Background
Due to educational policy initiatives over the last two decades, school districts across the United States have placed more emphasis on improving academic standards and accountability. Indeed, children are spending between 20 to 25 hours per year on meeting federal, state and local school-district testing requirements (Hart et al., 2015). This increased emphasis on academic standards and high stakes testing has, at least in part, been blamed for the increasing levels of stress and anxiety children are experiencing (APA, 2009; Merkangas et al., 2010; Pope, 2010). In addition to changes in education policy requiring an increased emphasis on academic standards and accountability, schools are increasingly expected to attend to the social, emotional, and behavioral needs and problems of students. Given that as many as 13% to 20% children in the U.S. are experiencing one or more mental disorders (Center for Disease Control and Prevention, 2013), schools are increasingly challenged to respond to the growing emotional and behavioral needs of their students. Moreover, socioemotional development and competencies have been linked to learning and academic achievement, and have thus become a target for school-based interventions as a means of improving learning and academic achievement (Durlak, Weisberg, Dymnicki, Taylor, & Schhellinger, 2011; Eisenberg, Spinrade, & Eggum, 2010; Zins & Elias, 2006).
One approach to supporting improvements in socioemotional development and competencies that has received growing interest for use in schools is mindfulness-based interventions (MBIs). Mindfulness is defined as “paying attention in a particular way: on purpose, in the present moment, nonjudgementally” (Zabat-Zinn, 1994, p. 4). Research suggests positive effects of MBIs for adults with chronic conditions, mental health diagnoses, psychiatric disorders, and stress (Chiesa, Calati, & Serretti, 2011; deVibe et al., 2012; Cramer, Haller, Lauche, & Dobos, 2012; Vollestad, B. Nielsen, & H. Nielsen, 2012. Moreover, studies suggest that mindfulness based practices may improve performance on a variety of socioemotional outcomes, including self-regulation, stress, and mood disturbance (Cheisa & Serretti, 2009; Regehr, Glancy, & Pitts, 2013). There has been increasing interest in MBIs with children and adolescents, and schools are often seen as a convenient setting to implement MBIs with children and youth. Some of the more popular MBIs used in schools are Mindfulness Based Stress Reduction (MBSR; Kabat-Zinn, 1979; Bootzin & Stevens, 2005), Mindfulness Based Cognitive Therapy for Children (MBCT-C; Segal, Williams, & Teasdale, 2002; Semple, Reid, & Miller, 2005), Meditation of the Soles of the Feet (SoF; Singh et al., 2007), and Learning to Breath (LTB; Broderick & Metz, 2009). Despite the dramatic increase in the use of MBIs in schools to affect socioemotional and academic outcomes, little effort has been invested to systematically examine the body of evidence using rigorous methods, particularly for behavioral and academic outcomes of MBIs implemented in school settings.
This review contributes to the literature by including the most recent research on MBIs in school settings, including outcomes that have not been examined in prior reviews of school-based mindfulness interventions, including behavioral and academic outcomes, and employing stringent criteria for search, selection, coding, and analysis as specified in the Campbell Collaboration policies and guidelines to inform educational policy and practice and identify gaps in the current evidence base to guide future research in this growing area of practice and research.
Objectives
The purpose of this review is to examine and synthesize evidence of mindfulness-based interventions implemented in school settings with primary and secondary school students on achievement, behavior and socioemotional outcomes to inform education practice and policy. Specifically, the primary aim of this review is designed to answer the following research question: 1. What are the effects of mindfulness-based interventions on cognition, academic achievement, behavior and socioemotional outcomes? Moreover, within the context of this review, we aim to describe:
The types of mindfulness-based interventions being evaluated in school settings.
The state and quality of evidence of intervention outcomes studies of mindfulness-based interventions in school settings?
Search methods
We attempted to identify and retrieve both published and unpublished studies through a comprehensive search that included multiple electronic databases, research registers, grey literature sources, and reference lists of reviews and relevant studies. We searched 13 electronic databases, research registers, relevant clearinghouse, government and research center websites, conference abstracts/proceedings, reference lists of prior reviews and included studies, and contact with experts and researchers in the area of school-based mindfulness interventions.
Selection criteria
Studies were included in this review if they met the following criteria:
Types of studies: Randomized controlled trial (RCT), quasi-experimental design (QED), single-group pre-post test design (SGPP) or single subject design (SSD). We only included RCT and QED studies in the meta-analyses.
Participants: Pre-school, primary and secondary school students
Interventions: Interventions of interest include those that are a) conducted in a school setting (during the school day or in a school-based after school program) and b) use a mindfulness component/strategy. Mindfulness is broadly defined as “self-regulation of attention to the conscious awareness of one's immediate experiences while adopting an attitude of curiosity, openness, and acceptance” (Bishop et al., 2004, p. 174).
Outcomes: Studies must have reported at least one of the following outcomes:
1. Cognition (e.g., executive function, memory, cognition, attention)
2. Academic performance (e.g., standardized achievement tests, measures of content mastery, reading, grades)
3. Behavior (e.g., disciplinary referrals, aggression and other externalizing behaviors, time on task, compliance, attendance)
4. Socioemotional (e.g., anxiety, stress, engagement, social skills, self-esteem, emotion regulation, grit, internalizing behaviors)
5. Physiological (e.g., cortisol, heart rate, brain activity)
Other criteria: Studies must have reported post-test data, interventions must have been conducted in a primary or secondary school setting, and must have been conducted or published between 1990 and 2015. The search was not restricted by geography, language, publication status or other study characteristics.
Data collection and analysis
Titles and abstracts of the studies found through the search procedures were screened for relevance by two reviewers for most electronic databases, with the exception of the Australian Education Index, the British Education Index, and CBCA Education which were reviewed by one author. Documents that were not obviously ineligible or irrelevant based on the title and abstract review were retrieved in full text for final eligibility screening. Two reviewers independently reviewed each full text report using a screening form to determine final inclusion. Any discrepancies between the reviewers were discussed and resolved through consensus. For all studies that passed the eligibility screening process described above, two reviewers independently coded each eligible study using a structured data extraction form. Following independent coding of studies, coders then compared coding and identified and discussed discrepancies, which were resolved through consensus. If consensus could not be reached between the two coders, a third member of the review team was consulted to resolve the discrepancy. We conducted descriptive analyses on variables of interest from all included studies to provide information regarding participant, setting, intervention characteristics for all studies that met eligibility criteria. For those that met criteria for inclusion in the meta-analysis, we estimated effect sizes for each included RCT and QED study when enough data was reported in the study or provided by study authors. For RCT and QED studies, we calculated the magnitude of effect using the standardized mean difference effect size with Hedges’ g correction for continuous outcomes and odds ratios for outcomes presented as dichotomous variables.
Following the estimation of individual study level effects, we conducted separate meta-analyses using Comprehensive Meta-Analysis, version 3.0 (CMA; Borenstein, Hedges, Higgins, & Rothstein, 2014) for the following outcome domains of interest: cognitive, academic, behavioral and socioemotional outcomes. To synthesize effects across studies, a weighted mean effect was calculated by weighting each study level effect size by the inverse of its variance. Random effects statistical models were used throughout. RCT and QED studies were pooled to allow for greater statistical power in all meta-analyses (heterogeneity between RCT and QED studies was not statistically significant in any of the meta-analyses). Following the estimation of summary effects, we conducted a test of homogeneity (Q-test) to compare the observed variance to what would be expected from sampling error. The I2 statistic was used to describe the percentage of total variation across studies due to the heterogeneity rather than chance. We also constructed a forest plot displaying study-level mean effect sizes and 95% confidence intervals for the included studies to provide opportunity for visual analysis of the precision of the estimated effect sizes, detection of studies with extreme effects, and information regarding heterogeneity of studies. Publication bias was assessed using funnel plots. When there was significant heterogeneity across studies, we conducted moderator analyses to examine the following variables: study type (RCT, QED), provider (classroom teacher, trained instructor), researcher involvement (yes/no), homework (yes/no), manualized program (yes/no), and number of weeks (# of weeks of the intervention).
Results
Of the 61 studies that met criteria for inclusion in this review, 25 were RCT studies, 19 were QED studies, 9 were SGPP studies and 8 were SSD studies. Of the 44 RCT and QED studies, 35 provided enough data to calculate an effect size and were included in one or more of the meta-analyses, depending on which outcomes of interest were reported in the studies. Of the 44 RCT and QED studies, seven were unpublished reports, with the remaining being published studies in peer-reviewed journals. The interventions represent a wide range of mindfulness intervention types, but most interventions were at least partially manualized and of shorter duration. The interventions were delivered by primarily a classroom teacher (31%) or by a mindfulness-trained interventionist (60%) external to the school system.
Meta-analytic findings indicate small, yet statistically significant effects on cognitive outcomes (k = 10; g = 0.25 (95% CI [0.06, 0.43], p = .01) and socioemotional outcomes (k = 28; g = 0.22 (95% CI [0.14, 0.30], p < .001), and small and non-significant effects on academic outcomes (k = 5; g = 0.27 (95% CI [-0.04, 0.57], p = .08.) and behavioral outcomes (k = 13; g = 0.14 (95% CI [-0.02, 0.30], p = .08). Heterogeneity was small and not statistically significant in all meta-analyses with the exception of behavioral outcomes (I2 = 48%; T2 = .034; Q = 22.96, p = .03). Six studies measured physiological factors, with three of those studies from the same author team. Due to the nature of these measures, the time dependency of some of these measures (cannot compare AM cortisol to PM cortisol for example), and that so few studies measured these outcomes, quantitatively synthesizing these outcomes across these studies was not warranted.
Overall, there was a moderate to high risk of bias across the 35 studies included in the meta-analyses, with variation in high risk areas across studies. Twenty-one studies (60%) were rated low risk and 14 studies (40%) were rated high risk of bias for random sequence generation. Only one study reported that participants and study and school staff were blinded to program allocation (Sibinga et al., 2013), although they did not specify how they blinded study and school staff to condition. The remaining studies were rated as high risk (86%) or unclear risk (11%). Most of the studies in this review were rated as low risk of attrition bias (74%). For the studies included in the meta-analyses, most studies were assessed as unclear risk for reporting bias as we could not find any protocols of studies with which to compare planned to actual reported outcomes. It must be noted, however, that several studies reported to use, and then only reported data for, subscales of measures rather than reporting the full measure score or all subscale scores. Thus, it is unclear whether the a priori intentions of these authors were to only use certain subscales or if the participants did complete the full measure and the study authors only described the use of and reported data for the subscale only. We also coded for additional factors related to potential bias in this corpus of studies: researcher allegiance bias, funding source bias, and confounding factors. Of the 35 studies included in the meta-analysis, we were able to clearly identify an author role in the development and/or delivery of the intervention in 18 (51%) of the studies. The funding source was often not reported, or the studies were not funded. In four of the studies (11%), the authors identified a funding source that was also an entity involved in the development or the delivery of the intervention. We also examined whether there were confounding factors with either the treatment or comparison groups. Thirteen studies (37%) were assessed as having a confound at the level of the instructor (n = 7; only one instructor in treatment, control or both conditions) or at the school/classroom level (n = 6; one classroom or school in the treatment, control or both conditions). A larger mean effect was observed for studies in which a confound was present, thus likely upwardly biasing results; however, we conducted sensitivity analysis and found that the difference in magnitude of effect between studies with and without confounds was not statistically significant.
Authors’ conclusions
Results indicate mixed results of school-based mindfulness interventions across the outcomes of interest in this review, with finding favorable impacts of mindfulness interventions on those processes that are likely more directly targeted by mindfulness interventions, namely cognitive and socioemotional outcomes. We found a lack of support at posttest to indicate that those positive effects on cognitive and socioemotional outcomes then translate into favorable outcomes for academic and behavioral outcomes as is hypothesized. The lack of heterogeneity for all outcomes with the exception of the behavioral outcomes indicate that the interventions in this review, although quite diverse in their characteristics, produced similar results across studies on cognitive, socioemotional and academic outcomes. These findings provide some support for the use of school-based mindfulness interventions for some outcomes, but do not provide overwhelming support of MBIs as being the panacea as some have advocated. The quality of the evidence varied, with some important risks of bias present across a large proportion of studies which threatens the internal validity of the included studies and is cause for caution in interpreting the results of this review.
Overall, the evidence from this review urges caution in the enthusiasm for, and widespread adoption of, school-based mindfulness interventions for children and youth. While the evidence points to positive effects on socioemotional and cognitive outcomes, there is a lack of evidence of effects on academic and behavioral outcomes. Despite the empirical support of mindfulness-based interventions for adults, children and adolescents may not benefit from mindfulness-based interventions similarly to adults. Children and adolescents may not be developmentally ready for the complex cognitive tasks, focus and level of awareness that mindfulness-based interventions require. Moreover, we know little about the costs and adverse effects of school-based mindfulness interventions—the costs of implementing these programs may not be justified, and there are some indications that mindfulness-based interventions may have some adverse effects on children and youth; however, these have not been adequately examined. If schools do want to implement mindfulness-based interventions, we urge schools to evaluate the practice in a rigorous way and monitor outcomes and costs.
Oversett med Google Translate
