Description

This is a non-randomized trial, with a 3 (time) x 3 (condition) within-between subjects design.

All children (and their parents) within the participating primary schools are asked to join this study. Independent of participation in this study, all children in these schools will receive the intervention as it is part of the curriculum focussing on the development of prosocial skills.

Schools will be called to ask if participants would like to participate in the study: schools choose in what condition participation is desired. In march/april 2023: schools share information letter with parents by email/in school app with a link to the informed consent. In june 2023: during end-of-the-year meetings with parents, teachers will approach parents that have not responded yet whether participation is approved. In addition, children in group 8 (with possible children of 12 years and older) will get information from teachers about the study and an information letter. Subsequently, children will be asked whether participation is approved. Assessements will take place in classrooms of the participating schools during school hours for children 8 years and older. For children younger than 8 years parent report will be used.

This study will obtain a data set with a hierarchical structure. The three observations are nested within students, students are nested within classrooms and classrooms are nested within schools. Therefore, a three-level growth model will be used. Lee and Hong (2021) provided sample size guidelines based on simulation studies for three-level growth models using Monte Carlo methods while considering various realistic ICC values at the three levels. The conservative estimation is made that only 10 out of 23 (average number of students per class) students per class participate in this study. Therefore the simulation results of Lee and Hong (2021) provided for 10 students per group are considered. Across all fixed effects and an intraclass correlation coeffcient at level 3 (schools) of .15, 2 out of 6 power rates were found to be below the .80 level, when 50 groups are included. With 100 groups included, all power rates were found to exceed .87. Based on these numbers, the aim is to include 75 groups in total, bringing the total number of participating students (again assuming 10 participants per group) to 10x75 = 750.

Multilevel regression analyses will be performed to evaluate the effect of HiRO on social-emotional development, self-esteem, and the perceived classroom peer context. Multilevel analyses is an intent-to-treat procedure that does not impute missing data but deals with incomplete data by assuming that the available data for a given subject are representative of that subject's deviation from the average trends across time. The level1 model includes the time variable, which captures within-person variation over time, in the level 2 model between person characteristics such an intervention condition are used to predict the slope estimates. In the level 3 model between group characteristics can be modeled.