Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2026 Aug 1;18(4):e70199. doi: 10.1111/aphw.70199

Longitudinal and dynamic associations of engagement in mindfulness training with emotion regulation difficulties, mindfulness, and self‐compassion in community adolescents

Megan J Moran 1,✉, Stephen Aichele 1, Addie Rzonca 1, Lauren B Shomaker 1, Rachel G Lucas‐Thompson 1
PMCID: PMC13428265  PMID: 42538848

Abstract

Despite enthusiasm for mindfulness‐based intervention (MBI) for preventing adolescent mental health problems, effects tend to be small. One potential explanation is that some adolescents are not engaging with MBI in ways that maximize benefits. In this study, we sought to understand the time‐ordered and interconnected changes in engagement and target intervention processes of emotion regulation difficulties, mindfulness, and self‐compassion. In a single‐arm trial, n = 73 community adolescents participating in a 6‐week, group MBI completed self‐report measures of engagement, emotion regulation difficulties, mindfulness, and self‐compassion at each weekly session. Caregivers reported weekly on their child's emotion regulation difficulties. Using structural equation models, we tested longitudinal and time‐ordered change associations for engagement and target processes, at both the between‐ and within‐person levels. Between‐person, increases in engagement were associated with decreases in adolescent‐reported emotion regulation difficulties and increases in self‐compassion. Within‐person, higher levels of engagement at any given session, relative to one's own average, were associated with decreased adolescent‐reported emotion regulation difficulties and increased self‐compassion at the subsequent session. Findings provide initial evidence that engagement is an important underlying process in adolescent MBI. Person‐centered, adaptive approaches to optimizing engagement in adolescent MBI warrant investigation to assess their potential to enhance MBI effects on health outcomes.

Keywords: adolescence, engagement, mental health, mindfulness‐based intervention, prevention

INTRODUCTION

Adolescence is a sensitive window for the onset of mental health problems, and early symptoms often persist or worsen over time (Jones, 2013; Prince et al., 2007). Mindfulness‐based intervention (MBI), which involves training in what mindfulness is and how to practice it, often with the goal of supporting participants in managing stress and strengthening their capacity for mindfulness, is typically delivered with nonclinical adolescents as a group‐based intervention, and often in school settings (Zoogman et al., 2014). MBI has been shown to strengthen capacities including emotion regulation, mindfulness, and self‐compassion, which are associated with better mental health in nonclinical adolescent samples (Bluth et al., 2016; Dumontheil, 2022; Galla, 2016; Gonçalves et al., 2019; Porter et al., 2024). Emotion regulation refers to the ability to understand, accept, and/or apply strategies to adaptively respond to emotions; mindfulness is present‐moment focused, nonjudgmental attention; and self‐compassion is the capacity to extend sympathy and kindness toward oneself (Gratz & Roemer, 2004; Kabat‐Zinn, 1990; Neff, 2003). Yet, meta‐analyses show small effects, and several recent large‐scale trials have reported null findings (Fulambarkar et al., 2023; Kuyken et al., 2022). Although null and small effects may reflect multiple factors, including variation in program content or delivery, engagement represents a particularly critical and understudied mechanism. Prior evidence suggests some adolescents do not engage fully with MBI (Montero‐Marin et al., 2022), which may limit benefits even when programs are well designed.

Engagement has been defined as productive involvement with an activity and is inherently multidimensional, encompassing both active participation and emotional response (Ben‐Eliyahu et al., 2018). Consistent with this work, we conceptualize engagement as comprising a cognitive‐behavioral component (reflecting combined attentional and participatory involvement) and an affective component. Prior research indicates that cognitive and behavioral aspects of engagement are empirically overlapping in adolescents and are best represented as a combined factor within a broader multidimensional engagement structure (Ben‐Eliyahu et al., 2018), which is the approach we use in this study. Engagement may be especially important in MBI because MBI benefits depend on practice rather than passive receipt of information (Kabat‐Zinn, 1990; Quach et al., 2017; Shapiro et al., 2018). Requiring adolescents to participate in MBI when they do not engage may, in some cases, exacerbate existing mental health concerns (Kuyken et al., 2022). However, no rigorous studies have examined adolescent engagement in MBI, leaving a critical gap in our understanding of how engagement relates to MBI target processes. Clarifying the timing and interplay of changes in engagement, emotion regulation, mindfulness, and self‐compassion is, therefore, essential for identifying how to optimize MBI for adolescent mental health.

Outside of the MBI literature, engagement predicts learning outcomes (Fredricks, 2011; Wong et al., 2024) and psychotherapy outcomes (Marker et al., 2013; Neimeyer et al., 2008). Given that adolescent MBI shares features with both school‐based learning (Broderick, 2021) and psychotherapy (Fresco & Mennin, 2019; Goldberg, 2022), there is a strong theoretical basis to expect that engagement may be an important mechanism in MBI as well. Engagement and outcomes may also be reciprocally related; for example, higher engagement predicts improvements in self‐regulation, which in turn predict higher engagement (Stefansson et al., 2018; see also Karababa, 2020; Labouliere et al., 2017; Low et al., 2014; Roebroek & Koning, 2016).

Yet, features of adolescence may make engagement with MBI particularly challenging. Although some MBI programs are designed to make mindfulness more accessible and concrete, mindfulness as a concept and process is paradoxical, which may make it challenging for adolescents to put into practice in the ways that, theoretically, afford the most benefit (Shapiro et al., 2018). For example, certain elements of practicing mindfulness may promote a relaxing retreat from stressful stimuli (e.g. closing eyes, lying down); yet, the intent of practice is actually to engage more fully with all of life's experiences, including those that are more unpleasant (Shapiro et al., 2018). Neurocircuitry supporting meta‐cognitive activities (e.g. prefrontal cortex and anterior cingulate cortex) are still maturing (Grayson & Fair, 2017; Tang et al., 2015), and mindfulness is inherently a meta‐cognitive activity; therefore, adolescents may struggle to remain mindful, particularly under “hot” cognitive conditions such as when they are under stress (Lucas‐Thompson et al., 2024). Thus, mindfulness practice may be more difficult for adolescents relative to adults. As a result, adolescents likely need more and/or different types of support to fully engage with MBI.

Research on engagement in adolescent MBI is limited and has largely relied on narrow proxies like home practice or attendance (Montero‐Marin et al., 2022; Tudor et al., 2022). The present study directly addresses this gap by using repeated measures of multidimensional engagement, including both cognitive‐behavioral and affective components in line with the learning/education literature (Ben‐Eliyahu et al., 2018; Fredricks, 2011), and target intervention outcomes (emotion regulation, mindfulness, and self‐compassion) in a sample of nonclinical adolescents participating in a single‐arm, open trial of MBI. We tested longitudinal and time‐ordered change associations between engagement and target outcomes. We hypothesized that increases in engagement are associated with improvements in outcomes, both in terms of overall change at the between‐person level (i.e. relative to the group average) and in terms of timepoint‐to‐timepoint change at the within‐person level (i.e. when adolescents reported engagement that was higher relative to their own average, they also report improvements in target processes at the subsequent timepoint). We also expected reciprocal effects, such that improvements in outcomes predict subsequent increases in engagement.

METHOD

Participants

Participants were n = 74 adolescents (M age = 14.08 years, SD = 1.97 years, range = 12–18 years) and a caregiver (86% female, M age = 45.11 years, SD = 6.18 years, range = 29–63 years). Adolescents participated across eight consecutive cohorts between June 2023 and July 2024. Cohorts ranged from 6 to 12 adolescents (median = 8.5). One adolescent who did not complete all study measures for at least one timepoint was excluded, resulting in a final analytic sample of n = 73.

Procedures

All procedures were approved by the Colorado State University Institutional Review Board. Participants were recruited through flyers, email listservs, and community/school referrals, which were directed to both adolescents and parents of adolescents. Interested adolescents and/or parents of adolescents contacted study staff, and adolescents completed a screener using the Center for Epidemiologic Studies of Depression Scale (Radloff, 1977). Adolescents with elevated symptoms (>20) were then screened for suicidal ideation and self‐injurious behavior using the Schedule for Affective Disorders and Schizophrenia for School‐Aged Youth Suicidality and Self‐Harm module (Kaufman et al., 1997). Adolescents endorsing active suicidal ideation with intent, method, or plan or self‐injurious behavior were ineligible. Eligible participants provided informed consent (≥18) or assent (12–17). Adolescents and caregivers completed demographic questionnaires prior to intervention.

All adolescents participated in Learning to BREATHE (L2B, Broderick, 2021), a 6‐week, manualized, adolescent‐focused group MBI. Each weekly, in‐person 60‐min session included didactic content, experiential activities that support participants in understanding mindfulness, and guided mindfulness practices (~2 to 8 min) (Broderick, 2021). Sessions correspond with the acronym, BREATHE, and cover the following themes: mindfulness of Body, mindfulness of Reflections (i.e. thoughts), mindfulness of Emotions, the role of Attention in stress management, Tenderness toward oneself, and healthy Habits to continue mindfulness practice beyond the group, all with the goal of Empowerment (see Table S1 for detailed session overviews). Participants were encouraged to practice mindfulness between sessions. L2B facilitators were two graduate students (MM and AR) who were trained (~24 h) by a certified master L2B trainer, completed supervised mock facilitation sessions, and had facilitated multiple previous cohorts of L2B in other studies. MM and AR co‐facilitated some groups and led others independently.

Adolescents completed self‐report measures via pen and paper at each weekly session. Adolescents completed measures of target intervention processes (emotion regulation difficulties, mindfulness, and self‐compassion) at the beginning of each session and measures of engagement at the end of each session. If an adolescent missed a session, study staff emailed their surveys via REDCap to complete measures for that week. Caregivers reported on their child's emotion regulation difficulties each week on the day their child was scheduled to attend L2B via a survey sent using REDCap (Harris et al., 2009). Adolescents were compensated for completing surveys, including for completing surveys if absent from the session that week.

Measures

Emotion regulation

Adolescents responded to the 18‐item Difficulties in Emotion Regulation Scale—Short Form (DERS‐SF; Gratz & Roemer, 2004), rating on a 5‐point Likert scale (1 = almost never to 5 = almost always) how often certain statements apply to them (e.g. “When I'm upset, it takes me a long time to feel better.”). Higher scores suggest greater problems with emotion regulation, and a mean was calculated to create an overall score. The DERS‐SF has demonstrated acceptable to excellent internal consistency (average α for the subscales = .81) and high convergent validity in adolescent samples (Neumann et al., 2010). Parents completed an adapted parallel version, which has demonstrated reliability (α = .84) and validity for parent reports (Bunford et al., 2020).

Mindfulness

Adolescents completed the 14‐item Mindful Attention Awareness Scale, a reliable (α = .85) and valid measure of mindful attention for this age group (de Bruin et al., 2011). They reported how often, in the past week, they engaged in mindful attention (e.g. “I rush through activities without being really attentive to them”) on a scale from 1 = almost always to 6 = almost never. A mean was derived, with higher scores reflecting higher mindful attention.

Self‐compassion

Adolescents responded to the 10‐item Self‐Compassion Scale—Short Form (SCS‐SF Raes et al., 2011), which has demonstrated reliability (α = .86). On a 5‐point Likert scale from 1 = almost never to 5 = almost always, they indicated how often, in the past week, certain statements applied (e.g. “I try to see my failings as part of the human condition”). Items were averaged, with higher scores indicating more self‐compassion.

Engagement

Adolescents completed the 16‐item Activity Engagement Scale (AES), which has been used to reliably (α = .64–.88) assess adolescent engagement in out‐of‐school contexts and includes items that capture both cognitive‐behavioral and affective dimensions of engagement (Ben‐Eliyahu et al., 2018). Items were modified to refer to “L2B” as the activity (e.g. “Today in L2B, I contributed to the discussion” and “Today in L2B, I felt frustrated or annoyed” [reverse scored]). On a scale of 1 = not at all true to 5 = very true, adolescents rated the extent to which statements applied to them. Participants also completed two subscales of the Intrinsic Motivation Inventory (IMI, 2022; Ryan & Deci, 2000), which have shown good internal consistency in adolescent samples (αs = .85–.94) (Bosch, 2024). Adolescents rated statements on a scale of 1 = not at all true to 7 = very true. They completed the seven‐item value subscale, which is related to cognitive‐behavioral engagement (e.g. “This week, I believed practicing mindfulness could be of some value to me.”), and the seven‐item interest/enjoyment subscale, which is related to affective engagement (e.g. “This week, L2B was boring,” reverse scored). As affective engagement is posited to be related to context, adolescents also completed the eight‐item Teacher and Classmate Support Scale, which demonstrates acceptable internal consistency (α = .74) (Torsheim et al., 2000). On a scale of 1 = strongly disagree to 5 = strongly agree they rated statements such as “The L2B facilitators are nice and friendly.” For all these measures, an average was calculated for each measure, whereby higher scores reflected higher engagement.

To facilitate comparability across measures and the examination of change over time, all scale mean scores were standardized using the mean and standard deviation from Session 1, the initial timepoint. These scaled scores were then averaged to create a composite engagement index representing overall engagement across cognitive‐behavioral and affective dimensions.

Attendance

Attendance was recorded at each of the six sessions by facilitators and summed to create a total attendance variable (range: one to six sessions).

Data analyses

Analyses were conducted in R (R Core Team, 2022). Variables were examined for normality and outliers. Bivariate correlations tested whether age or cohort were associated with engagement and any target process at Session 1 or 6; point‐biserial correlations tested associations with sex. Variables correlated at r > |.3| with engagement and at least one target intervention process were included as covariates.

We considered potential cohort effects, but the presence of eight cohorts and limited statistical power restricted our ability to detect these effects reliably. To address this, we examined intraclass correlation coefficients (ICCs) to assess the degree to which cohort accounted for variability in the outcomes.

Structural equation models (SEMs) were estimated in lavaan (Rosseel, 2012) using maximum likelihood estimation to accommodate missingness (Enders, 2023). Model fit was evaluated based on a holistic examination of overall (global) model fit indices, including the comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR) (Little, 2023)—and parameter estimates (local fit). Alpha was set at α = .05 (Thiese et al., 2016).

Latent trajectory models (LTMs)

For each target process and engagement, we estimated intercept‐only (i.e. no growth specified), linear, latent basis, and nonlinear spline growth models and compared fit. Bivariate LTMs were then estimated pairing engagement with each target process (four models total) to test associations between stable individual differences (levels) and longitudinal change (slopes) and between‐process slope–slope associations. Figure 1 presents a general path diagram.

FIGURE 1.

FIGURE 1

Open in a new tab

Bivariate latent trajectory model. Target intervention outcome (e.g. emotion regulation, mindfulness, and self‐compassion) and engagement are modeled across six weekly intervals (T1:T6). Labels on paths show model constraints. Outcomes were modeled using latent basis models, whereby the loadings for Intervals 2–4 are freely estimated, between 0 and 1. Unlabeled paths with double‐headed arrows show freely estimated variances and covariances.

Random‐intercept cross‐lagged panel models (RI‐CLPMs)

To test time‐ordered, within‐person associations—including potential reciprocal effects—we estimated RI‐CLPMs (Hamaker et al., 2015). These models separate stable between‐person differences (random intercepts) from within‐person fluctuations (Mulder & Hamaker, 2021). Cross‐lagged paths therefore represent whether deviations from an individual's own average on one variable predict deviations in another variable at the subsequent timepoint. A general path diagram is presented in Figure 2.

FIGURE 2.

FIGURE 2

Open in a new tab

Random‐intercept cross‐lagged panel model. Target intervention outcome (e.g. emotion regulation, mindfulness, and self‐compassion) and engagement are modeled across six weekly intervals (T1:T6). Labels on paths show model constraints. Unlabeled paths with double‐headed arrows show freely estimated variances and covariances.

RESULTS

Preliminary analyses

Sample demographics are summarized in Table 1. Summary statistics for repeated measures are provided in Table S2. Adolescents attended an average of 4.6 sessions (SD = 1.4; range = 1–6), with 64% attending at least five sessions. Across the six measurement occasions, completion rates ranged from 75% to 95% (see Table S2). The correlations of cohort, sex, and age with engagement and target processes did not exceed r = |.3| (ps > .05); thus, these variables were not included as covariates in analyses. We also examined the correlations of attendance with engagement at Sessions 1 and 6. These were very small (r = −.01 and r = −.02) and not statistically significant (ps > .05).

TABLE 1.

Demographic information.

n %
Adolescents
Sex
Female 40 54
Male 34 46
Gender
Girl 36 49
Boy 30 41
Non‐binary, transgender, or gender fluid 4 5
Not provided 4 5
Ethnicity
Hispanic 13 18
Non‐Hispanic 58 78
Not reported 2 3
Race
Native American 3 4
Asian 9 12
Black/African American 1 1
White 62 84
Another race 5 7
2 or more races 9 9
Age, years (M ± SD) 14.1 ± 2.0
Caregivers
Educational attainment
Less than a 4‐year degree 21
Bachelor's degree 20
Graduate degree 55
Sex
Female 86
Male 9
Not provided 5
Age, years (M ± SD) 45.1 ± 6.2
Open in a new tab

Primary analyses Part 1: Trajectories via latent growth curve modeling

Fit information for all estimated univariate models is shown in Table S3. Fit statistics for best fitting univariate and bivariate LTMs are summarized in Table 2. In univariate models of outcomes, latent basis models demonstrated improved fit relative to linear models for parent‐reported DERS and self‐compassion, as indicated by χ 2 difference tests (Δχ 2s = 12.44–15.71, Δdfs = 4–6, ps = .003–.050). In contrast, for adolescent‐reported DERS and engagement, the latent basis models did not improve fit relative to the linear models (Δχ 2s = 5.33 and 1.26, Δdf = 4, ps = .26 and .87; ΔCFIs = .00, respectively), and the more parsimonious linear model was retained for bivariate models. For mindfulness, the linear model did not improve fit relative to the intercept‐only model (Δχ 2 = 6.57, Δdf = 3, p = .09); thus, the intercept‐only model was retained to include in bivariate analysis. Overall, these findings support the selected models despite some limitations in absolute fit. Thus, in bivariate LTMs, to model change in outcomes, we used the latent basis model, a flexible model that allows for complex, nonlinear patterns of change (Grimm et al., 2011; Grimm et al., 2016). To model change in engagement, we used the linear growth model. Parameter estimates are shown in Table 3.

TABLE 2.

Model fit indices.

Outcome Model a χ 2 df RMSEA 90% CI SRMR CFI BIC
Univariate trajectory models
DERS‐A Linear 33.76 16 .12 .06, .18 .10 .96 623.28
DERS‐P Latent basis 12.94 12 .03 .00, .13 .04 .99 541.72
Mindfulness Intercept‐only 26.27 19 .07 .00, .13 .06 .96 828.82
Self‐compassion Latent basis 11.45 12 .00 .00, .12 .08 .99 833.38
Engagement Linear 43.46 16 .15 .10, .21 .10 .91 889.23
Bivariate trajectory models of outcomes with engagement
DERS‐A 138.06 64 .13 .09, .16 .09 .90 1510.22
DERS‐P 91.41 60 .09 .05, .12 .07 .96 1439.40
Mindfulness 112.94 60 .11 .07, .13 .07 .93 1725.12
Self‐compassion 114.87 60 .11 .08, .14 .08 .92 1697.60
Random‐intercept cross‐lagged panel models of outcomes with engagement
DERS‐A 112.04 53 .12 .09, .16 .09 .93 1531.39
DERS‐P 69.06 53 .06 .00, .10 .07 .98 1447.08
Mindfulness 107.52 53 .12 .09, .15 .12 .92 1750.05
Self‐compassion 91.09 53 .10 .06, .13 .11 .95 1793.85
Open in a new tab

Note: Fit statistics for all estimated univariate models are shown in Table S3.

Abbreviations: 90% CI, 90% confidence interval; BIC, Bayesian information criteria; CFI, comparative fit index; DERS‐A, adolescent‐reported difficulties in emotion regulation; DERS‐P, parent‐reported difficulties in emotion regulation; df, degrees of freedom; p, p‐value for chi‐squared test of model fit; RMSEA, root mean square error of approximation; SRMR, standardized root mean squared residual; χ 2, chi square.

a

Best fitting univariate models were latent basis models for DERS‐P and self‐compassion; for DERS‐A and engagement, the linear model was best fitting; for mindfulness, the intercept‐only model was best fitting.

TABLE 3.

Parameter estimates for bivariate trajectory models of outcomes with engagement.

DERS‐A and engagement DERS‐P and engagement Mindfulness and engagement Self‐compassion and engagement
Slope factor loadings for DERS‐A Est. SE p Std.Est. Slope factor loadings for DERS‐P Est. SE p Std.Est. Est. SE p Std.Est. Slope factor loadings for SCS Est. SE p Std.Est.
T1 0 0 T1 0 0 T1 0 0
T2 1 .12 T2 .14 .14 .29 .07 T2 −.15 .13 .26 −.12
T3 2 .23 T3 .65 .12 <.001 .29 T3 .01 .11 .95 .01
T4 3 .36 T4 .87 .13 <.001 .41 T4 .5 .11 <.001 .34
T5 4 .49 T5 .83 .11 <.001 .41 T5 .66 .11 <.001 .43
T6 5 .58 T6 1 .48 T6 1 .64
Means Means Means Means
Level of DERS‐A .01 .11 .91 .01 Level of DERS‐P .03 .12 .81 .03 Level of MAAS −.01 .11 .92 −.01 Level of SCS .04 .11 .75 .04
Slope of DERS‐A −.04 .02 .02 −.35 Slope of DERS‐P −.42 .09 <.001 −1.05 Level of engage .04 .11 .75 .04 Slope of SCS .26 .1 .01 .33
Level of engage .04 .11 .72 .05 Level of engage .04 .11 .73 .05 Slope of engage .11 .02 <.001 .87 Level of engage .03 .11 .74 .04
Slope of engage .11 .02 <.001 .91 Slope of engage .11 .02 <.001 .91 Slope of engage .11 .02 <.001 .84
Variances Variances Variances Variances
Level DERS‐A .82 .15 <.001 1 Level DERS‐P .73 .14 <.001 1 Level MAAS .87 .16 <.001 1 Level SCS .74 .14 <.001 1
Slope DERS‐A .01 .01 <.001 1 Slope DERS‐P .16 .07 .02 1 Level engage .65 .15 <.001 1 Slope SCS .6 .16 <.001 1
Level engage .64 .14 <.001 1 Level engage .65 .14 <.001 1 Slope engage .02 .01 .01 1 Level engage .67 .15 <.001 1
Slope engage .02 .01 .02 1 Slope engage .02 .01 .02 1 Slope engage .02 .01 .01 1
Covariances Covariances Covariances Covariances
Level DERS‐A–slope DERS‐A −.03 .02 .06 −.34 Level DERS‐P with slope DERS‐P −.14 .08 .07 −.4 Level MAAS–slope engage .02 .02 .38 .16 Level SCS–slope SCS .08 .1 .4 .13
Level DERS‐A–slope engage −.01 .02 .75 −.06 Level DERS‐P with slope engage −.02 .02 .37 −.2 Level MAAS–level engage .15 .11 .16 .2 Level SCS–slope engage 0 .02 .89 −.02
Level DERS‐A–level engage −.07 .1 .52 −.09 Level DERS‐P with level engage −.01 .1 .9 −.18 Level engage–slope engage .01 .02 .59 .12 Level SCS–intercept engage .21 .1 .04 .3
Slope DERS‐A–slope engage −.01 0 .01 −.59 Slope DERS‐P with slope engage −.03 .02 .07 −.55 Slope SCS–slope engage .09 .02 <.001 .85
Level engage–slope engage −.02 .02 .13 −.27 Level engage with slope engage .01 .02 .67 .09 Level engage–slope engage 0 .02 .89 .03
Level engage–slope DERS‐A .01 .02 .62 .11 Level engage with slope DERS‐P .04 .07 .59 .11 Level engage–slope SCS .12 .1 .21 .19
Open in a new tab

Abbreviations: DERS‐A, adolescent‐reported difficulties in emotion regulation; DERS‐P, parent‐reported difficulties in emotion regulation; Est., estimate; MAAS, mindfulness; SCS, self‐compassion; SE, standard error; Std.Est., standardized estimate; T, time.

Engagement

Across models, on average, engagement showed a small but consistent increase over time (Ests. ≈ .11, SEs ≈ .02, ps < .001).

Adolescent‐reported emotion regulation difficulties (DERS)

In the bivariate model of adolescent‐reported DERS and engagement, on average, DERS showed a significant decrease over the course of the intervention (Est. = −.04, SE = .02, p = .02). Across processes, level–level and level–slope associations were nonsignificant; however, overall change in engagement was significantly, negatively associated with overall change in DERS (Est. = −.01, SE = .01, p = .014). That is, decreases in engagement over time tracked with decreases in DERS.

Parent‐reported emotion regulation difficulties

On average, parents' reports of adolescent DERS showed a significant reduction across the intervention (Est. = −.42, SE = .09, p < .001), with approximately 50% of the change occurring between Sessions 2 and 3. Yet, parent‐reported DERS and engagement were not significantly associated, either at baseline or with respect to changes over time (ps > .07).

Mindfulness

On average, mindfulness remained stable across the intervention period. Level of mindfulness was not significantly associated with level of engagement or change in engagement.

Self‐compassion

On average, self‐compassion increased significantly over the course of the intervention (Est. = .26, SE = .10, p = .01), with the largest portion of change (approximately 50%) occurring between Sessions 3 and 4. Baseline levels of engagement and self‐compassion were positively associated (r = .30, SE = .10, p = .04). Level–slope associations were nonsignificant. Further, changes in engagement and changes in self‐compassion were significantly, strongly positively associated with each other (r = .85, SE = .02, p < .001), such that engagement and self‐compassion increased in tandem.

Primary analyses Part 2: Dynamic change associations using RI‐CLPMs

Fit indices for the RI‐CLPMs are shown in Table 2. RI‐CLPMs demonstrated acceptable to good fit across outcomes (χ 2s = 69.06–112.04, dfs = 53, ps < .001–.068), with RMSEA values ranging from .06 to .12 (90% CIs = .00–.16) and SRMRs from .07 to .12, and CFIs from .92 to .98. Parameter estimates are provided in Table 4.

TABLE 4.

Parameter estimates for random‐intercept cross‐lagged panel models.

Adolescent‐reported difficulties in emotion regulation (DERS‐A) and engagement (engage) Parent‐reported difficulties in emotion regulation (DERS‐P) and engagement
Autoregressive effects Est. SE p Std.Est. Autoregressive effects Est. SE p Std. Est.
DERS‐A at T ⟶ DERS‐A at T + 1 .28 .08 <.001 .28 DERS‐P at T ⟶ DERS‐P at T + 1 .31 .08 <.001 .34
Engage at T ⟶ engage at T + 1 .63 .07 <.001 .60 Engage at T ⟶ engage at T + 1 .64 .07 <.001 .61
Cross‐lagged effects Cross‐lagged effects
DERS‐A at T ⟶ engage at T + 1 −.16 .15 .295 −.08 DERS‐P at T ⟶ engage at T + 1 .06 .18 .729 .03
Engage at T ⟶ DERS‐A at T + 1 −.21 .05 <.001 −.39 Engage at T ⟶ DERS‐P at T + 1 −.09 .06 .143 −.19
Variances Variances
Ran. Int. DERS‐A .67 .12 <.001 1.00 Ran. Int. DERS‐P .59 .11 <.001 1.00
Ran. Int. engage .47 .14 <.001 1.00 Ran. Int. engage .53 .13 <.001 1.00
Covariances Covariances
Ran. Int. DERS‐A–Ran. Int. engage −.09 .10 .344 −.16 Intercept DERS‐P with slope DERS‐P −.03 .09 .747 −.05
Mindfulness (MAAS) and engagement Self‐compassion (SCS) and engagement
Autoregressive effects Est. SE p Std.Est. Autoregressive effects Est. SE p Std.Est.
MAAS at T ⟶ MAAS at T + 1 .15 .08 .064 .15 SCS at T ⟶ SCS at T + 1 .33 .08 <.001 .32
Engage at T ⟶ engage at T + 1 .26 .15 .089 .28 Engage at T ⟶ engage at T + 1 .59 .09 <.001 .57
Cross‐lagged effects Cross‐lagged effects
MAAS at T ⟶ engage at T + 1 .01 .09 .928 .01 SCS at T ⟶ engage at T + 1 .10 .10 .340 .08
Engage at T ⟶ MAAS at T + 1 .03 .07 .694 .03 Engage at T ⟶ SCS at T + 1 .28 .07 <.001 .34
Variances Variances
Ran. Int. MAAS .83 .15 <.001 1.00 Ran. Int. SCS .76 .15 <.001 1.00
Ran. Int. engage .81 .18 <.001 1.00 Ran. Int. engage .50 .13 <.001 1.00
Covariances Covariances
Ran. Int. MAAS–Ran. Int. engage .20 .11 .075 .25 Ran. Int. SCS–Ran. Int. engage .16 .11 .149 .25
Open in a new tab

Abbreviations: DERS‐A, adolescent‐reported difficulties in emotion regulation; DERS‐P, parent‐reported difficulties in emotion regulation; Est, estimate; MAAS, mindfulness; Ran. Int., random intercept; SCS, self‐compassion; SE, standard error; Std.Est., standardized estimate.

Adolescent‐reported emotion regulation difficulties

The variances of the random intercepts of both adolescent‐reported DERS (Est. = .67, SE = .12, p < .001) and engagement (Est. = .47, SE = .14, p < .001) were significant, indicating that there were stable (i.e. relatively invariant across the repeated measures) between‐person differences in each of these processes. However, the intercepts of adolescent‐reported DERS and engagement did not significantly covary (p = .34), meaning that there was not a stable, between‐person association in these processes. In terms of the cross‐lagged associations, adolescent engagement at one timepoint significantly, negatively predicted adolescent‐reported DERS at the next timepoint (Est. = −.21, SE = .05, p < .001), meaning that higher engagement at any given timepoint, relative to one's own average, related to lower DERS a week later. The reverse path (DERS ⟶ engagement) was not significant (p = .30).

Parent‐reported emotion regulation difficulties

As with adolescent‐reported DERS, there were stable between‐person differences in levels of parent‐reported DERS, as evidenced by significant variance in the random intercept of this process (Est. = .59, SE = .11, p < .001). The random intercepts of parent‐reported DERS and engagement were not significantly associated (p = .75), meaning there was no stable, between‐person association between these two constructs across the study period. Additionally, there were no significant cross‐lagged paths between these two processes (ps > .14), meaning there was no evidence for dynamic within‐person associations between these two processes.

Mindfulness

The random intercepts of both processes showed significant variances (mindfulness: Est. = .83, SE = .15, p < .001, engagement: Est. = .81, SE = .18, p < .001), meaning that there were stable between‐person differences in the baseline levels for both of these processes. The random intercepts of mindfulness and engagement were not significantly associated (p = .08), meaning that there was not a stable, between‐person association between these constructs across time. There was no evidence for cross‐lagged associations between these two processes (ps > .69).

Self‐compassion

The variance of the random intercepts of self‐compassion and engagement were each significant (self‐compassion: Est. = .76, SE = .15, p < .001; engagement: Est. = .50, SE = .13, p < .001), providing evidence of stable between‐person differences in these processes. Yet, as with all of the other models, the random intercepts were not significantly related to each other (p = .15). Instead, engagement at a given timepoint significantly predicted self‐compassion at the next timepoint (Est. = .28, SE = .07, p < .001), indicating that when adolescents reported higher engagement relative to their own average, at the subsequent session, they reported greater self‐compassion relative to their average level of self‐compassion. In contrast, the cross‐lagged path in which self‐compassion was modeled to predict change in engagement was not significant (p = .34), indicating that adolescents' self‐compassion at a given session was not associated with changes in engagement at the next session.

Sensitivity analyses

ICCs indicated negligible clustering by study cohort for engagement (ICC = .002), but small‐to‐moderate clustering for target processes, including adolescent‐reported difficulties in emotion regulation (DERS) (ICC = .22), mindfulness (ICC = .21), self‐compassion (ICC = .15), and parent‐reported DERS (ICC = .11). In light of these results, and given limited power to model cohort effects directly, we conducted sensitivity analyses to assess the robustness of results (see Table S4).

Across multiple tests per outcome (i.e. seven cohort effects on four trajectory parameters = 28 tests per outcome), only a very small percentage reached statistical significance, a rate consistent with chance and likely reflecting the limited precision available when estimating seven cohort contrasts in a modest sample. Across the LTMs, adjustment for cohort differentially influenced some of the parameter estimates, but the overall pattern of associations was largely preserved. For the RI‐CLPMs, the pattern, magnitude, and statistical significance of focal parameters were unchanged after adjustment for cohort. Taken together, results from the sensitivity analyses indicated that cohort effects were variable and introduced additional statistical noise, but they did not systematically alter the primary conclusions.

DISCUSSION

This study examined trajectories of engagement, difficulties in emotion regulation, mindfulness, and self‐compassion among nonclinical adolescents participating in a 6‐week group MBI, testing associations at both between‐ and within‐person levels, including reciprocal effects. Between‐persons, adolescents who demonstrated greater increases in engagement relative to the sample average showed more pronounced improvements in self‐reported self‐compassion, as well as reductions in difficulties in adolescent‐reported emotion regulation, but not in mindfulness or parent‐reported difficulties in emotion regulation. Within‐person, higher‐than‐usual engagement predicted subsequent improvements in self‐reported difficulties in emotion regulation and self‐compassion. We found no evidence of reciprocal effects; changes in target processes did not predict subsequent changes in engagement.

These results provide initial support for multidimensional engagement as a potential driver of change in key MBI targets. Findings are consistent with a broader literature demonstrating that engagement is a key determinant of intervention effectiveness across both educational and clinical contexts (Fredricks, 2011; Marker et al., 2013; Wong et al., 2024). Prior research on engagement in MBI has relied primarily on home practice as a proxy for engagement (Tudor et al., 2022). Our study extends this work by examining longitudinal change–change associations and demonstrating temporal precedence using a multidimensional measure. Results suggest that engagement may potentially function as an explanatory mediator of intervention effects, though future randomized controlled trials are needed to formally test such mediating influences and evaluate adaptations to MBI explicitly designed to increase engagement. For example, program features such as interactivity and developmental relevance could be experimentally manipulated to test their causal impact on engagement and outcomes. Prior research from other contexts suggests several strategies for increasing adolescent engagement that may be relevant for MBI, such as enhancing autonomy and outcome expectancy and fostering supportive facilitator relationships (Eccles & Roeser, 2011; Fredricks, 2011; Gergov et al., 2021; O'Keeffe et al., 2018). These approaches may help adolescents more fully participate in and connect with mindfulness practices, thereby increasing the likelihood of benefit.

The absence of reciprocal effects was unexpected, given evidence of bidirectional relations between engagement and outcomes in educational and clinical contexts (Karababa, 2020; Stefansson et al., 2018). One possibility is that the 1‐week measurement intervals were too short for improvements in target processes to meaningfully boost engagement. Alternatively, adolescent engagement may be more strongly influenced by external factors (e.g. group dynamics and life stressors), which we did not assess, than by internal improvements in target outcomes alone. For example, rapport with the facilitator may influence adolescents' willingness to participate and their experience of the intervention as positive and helpful (Garcia et al., 2011; Gee et al., 2021). Peer norms and group dynamics also likely shape engagement, as participation may depend on whether it is perceived as socially acceptable within the group (Arias‐Pujol & Anguera, 2017; Shechtman, 2025). Future research is needed to understand these types of potentially modifiable influences on engagement, including contextual factors internal to the MBI (e.g. facilitator rapport, group dynamics) and external to the MBI (e.g. school stress). If we can target some of these processes, we may be better able to increase engagement and thereby improve MBI efficacy. Importantly, it is possible that low engagement covaries with higher baseline distress or avoidant coping, meaning low engaged adolescents could be a subgroup that would benefit most from MBI, because these are directly targeted by mindfulness‐based approaches (de Vibe et al., 2018; Dunning et al., 2022; Kwon et al., 2018).

Adolescent engagement did not predict parent‐reported adolescent difficulties in emotion regulation, suggesting that parent and adolescent reports may capture different facets of this construct. Specifically, parents are better reporters of observable behavior and externalizing symptoms, rather than internalizing symptoms, whereas adolescent reports can better reflect internal emotional processes and regulatory efforts (Hourigan et al., 2011). Observational methods could offer an additional perspective.

Our data did not support systematic mean‐level change in self‐reported mindfulness across the intervention period. There are many potential reasons for this. First, the MAAS primarily assesses general attentional awareness in daily life and may be relatively insensitive to early changes occurring over a brief 6‐week intervention. Second, participants entered the study with relatively high baseline mindfulness, potentially limiting detectable improvement. Finally, it is possible that participating in an MBI makes participants more aware of the ways in which they tend to be mindless, which may lead to lower self‐reported mindfulness (Baer et al., 2012; Grossman, 2011). Additionally, it has been argued that self‐report measures of mindfulness do not adequately capture the construct itself (Visted et al., 2014) and thus may be less likely to show change in response to MBI. Specifically, the measure of mindfulness used here, the MAAS, actually measures mindlessness and is unidimensional, two features for which it has been critiqued (Grossman & Van Dam, 2011; Visted et al., 2014).

Latent basis models revealed that most change in parent‐reported DERS and self‐compassion occurred early to mid‐intervention (e.g. Sessions 2–4). Timing effects are rarely examined in adolescent MBI, with most studies evaluating baseline to post‐intervention change only. This pattern of findings may be L2B specific: It is possible that the material covered in Sessions 2–4 is more impactful relative to other sessions in terms of supporting change in these specific target processes. However, the pattern of change observed here is also consistent with research investigating dose–response in other types of psychological interventions, which finds a nonlinear association: Early improvement is common, with rates of improvement slowing later on (Klein et al., 2024). These findings raise questions about optimal MBI length, particularly for adolescents of varied ages who may require shorter or longer training to benefit.

Our findings have several practical implications for the design and delivery of adolescent MBI. First, because engagement appears to drive change in outcomes, programs should prioritize strategies to actively monitor and support engagement throughout the intervention (e.g. through daily monitoring of engagement by facilitators with ecological momentary assessment methods). Second, given that changes occurred early in the intervention, practitioners may consider front‐loading core skills and ensuring that initial sessions are highly engaging and accessible. It may be particularly helpful to identify adolescents with low early engagement so that targeted supports (e.g. motivational enhancement and individualized outreach) could be introduced to increase participation. Additionally, program developers should consider enhancing program features that increase relevance and interactivity, as doing so may help sustain engagement and improve outcomes.

LIMITATIONS

Several models showed some degree of misfit (e.g. elevated RMSEA). However, prior methodological work indicates that RMSEA can be upwardly biased in models with small sample sizes, low degrees of freedom, and complex longitudinal structures, which can lead to over‐rejection of otherwise reasonable models (Shi et al., 2022). Accordingly, we interpreted RMSEA alongside additional indices (CFI and SRMR) and overall model behavior rather than in isolation. Taken together with these other indices and considering the context of a longitudinal study with a small sample, we considered these degrees of fit acceptable (Little, 2023). The single‐arm study design is also a key limitation to our ability to establish causal associations; randomized controlled trials with active comparators are needed to test engagement as a mechanism. Nonetheless, the use of repeated measurement timepoints and the RI‐CLPM framework allowed us to examine temporal precedence—a necessary condition for establishing etiological models—which is a strength.

In addition, we did not directly assess home practice adherence, which represents an important potential confounder. Homework completion reflects engagement with MBI outside of sessions and may independently contribute to improvements in target processes. It is possible that adolescents who were more engaged during sessions were also more likely to complete home practice, which could partially account for the observed associations. It is also likely that different factors affect engagement in group sessions relative to engagement outside of group sessions, and exploring these distinctions is an important area for future work. Future research should assess both in‐session and out‐of‐session engagement to disentangle their unique and combined effects.

The sample was relatively small, and demographically homogenous, and all adolescents participated in this MBI voluntarily, thus limiting the ecological validity and generalizability of our findings. Engagement processes may differ in more diverse adolescent samples and in mandated contexts (e.g. school‐based MBI), as structural barriers may constrain participation for some youth, while culturally relevant and identity‐affirming environments may enhance engagement (Fredricks, 2011; Whitehead et al., 2023; Wong et al., 2024). Although power to detect cohort effects was limited, the consistency of within‐person associations is notable and suggests these processes may be relatively robust across samples. In contrast, between‐person effects appeared more sensitive to baseline differences, underscoring the importance of accounting for cohort heterogeneity in future work. Larger, more diverse studies, utilizing approaches such as multilevel SEM, are needed to more rigorously evaluate these effects and clarify the role of baseline variability in shaping developmental trajectories.

CONCLUSION

Although MBI is a promising preventive approach for adolescent mental health, effects are typically small, and recent large‐scale trials have reported null effects (Fulambarkar et al., 2023; Kuyken et al., 2022). Increasingly, researchers have questioned whether adolescents are engaging with MBI in ways that allow them to benefit (Galla, 2024; Kuyken et al., 2022). The current findings provide initial evidence that multidimensional engagement is an important underlying process for nonclinical adolescents in group MBI. Effects were strongest at the within‐person level, suggesting that person‐centered, adaptive approaches to optimizing engagement for adolescents in MBI warrant investigation.

CONFLICT OF INTEREST STATEMENT

None to report.

ETHICS STATEMENT

All participants provided informed consent/assent. Study procedures were approved by the Colorado State University Institutional Review Board (July 2023, IRB 4219).

Supporting information

Table S1. Learning to BREATHE 6‐week Mindfulness‐Based Intervention: Weekly Session Content.

APHW-18-0-s001.docx (20.3KB, docx)

Table S2. Descriptive Statistics for Study Variables Across Measurement Occasions.

APHW-18-0-s002.docx (16.3KB, docx)

Table S3: Univariate Structural Equation Model Comparisons.

APHW-18-0-s003.docx (39.5KB, docx)

Table S4. Follow‐Up Sensitivity Analyses: Cohort Effects.

APHW-18-0-s004.docx (49.7KB, docx)

DATA AVAILABILITY STATEMENT

Data are available by request from the corresponding author.

REFERENCES

  1. Arias‐Pujol, E. , & Anguera, M. T. (2017). Observation of interactions in adolescent group therapy: A mixed methods study. Frontiers in Psychology, 8, 1188. 10.3389/fpsyg.2017.01188 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Baer, R. A. , Carmody, J. , & Hunsinger, M. (2012). Weekly change in mindfulness and perceived stress in a mindfulness‐based stress reduction program. Journal of Clinical Psychology, 68(7), 755–765. 10.1002/jclp.21865 [DOI] [PubMed] [Google Scholar]
  3. Ben‐Eliyahu, A. , Moore, D. , Dorph, R. , & Schunn, C. D. (2018). Investigating the multidimensionality of engagement: Affective, behavioral, and cognitive engagement across science activities and contexts. Contemporary Educational Psychology, 53(April), 87–105. 10.1016/j.cedpsych.2018.01.002 [DOI] [Google Scholar]
  4. Bluth, K. , Gaylord, S. A. , Campo, R. A. , Mullarkey, M. C. , & Hobbs, L. (2016). Making friends with yourself: A mixed methods pilot study of a mindful self‐compassion program for adolescents. Mindfulness, 7(2), 479–492. 10.1007/s12671-015-0476-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bosch, C. (2024). Assessing the psychometric properties of the Intrinsic Motivation Inventory in blended learning environments. Journal of Education and E‐Learning Research, 11(2), 263–271. 10.20448/jeelr.v11i2.5468 [DOI] [Google Scholar]
  6. Broderick, P. C. (2021). Learning to BREATHE: A mindfulness curriculum for adolescents to cultivate emotion regulation, attention, and performance. New Harbinger Publications. [Google Scholar]
  7. Bunford, N. , Dawson, S. W. , Evans, A. , Ray, A. R. , Langberg, J. M. , Owens, J. S. , DuPaul, G. J. , & Allan, D. M. (2020). The Difficulties in Emotion Regulation Scale‐Parent Report: A psychometric investigation examining adolescents with and without ADHD. Assessment, 27(5), 921–940. 10.1177/1073191118792307 [DOI] [PubMed] [Google Scholar]
  8. de Bruin, E. I. , Zijlstra, B. J. H. , van de Weijer‐Bergsma, E. , & Bögels, S. M. (2011). The Mindful Attention Awareness Scale for Adolescents (MAAS‐A): Psychometric properties in a Dutch sample. Mindfulness, 2(3), 201–211. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. de Vibe, M. , Solhaug, I. , Rosenvinge, J. H. , Tyssen, R. , Hanley, A. , & Garland, E. (2018). Six‐year positive effects of a mindfulness‐based intervention on mindfulness, coping and well‐being in medical and psychology students; results from a randomized controlled trial. PLoS ONE, 13(4), e0196053. 10.1371/journal.pone.0196053 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dumontheil, I. (2022). Randomized controlled trial of a mindfulness‐based intervention in adolescents: The Mindfulteen study. Journal of Child Psychology and Psychiatry, 63(10), 1183–1193. [Google Scholar]
  11. Dunning, D. , Tudor, K. , Radley, L. , Dalrymple, N. , Funk, J. , Vainre, M. , Ford, T. , Montero‐Marin, J. , Kuyken, W. , & Dalgleish, T. (2022). Do mindfulness‐based programmes improve the cognitive skills, behaviour and mental health of children and adolescents? An updated meta‐analysis of randomised controlled trials. Evidence‐Based Mental Health, 25(3), 135–142. 10.1136/ebmental-2022-300464 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Eccles, J. S. , & Roeser, R. W. (2011). Schools as developmental contexts during adolescence. Journal of Research on Adolescence: The Official Journal of the Society for Research on Adolescence, 21(1), 225–241. 10.1111/j.1532-7795.2010.00725.x [DOI] [Google Scholar]
  13. Enders, C. K. (2023). Missing data: An update on the state of the art. Psychological Methods, 30, 322–339. 10.1037/met0000563 [DOI] [PubMed] [Google Scholar]
  14. Fredricks, J. A. (2011). Engagement in school and out‐of‐school contexts: A multidimensional view of engagement. Theory into Practice, 50(4), 327–335. 10.1080/00405841.2011.607401 [DOI] [Google Scholar]
  15. Fresco, D. M. , & Mennin, D. S. (2019). All together now: Utilizing common functional change principles to unify cognitive behavioral and mindfulness‐based therapies. Current Opinion in Psychology, 28(August), 65–70. 10.1016/j.copsyc.2018.10.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fulambarkar, N. , Seo, B. , Testerman, A. , Rees, M. , Bausback, K. , & Bunge, E. (2023). Review: Meta‐analysis on mindfulness‐based interventions for adolescents' stress, depression, and anxiety in school settings: A cautionary tale. Child and Adolescent Mental Health, 28(2), 307–317. 10.1111/camh.12572 [DOI] [PubMed] [Google Scholar]
  17. Galla, B. (2024). How motivation restricts the scalability of universal school‐based mindfulness interventions for adolescents. Child Development Perspectives, 18(3), 129–136. 10.1111/cdep.12508 [DOI] [Google Scholar]
  18. Galla, B. M. (2016). Within‐person changes in mindfulness and self‐compassion predict enhanced emotional well‐being in healthy, but stressed adolescents. Journal of Adolescence, 49(1), 204–217. 10.1016/j.adolescence.2016.03.016 [DOI] [PubMed] [Google Scholar]
  19. Garcia, C. , Lindgren, S. , & Pintor, J. K. (2011). Knowledge, skills, and qualities for effectively facilitating an adolescent girls' group. The Journal of School Nursing: The Official Publication of the National Association of School Nurses, 27(6), 424–433. 10.1177/1059840511419369 [DOI] [PubMed] [Google Scholar]
  20. Gee, B. , Wilson, J. , Clarke, T. , Farthing, S. , Carroll, B. , Jackson, C. , King, K. , Murdoch, J. , Fonagy, P. , & Notley, C. (2021). Review: Delivering mental health support within schools and colleges—A thematic synthesis of barriers and facilitators to implementation of indicated psychological interventions for adolescents. Child and Adolescent Mental Health, 26(1), 34–46. 10.1111/camh.12381 [DOI] [PubMed] [Google Scholar]
  21. Gergov, V. , Lindberg, N. , Lahti, J. , Lipsanen, J. , & Marttunen, M. (2021). Effectiveness and predictors of outcome for psychotherapeutic interventions in clinical settings among adolescents. Frontiers in Psychology, 12(February), 628977. 10.3389/fpsyg.2021.628977 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Goldberg, S. B. (2022). A common factors perspective on mindfulness‐based interventions. Nature Reviews Psychology, 1(10), 605–619. 10.1038/s44159-022-00090-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Gonçalves, S. F. , Chaplin, T. M. , Turpyn, C. C. , Niehaus, C. E. , Curby, T. W. , Sinha, R. , & Ansell, E. B. (2019). Difficulties in emotion regulation predict depressive symptom trajectory from early to middle adolescence. Child Psychiatry and Human Development, 50(4), 618–630. 10.1007/s10578-019-00867-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Gratz, K. L. , & Roemer, L. (2004). Multidimensional assessment of emotion regulation and dysregulation: Development, factor structure, and initial validation of the Difficulties in Emotion Regulation Scale. Journal of Psychopathology and Behavioral Assessment, 26, 41–54. 10.1023/b:joba.0000007455.08539.94 [DOI] [Google Scholar]
  25. Grayson, D. S. , & Fair, D. A. (2017). Development of large‐scale functional networks from birth to adulthood: A guide to the neuroimaging literature. NeuroImage, 160(October), 15–31. 10.1016/j.neuroimage.2017.01.079 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Grimm, K. J. , Ram, N. , & Estabrook, R. (2016). Growth modeling: Structural equation and multilevel modeling approaches. Guilford Publications. [Google Scholar]
  27. Grimm, K. J. , Ram, N. , & Hamagami, F. (2011). Nonlinear growth curves in developmental research. Child Development, 82(5), 1357–1371. 10.1111/j.1467-8624.2011.01630.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Grossman, P. (2011). Defining mindfulness by how poorly I think I pay attention during everyday awareness and other intractable problems for psychology's (re)invention of mindfulness: Comment on Brown et al. (2011). Psychological Assessment, 23(4), 1034–1046. [DOI] [PubMed] [Google Scholar]
  29. Grossman, P. , & Van Dam, N. T. (2011). Mindfulness, by any other name…: Trials and tribulations of sati in Western psychology and science. Contemporary Buddhism, 12(1), 219–239. 10.1080/14639947.2011.564841 [DOI] [Google Scholar]
  30. Hamaker, E. L. , Kuiper, R. M. , & Grasman, R. P. P. P. (2015). A critique of the cross‐lagged panel model. Psychological Methods, 20(1), 102–116. 10.1037/a0038889 [DOI] [PubMed] [Google Scholar]
  31. Harris, P. A. , Taylor, R. , Thielke, R. , Payne, J. , Gonzalez, N. , & Conde, J. G. (2009). Research electronic data capture (REDCap)—A metadata‐driven methodology and workflow process for providing translational research informatics support. Journal of Biomedical Informatics, 42(2), 377–381. 10.1016/j.jbi.2008.08.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Hourigan, S. E. , Goodman, K. L. , & Southam‐Gerow, M. A. (2011). Discrepancies in parents' and children's reports of child emotion regulation. Journal of Experimental Child Psychology, 110(2), 198–212. 10.1016/j.jecp.2011.03.002 [DOI] [PubMed] [Google Scholar]
  33. Intrinsic Motivation Inventory . (2022). Self‐Determination Theory. selfdeterminationtheory.org/intrinsic‐motivation‐inventory.
  34. Jones, P. B. (2013). Adult mental health disorders and their age at onset. The British Journal of Psychiatry, 54, s5–s10. [DOI] [PubMed] [Google Scholar]
  35. Kabat‐Zinn, J. (1990). Full catastrophe living: Using the wisdom of your body and mind to face stress, pain, and illness. Delacorte Press. [Google Scholar]
  36. Karababa, A. (2020). The reciprocal relationships between school engagement and self‐esteem among Turkish early adolescents: A three‐wave cross‐lagged model. Children and Youth Services Review, 116, 105114. 10.1016/j.childyouth.2020.105114 [DOI] [Google Scholar]
  37. Kaufman, J. , Birmaher, B. , Brent, D. , Rao, U. , Flynn, C. , Moreci, P. , Williamson, D. , & Ryan, N. (1997). Schedule for affective disorders and schizophrenia for school‐age children‐present and lifetime version (K‐SADS‐PL): Initial reliability and validity data. Journal of the American Academy of Child and Adolescent Psychiatry, 36(7), 980–988. 10.1097/00004583-199707000-00021 [DOI] [PubMed] [Google Scholar]
  38. Klein, T. , Breilmann, J. , Schneider, C. , Girlanda, F. , Fiedler, I. , Dawson, S. , Crippa, A. , Priebe, S. , Barbui, C. , Becker, T. , & Kösters, M. (2024). Dose‐response relationship in cognitive behavioral therapy for depression: A nonlinear metaregression analysis. Journal of Consulting and Clinical Psychology, 92(5), 296–309. 10.1037/ccp0000879 [DOI] [PubMed] [Google Scholar]
  39. Kuyken, W. , Ball, S. , Crane, C. , Ganguli, P. , Jones, B. , Montero‐Marin, J. , Nuthall, E. , Raja, A. , Taylor, L. , Tudor, K. , Viner, R. M. , Allwood, M. , Aukland, L. , Dunning, D. , Casey, T. , Dalrymple, N. , De Wilde, K. , Farley, E. R. , Harper, J. , … Williams, J. M. G. (2022). Effectiveness and cost‐effectiveness of universal school‐based mindfulness training compared with normal school provision in reducing risk of mental health problems and promoting well‐being in adolescence: The MYRIAD cluster randomised controlled trial. Evidence‐Based Mental Health, 25, 99–109. 10.1136/ebmental-2021-300396 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Kwon, K. , Kupzyk, K. , & Benton, A. (2018). Negative emotionality, emotion regulation, and achievement: Cross‐lagged relations and mediation of academic engagement. Learning and Individual Differences, 67(October), 33–40. 10.1016/j.lindif.2018.07.004 [DOI] [Google Scholar]
  41. Labouliere, C. D. , Reyes, J. P. , Shirk, S. , & Karver, M. (2017). Therapeutic alliance with depressed adolescents: Predictor or outcome? Disentangling temporal confounds to understand early improvement. Journal of Clinical Child and Adolescent Psychology: The Official Journal for the Society of Clinical Child and Adolescent Psychology, 46(4), 600–610. 10.1080/15374416.2015.1041594 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Little, T. D. (2023). Longitudinal structural equation modeling. Guilford Publications. [Google Scholar]
  43. Low, S. , Van Ryzin, M. J. , Brown, E. C. , Smith, B. H. , & Haggerty, K. P. (2014). Engagement matters: Lessons from assessing classroom implementation of steps to respect: A bullying prevention program over a one‐year period. Prevention Science: The Official Journal of the Society for Prevention Research, 15(2), 165–176. 10.1007/s11121-012-0359-1 [DOI] [PubMed] [Google Scholar]
  44. Lucas‐Thompson, R. G. , Prince, M. A. , Adams, M. S. , Miller, R. L. , Moran, M. J. , Rayburn, S. R. , & Seiter, N. S. (2024). Does a mindfulness‐based intervention strengthen mindfulness stress buffering effects in adolescence? A preliminary investigation. Current Psychology, 43(4), 3440–3454. 10.1007/s12144-023-04520-5 [DOI] [Google Scholar]
  45. Marker, C. D. , Comer, J. S. , Abramova, V. , & Kendall, P. C. (2013). The reciprocal relationship between alliance and symptom improvement across the treatment of childhood anxiety. Journal of Clinical Child and Adolescent Psychology: The Official Journal for the Society of Clinical Child and Adolescent Psychology, 42(1), 22–33. 10.1080/15374416.2012.723261 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Montero‐Marin, J. , Allwood, M. , Ball, S. , Crane, C. , De Wilde, K. , Hinze, V. , Jones, B. , Lord, L. , Nuthall, E. , Raja, A. , Taylor, L. , Tudor, K. , MYRIAD Team , Blakemore, S.‐J. , Byford, S. , Dalgleish, T. , Ford, T. , Greenberg, M. T. , Ukoumunne, O. C. , … Kuyken, W. (2022). School‐based mindfulness training in early adolescence: What works, for whom and how in the MYRIAD trial? Evidence‐Based Mental Health, 25, 117–124. 10.1136/ebmental-2022-300439 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Mulder, J. D. , & Hamaker, E. L. (2021). Three extensions of the random intercept cross‐lagged panel model. Structural Equation Modeling: A Multidisciplinary Journal, 28(4), 638–648. 10.1080/10705511.2020.1784738 [DOI] [Google Scholar]
  48. Neff, K. (2003). Self‐compassion: An alternative conceptualization of a healthy attitude toward oneself. Self and Identity: The Journal of the International Society for Self and Identity, 2(2), 85–101. 10.1080/15298860309032 [DOI] [Google Scholar]
  49. Neimeyer, R. A. , Kazantzis, N. , Kassler, D. M. , Baker, K. D. , & Fletcher, R. (2008). Group cognitive behavioural therapy for depression outcomes predicted by willingness to engage in homework, compliance with homework, and cognitive restructuring skill acquisition. Cognitive Behaviour Therapy, 37(4), 199–215. 10.1080/16506070801981240 [DOI] [PubMed] [Google Scholar]
  50. van Neumann, A., Lier, P. A. , Gratz, K. L. , & Koot, H. M. (2010). Multidimensional assessment of emotion regulation difficulties in adolescents using the Difficulties in Emotion Regulation Scale. Assessment, 17(1), 138–149. 10.1177/1073191109349579 [DOI] [PubMed] [Google Scholar]
  51. O'Keeffe, S. , Martin, P. , Goodyer, I. M. , Wilkinson, P. , Impact Consortium , & Midgley, N. (2018). Predicting dropout in adolescents receiving therapy for depression. Psychotherapy Research: Journal of the Society for Psychotherapy Research, 28(5), 708–721. 10.1080/10503307.2017.1393576 [DOI] [PubMed] [Google Scholar]
  52. Porter, B. , Oyanadel, C. , Betancourt, I. , Worrell, F. C. , & Peñate, W. (2024). Effects of two online mindfulness‐based interventions for early adolescents for attentional, emotional, and behavioral self‐regulation. Pediatric Reports, 16(2), 254–270. 10.3390/pediatric16020022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Prince, M. , Patel, V. , Saxena, S. , Maj, M. , Maselko, J. , Phillips, M. R. , & Rahman, A. (2007). No health without mental health. The Lancet, 370(9590), 859–877. 10.1016/S0140-6736(07)61238-0 [DOI] [PubMed] [Google Scholar]
  54. Quach, D. , Gibler, R. C. , & Jastrowski Mano, K. E. (2017). Does home practice compliance make a difference in the effectiveness of mindfulness interventions for adolescents? Mindfulness, 8(2), 495–504. 10.1007/s12671-016-0624-7 [DOI] [Google Scholar]
  55. Radloff, L. S. (1977). CES‐D scale: Self report depression scale research general population. Applied Psychological Measurement, 1, 385–401. [Google Scholar]
  56. Raes, F. , Pommier, E. , Neff, K. D. , & Van Gucht, D. (2011). Construction and factorial validation of a short form of the self‐compassion scale. Clinical Psychology & Psychotherapy, 18, 250–255. 10.1002/cpp.702 [DOI] [PubMed] [Google Scholar]
  57. R Core Team . (2022). R: A language and environment for statistical computing. http://www.R-project.org [Google Scholar]
  58. Roebroek, L. , & Koning, I. M. (2016). The reciprocal relation between adolescents' school engagement and alcohol consumption, and the role of parental support. Prevention Science: The Official Journal of the Society for Prevention Research, 17(2), 218–226. 10.1007/s11121-015-0598-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Rosseel, Y. (2012). lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software, 48(2), 1–36. 10.18637/jss.v048.i02 [DOI] [Google Scholar]
  60. Ryan, R. M. , & Deci, E. L. (2000). Self‐determination theory and the facilitation of intrinsic motivation, social development, and well‐being. The American Psychologist, 55(1), 68–78. 10.1037/0003-066X.55.1.68 [DOI] [PubMed] [Google Scholar]
  61. Shapiro, S. , Siegel, R. , & Neff, K. D. (2018). Paradoxes of mindfulness. Mindfulness, 9(6), 1693–1701. 10.1007/s12671-018-0957-5 [DOI] [Google Scholar]
  62. Shechtman, Z. (2025). Group therapy to promote adolescents' mental health: Clinical and empirical evidence. Adolescents, 5(4), 57. 10.3390/adolescents5040057 [DOI] [Google Scholar]
  63. Shi, D. , DiStefano, C. , Maydeu‐Olivares, A. , & Lee, T. (2022). Evaluating SEM model fit with small degrees of freedom. Multivariate Behavioral Research, 57(2–3), 179–207. 10.1080/00273171.2020.1868965 [DOI] [PubMed] [Google Scholar]
  64. Stefansson, K. K. , Gestsdottir, S. , Birgisdottir, F. , & Lerner, R. M. (2018). School engagement and intentional self‐regulation: A reciprocal relation in adolescence. Journal of Adolescence, 64, 23–33. 10.1016/j.adolescence.2018.01.005 [DOI] [PubMed] [Google Scholar]
  65. Tang, Y.‐Y. , Hölzel, B. K. , & Posner, M. I. (2015). The neuroscience of mindfulness meditation. Nature Reviews. Neuroscience, 16(4), 213–225. 10.1038/nrn3916 [DOI] [PubMed] [Google Scholar]
  66. Thiese M. S., Ronna, B. , & Ott, U. (2016). P value interpretations and considerations. Journal of Thoracic Disease, 8(9), E928–E931. 10.21037/jtd.2016.08.16 [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Torsheim, T. , Wold, B. , & Samdal, O. (2000). The teacher and classmate support scale: Factor structure, test–retest reliability and validity in samples of 13‐ and 15‐year‐old adolescents. School Psychology International, 21(2), 195–212. [Google Scholar]
  68. Tudor, K. , Maloney, S. , Raja, A. , Baer, R. , Blakemore, S.‐J. , Byford, S. , Crane, C. , Dalgleish, T. , de Wilde, K. , Ford, T. , Greenberg, M. , Hinze, V. , Lord, L. , Radley, L. , Opaleye, E. S. , Taylor, L. , Ukoumunne, O. C. , Viner, R. , MYRIAD Team , … Montero‐Marin, J. (2022). Universal mindfulness training in schools for adolescents: A scoping review and conceptual model of moderators, mediators, and implementation factors. Prevention Science: The Official Journal of the Society for Prevention Research, 23(6), 934–953. 10.1007/s11121-022-01361-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Visted, E. , Vøllestad, J. , Nielsen, M. B. , & Nielsen, G. H. (2014). The impact of group‐based mindfulness training on self‐reported mindfulness: A systematic review and meta‐analysis. Mindfulness, 6(3), 501–522. 10.1007/s12671-014-0283-5 [DOI] [Google Scholar]
  70. Whitehead, L. , Talevski, J. , Fatehi, F. , & Beauchamp, A. (2023). Barriers to and facilitators of digital health among culturally and linguistically diverse populations: Qualitative systematic review. Journal of Medical Internet Research, 25(1), e42719. 10.2196/42719 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Wong, Z. Y. , Liem, G. A. D. , Chan, M. , & Datu, J. A. D. (2024). Student engagement and its association with academic achievement and subjective well‐being: A systematic review and meta‐analysis. Journal of Educational Psychology, 116(1), 48–75. 10.1037/edu0000833 [DOI] [Google Scholar]
  72. Zoogman, S. , Goldberg, S. B. , Hoyt, W. T. , & Miller, L. (2014). Mindfulness interventions with youth: A meta‐analysis. Mindfulness, 6(2), 290–302. 10.1007/s12671-013-0260-4 [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Table S1. Learning to BREATHE 6‐week Mindfulness‐Based Intervention: Weekly Session Content.

APHW-18-0-s001.docx (20.3KB, docx)

Table S2. Descriptive Statistics for Study Variables Across Measurement Occasions.

APHW-18-0-s002.docx (16.3KB, docx)

Table S3: Univariate Structural Equation Model Comparisons.

APHW-18-0-s003.docx (39.5KB, docx)

Table S4. Follow‐Up Sensitivity Analyses: Cohort Effects.

APHW-18-0-s004.docx (49.7KB, docx)

Data Availability Statement

Data are available by request from the corresponding author.

Articles from Applied Psychology. Health and Well-Being are provided here courtesy of Wiley

RESOURCES