Recruitment: The study was conducted from July 2016 to May 2017. Children from two preschools in Rovigo (Italy) were recruited for the study. Otherwise, little information on recruitment was provided, except that 100% of children within participating classes received informed parental consent to participate in the research. The initial sample comprised 102 students from six classrooms.

Assignment: Classrooms were randomly assigned to treatment (n=48 children within three classrooms) and control (n=54 children within three classrooms) conditions. Randomization was achieved using a computer-generated random allocation sequence. The control group followed the "standard curriculum activities," meaning that they received education for preschool activities on reading, handwriting and mathematics, as well as motor and play activities and did not receive any psychological activities. 

Assessments/Attrition: Assessments occurred in September prior to the start of the intervention and in May after the intervention ended. Attrition was not discussed. However, there was the possibility of no attrition, which could be the reason the authors did not report explicitly on attrition.

Sample: Slightly more than half (56%) of the sample were girls, and 44% were boys. The mean age was 5.32 years old. Six children in the sample had an intellectual or sensory disability certificate. Teachers had a mean number of ten years of experience with preschool children.

Measures: Teachers and parents completed the Strengths and Difficulties Questionnaire to assess students' behavioral problems (i.e., conduct problems, hyperactive behaviors, and emotional symptoms; Cronbach's alphas = 0.83-0.86 for the current sample). Teachers also completed the Questionnaire for Early Identification of Learning Disabilities to assess pre-academic abilities (Cronbach's alpha = 0.77 for the current sample).

Analysis: Researchers used a 3-level linear mixed-effects modeling (MIXED) procedure in SPSS (measurement occasion within individuals within classes) to assess problem behavior.

Pre-academic skills were assessed as subgroups using chi-square tests since scores were converted into percentiles and ranked according to four categories (lower than 10th percentile; between 11th and 25th percentile; between 26th and 50th percentile; and between 51st and 99th percentile).

Intent-to-Treat: Intent-to-treat was not discussed. Researchers used maximum likelihood (ML) estimation, which following What Works Clearinghouse (WWC) guidelines is an acceptable strategy for handling missing data and assumes no missing data.

Outcomes

Implementation Fidelity: At the end of the program, researchers interviewed the control group teachers to verify that they had not received information about the intervention. All teachers from the intervention classes attended an initial eight-hour training workshop in October 2016 and had two-hours monthly meetings in small groups for the duration of the program. Treatment group teachers were asked to complete a four-point Likert scale to assess adherence, with higher scores indicating better adherence. The mean score of the intervention adherence checklist was 3.55 (SD=0.35).

Baseline Equivalence: There were no significant baseline differences between conditions among four parent-reported and four teacher-reported assessments of problem behavior (total stress, conduct problems, hyperactive behavior, emotional symptoms). There were also no significant baseline differences in pre-academic skills. Differences between conditions in socio-demographic characteristics were not discussed. 

Differential Attrition: Attrition was not discussed.

Posttest: Results showed differences in favor of the treatment group for three of eight outcome measures. Compared to the control group, children in the treatment group showed significant reductions in parent-reported overall stress (a risk and protective factor) and emotional problems, as well as teacher-rated hyperactivity.

Subgroup chi-square tests showed that relative to the control group, a significantly greater proportion of students in the treatment group were rated by their teachers as being in the highest category of percentiles in the pre-academic skills of pre-math and language abilities.

Long-Term: Not reported.