Testing the effectiveness of the N-PALS model – a school-wide framework to prevent externalizing student problem behavior

Abstract

Intervention effects on moderate and severe problem behavior in and outside the classroom context, on classroom climate and inclusion, and on perceived collective efficacy, self-efficacy, and disciplinary practices among the school staff were examined using a strengthened nonequivalent control group (NEC) design and multi-level analyses. Differential intervention effects were examined with implementation quality (fidelity), school size, program training dosage, or portion of untrained staff as moderators. Selected data from NUBUʼs (Norwegian Center for Child Development) longitudinal effectiveness and development study ʽPositive Behavior Support in Schoolʼ were used. This study involved totally 12,050 students (Grades 4-7), 3,695 teachers, after school personnel, and principals, and a random sample of 2,750 parents from 65 elementary schools around the country. Twenty-eight of the schools implemented the full-scale N-PALS model, 17 schools implemented the abbreviated version (PPBS), and 20 schools doing ʽpractice-asusual ʼ (PAU) made up the control group. The study was conducted over five successive school years. The analyses and results included in the dissertation are based on staff and student ratings from five of the totally six assessment points (double pre-test, two intermediate tests, post-test), and cover four of the five study years. The N-PALS model seemed effective in reducing the level of both moderate and severe externalizing student problem behaviors (particularly on common school areas) and in promoting qualitatively better social and academic learning conditions in class. The results also indicate increased ability in the intervention schools to reach out to all students and to reduce the number of students singled out for individual intervention plans due to conduct problems. N-PALS also appeared effective in promoting the school staffs’ perceived individual and collective efficacy. Besides, this three-level prevention framework appeared effective in changing staff behavior in positive ways. The schools with high implementation quality benefitted the most from the intervention model, and small-moderately sized schools reached higher implementation scores than large schools. Outcome analyses of the abbreviated PPBS intervention indicated promising immediate effects on several but not on all of the same outcome variables as the full threelevel model. Moderation analyses indicated better outcomes for schools with high fidelity scores, high mean training dosage, higher portion of untrained staff, and for small to medium sized schools. The significant effect sizes for the two intervention frameworks were generally large enough to be considered of practical significance, and the confidence intervals were narrow, indicating high precision of the estimates. The study results indicate that SWPBS was successfully transferred across the Atlantic without major adaptations, without violating the national curriculum or the norm of inclusion. The current study adds to the existing base of knowledge in several ways, including by being the first to examine the cumulative effects of the three-level model and by being the first effectiveness study of the SWPBS model implemented in Europe and in a non-English speaking context. Moreover, the Norwegian study is the first investigating effects of SWPBS/N-PALS on perceived teacher collective efficacy and on school staffs’ disciplinary practices, and also one of very few longitudinal effectiveness studies to date that have been conducted on this school-wide prevention model. Besides, our study is the first to investigate the relationship between collective efficacy and problem behavior in school, and on an empirical basis suggesting that teacher collective efficacy is a significant and reliable indicator of a school’s culture. Ideally, the study results should be replicated in a Scandinavian randomized controlled trial before more firm conclusions are drawn on the effectiveness of the N-PALS model.