The Benefits of Mindfulness Training for Momentary Mindfulness and Emotion Regulation: A Randomized Controlled Trial for Adolescents Exposed to Chronic Stressors

Abstract

Objective:

The goal of this study was to test if a mindfulness-based intervention (MBI) compared to an active control ameliorates the impacts of life stressors on momentary mindfulness and emotion regulation difficulties among adolescents exposed to chronic stressors.

Method:

Adolescents exposed to chronic stressors (N=81, M_age=13.75 years; 56% boys; 24% Hispanic/Latino, 57% White) were randomized to receive MBI within the context of a community-based mentoring program (MBI+mentoring) or mentoring-alone. Participants completed ecological momentary assessments (EMA) three times each day for seven days at three intervals/bursts (baseline, mid-intervention, and post-intervention), contributing to a total of 3,178 EMA reports. EMA assessed momentary exposure to life stressors, mindfulness (vs. mindlessness), and emotion regulation difficulties.

Results:

Linear mixed-effects models revealed that the interaction between intervention arm (MBI+mentoring vs. mentoring-alone) and burst was significantly associated with the random slopes of life stressor exposure predicting mindful attention (b=−.05, SE=.01, p<.001), mindful non-judgment (b=−.03, SE=.01, p<.001), and emotion regulation difficulties (b=−.04, SE=.01, p<.001). Estimated marginal means revealed that MBI+mentoring, compared to mentoring-alone, produced small, but significant attenuation in the association of life stressors with mindful attention, mindful non-judgment, and emotion regulation difficulties at post-intervention.

Conclusion:

Mindfulness training may buffer adolescents exposed to chronic stressors against the negative impacts of life stressors on mindfulness and emotion regulation in daily life. Going forward, it will be important to investigate these relationships in the context of mental/physical health outcomes and to include longer periods of follow-up to determine the sustainable benefits of MBI for adolescent health.

Adolescents embedded in communities with high exposure to chronic stressors are disproportionately at risk for developing mental and physical health problems (Sheth et al., 2017). According to the allostatic load hypothesis, exposure to chronic stressors, which are defined as ongoing or repeated life events (e.g., living on a low income, community/interpersonal violence) that elicit the stress response, can contribute to the prolonged activation of the emotional, endocrine, and cardiovascular stress regulation systems (McEwen, 2017). Chronic stressors are distinct, but overlapping with everyday life stressors, which are defined as everyday situations (e.g., conflict with family, earning a poor grade on a test) that elicit the stress response (Anisman & Merali, 1999); life stressors can become chronic stressors when they become repeated and ongoing. The prolonged activation of stress response systems after experiencing chronic stressors can contribute to wear and tear on the body and ultimately increase the vulnerability for developing mental health problems (McEwen, 2017). As such, exposure to chronic stressors is a well-documented risk factor for the development of anxiety and depression (McEwen, 2017). One pathway through which chronic stressors negatively influence adolescent mental health is by inducing changes in neural circuity necessary for adaptive emotion regulation (Sheth et al., 2017). Impairments in emotion regulation after experiencing chronic stressors may then exacerbate risk for mental health problems (Sheth et al., 2017). Despite increasing knowledge about risk factors and the long-term consequences of adolescent mental health (Lancefield et al., 2016), rates of anxiety and depression have only continued to increase in youth over the past decade (Bitsko et al., 2018; Whitney & Peterson, 2019), especially following the COVID-19 global pandemic (Chavira et al., 2022). This trend highlights the need for additional research on approaches for protecting adolescents from the negative impacts of stress.

Extant research suggests that mindfulness-based intervention (MBI) may be an effective approach for improving coping and mental health among adolescents at risk for emotional, behavioral, and academic concerns (Rawlett & Scrandis, 2015). Despite strong theoretical underpinnings and evidence from pilot randomized controlled trials (Bluth et al., 2016; Broderick & Metz, 2016; Rawlett & Scrandis, 2015), this work is in the early stages and most existing data have not fully elucidated the mechanisms by which MBIs may alter the impact of stressors in one’s daily life (Rawlett & Scrandis, 2015). The current study aims to address these gaps in the literature using real-time, real-world data on momentary life experiences to evaluate the potential benefits of MBI for ameliorating the impacts of life stressors on momentary mindfulness and emotion regulation difficulties among adolescents exposed to chronic stressors.

Mindfulness, Stressors, and Emotion Regulation

According to the mindfulness stress buffering hypothesis, mindfulness—present-moment, non-judgmental, purposeful attention (Brown & Ryan, 2004)—is most beneficial under conditions of high stress, such that populations who experience chronic stressors may benefit the most from mindfulness training (Creswell & Lindsay, 2014). Among adults, higher dispositional mindfulness is associated with lower physiological reactivity to stressors (Creswell & Lindsay, 2014) and it has provided adults with a buffer against the harmful effects of stressors on mental health (e.g., Bergin & Pakenham, 2016). Although evidence for adolescents is much more limited, cross-sectional evidence suggests that trait mindfulness buffers youth from the psychological effects of life hassles (Marks et al., 2010) as well as from the negative effects of stressors on internalizing symptoms (Lucas-Thompson et al., 2021). Taken together, mindfulness is thought to provide both adults and adolescents with an effective buffer against the negative effects of stressors on mental health.

One process through which greater mindfulness may give rise to improvements in mental health is through improvements in emotion regulation. Emotion regulation has been conceptualized as the process through which individuals effectively manage and express their emotions and it is central to stress responding (Gross, 2014). Individuals who experience difficulties adaptively regulating their emotions often experience more persistent distress in the form of anxiety and depression symptoms (Sheppes et al., 2015). Unfortunately, exposure to high levels of prolonged stress or chronic stressors during childhood and adolescence can result in neural changes that contribute to emotion regulation difficulties (Sheth et al., 2017). Greater emotion regulation difficulties then contribute to the development and exacerbation of mental health problems (Sheppes et al., 2015). In contrast, practicing mindfulness theoretically interrupts these processes by increasing the “top-down” recruitment and activation of prefrontal regulatory regions of the brain that underlie emotion regulation (e.g., see reviews by Creswell & Lindsay, 2014; Guendelman et al., 2017). Mindfulness and practices intended to enhance mindfulness are also thought to target “bottom-up” emotion processes that are more subconscious in nature and closely associated with activity in the amygdala (Guendelman et al., 2017). According to a review of the literature, adults who engage in practices intended to enhance mindfulness (e.g., meditation), also exhibit structural differences within brain regions central to emotion regulation when compared to controls (Wheeler et al., 2017). Higher mindfulness is also associated with higher emotion regulation among adolescents (Ma & Fang, 2019; Pepping et al., 2016). As such, greater mindfulness may give rise to more adaptive emotion regulation, which may ultimately contribute to fewer mental health problems among adolescents exposed to chronic stressors.

However, evidence suggests that, in the absence of mindfulness training, adolescents may not always be capable of remaining mindful and emotionally regulated when experiencing life stressors (R. G. Lucas-Thompson et al., 2021). During adolescence, experiencing everyday life stressors (e.g., peer conflict, academic stress) is relatively common (LaRue & Herrman, 2008; Reddy et al., 2017), but most US adolescents report that their stress levels are higher than what they believe to be healthy (Bethune, 2014). In the wake of COVID-19, many teens also reported significant increases in their stress levels (Goodman et al., 2023). These trends are particularly concerning for those who are exposed to chronic stressors (e.g., ongoing family conflict, poverty) as teens exposed to chronic stressors tend to report greater experiences of everyday life stressors (Booth & Anthony, 2015; Evans et al., 2009; Thoits, 2010) and high levels of perceived stress (Glasscock et al., 2013).

The presence of life stressors can degrade meta-cognitive processes such as executive functioning that support attention allocation and self-regulation of cognition, emotion, and behavior among adolescents, which are associated with mindfulness and emotion regulation (Jankowski & Holas, 2014; Steinberg, 2014). Within situations of high emotional arousal, adolescents’ cognitive abilities and executive functioning also can be negatively influenced by emotionality and situational factors (Steinberg, 2014; Zelazo & Carlson, 2012). As such, experiencing high levels of stress in response to stressful events may degrade adolescents’ mindfulness and emotion regulation. Developmentally, adolescence is characterized by ongoing maturation of cognitive capacities (Jankowski & Holas, 2014; Steinberg, 2014) and thus, without training, mindfulness skills may not be sufficiently developed to effectively buffer adolescents from the impacts of stressors. Recent evidence from an observational daily diary study of adolescents supports this notion; daily stressors were found to degrade adolescent daily mindfulness (when mindfulness was untrained) and contribute to greater daily psychological distress (Lucas-Thompson et al., 2021). Likewise, among adolescents who have not received mindfulness training, momentary life stressors were associated with low levels of mindfulness and high levels of emotion regulation difficulties within the same moment (Miller et al., 2024). The presence of life stressors also predicted less mindfulness and more emotion regulation difficulties within the next moment (Miller et al., 2024). Thus, it may be particularly challenging for adolescents to remain mindful and emotionally regulated during life stressors. Given that training can support individuals in learning to effectively apply new skills under stressful circumstances (Collyer & Malecki, 1998; Salas et al., 2012), mindfulness training may be particularly beneficial for adolescents exposed to chronic stressors.

Mindfulness-Based Interventions (MBIs)

MBIs were designed to provide individuals with training and opportunities to practice and cultivate mindfulness within daily life, especially during times of stress (Cullen, 2011). MBIs have also been shown to produce a wide range of health benefits for adolescents (for a review see Dunning et al., 2019; Hughes et al., 2023; Kostova et al., 2019). Further, a systematic review of the literature suggests that MBIs exert such benefits on health through the enhancement of mindfulness and improvements in one’s regulatory abilities, such as improvements in emotion regulation (Kostova et al., 2019). For adolescents experiencing problem behaviors, delinquency, mental health problems, and poor academic achievement, similar findings have been observed (Rawlett & Scrandis, 2015). A review of the literature on the use of MBIs with adolescents at-risk for poor future outcomes such as poor academic achievement suggests that MBIs can be effective at reducing stress and increasing mindfulness (Rawlett & Scrandis, 2015). Also, within three pilot randomized controlled trials (Bluth et al., 2015; Miller et al., 2021; Sibinga et al., 2013) and one larger randomized controlled trial (Sibinga et al., 2016), an MBI also reduced emotion regulation difficulties and depression symptoms and improved coping skills (Sibinga et al., 2013, 2016) among adolescents at-risk for behavioral, emotional, and/or academic concerns. As such, MBI may beneficially impact adolescents by providing them with the necessary skills and tools to practice mindfulness in the face of stressful life events.

Need for Innovative Assessment Approaches

Most evidence on MBI outcomes among adolescents is based on traditional self-report questionnaires that are completed before and after an intervention (Bai et al., 2020; Galla, 2016). Although data from these measures can characterize average, trait-like levels of socio-emotional and behavioral functioning, they are limited in terms of ecological validity, recall bias, and the inability to characterize momentary associations between dynamic constructs (including mindfulness and emotion regulation; Bai et al., 2020; Kabat-Zinn, 2003). In addition, traditional assessment approaches do not provide data that allow for the examination of process-oriented changes that can occur in response to real-time stimuli (e.g., life stressors; Schwarz, 2007; Shiffman et al., 2008). Understanding responses to life stressors in real-time is important because the nature and degree of changes in emotions, thoughts, and behaviors following exposure to stressors can impact mental health (e.g., Bai et al., 2020; Lucas-Thompson et al., 2021). For example, lower momentary emotion regulation is associated with greater depression symptoms among adults (Vanderlind et al., 2022), and momentary changes in mindfulness have been associated with increases in psychological distress among adolescents (Lucas-Thompson et. al., 2021).

Ecological momentary assessment (EMA) offers an alternative to traditional retrospective self-report measures, and involves intensive, repeated measurements of socio-emotional, cognitive, and behavioral processes in real-time as individuals go about their daily lives. EMA can minimize recall bias, maximize ecological validity, and support examinations of momentary relationships (e.g., responses to stressor exposure; (Bai et al., 2020; Miller et al., 2024; Schwarz, 2007). EMA studies have found that mindfulness training (versus waitlist control) mitigates the depletion of state emotion regulation among college students when experiencing various life stressors (Bai et al., 2020). Also, within a randomized controlled trial, mindfulness training (versus an active control) for emotionally distressed older adults (Moore et al., 2016) was shown to improve momentary reports of depression, anxiety, and mindfulness. Although there has been a call to use EMA within investigations of adolescent MBI (Goodman et al., 2017), such research has thus far been limited. Therefore, using data obtained via EMA to characterize theoretically important mechanisms of MBIs among adolescents is an important next step for research.

The Current Study

Using EMA data collected as part of a randomized controlled trial for adolescents exposed to chronic stressors, the aim of the present study was to understand how mindfulness training may attenuate the associations of life stressor exposure with lower mindfulness and greater emotion regulation difficulties at the momentary-state level. Adolescent participants were randomized to receive mindfulness training (MBI+mentoring) or an active control (mentoring-alone). In line with existing literature (Bai et al., 2020; Bluth et al., 2015; Cotton et al., 2016; Miller et al., 2021), we hypothesized that the strength of the relationships between 1) life stressor exposure and lower mindfulness and 2) life stressor exposure and greater emotion regulation difficulties would be significantly mitigated over time for those who received MBI+mentoring compared to those who received mentoring-alone. We also expected that there would be smaller associations between 1) life stressor exposure and lower mindfulness and 2) life stressor exposure and greater emotion regulation difficulties measured at mid-intervention and at post-intervention, compared to pre-intervention, as well as at post-intervention when compared to mid-intervention, among adolescents who received MBI+mentoring. In contrast, for adolescents who received mentoring-alone, we expected a consistent degree of association of life stressors with lower mindfulness and greater emotion regulation difficulties across the pre-intervention, mid-intervention, and post-intervention periods. The overarching aim of this study was to provide knowledge about how mindfulness training may alter the impacts that life stressors have on momentary-state mindfulness and emotion regulation difficulties among adolescents exposed to chronic stressors.

Method

All procedures of this study were approved by the Colorado State University Institutional Review Board. Adolescents and their parents/guardians provided assent and informed consent, respectively, at their intake meeting with the mentoring program after all risks and study procedures were explained by trained staff. Adolescents also selected the night of the mentoring program that worked best with their schedule during the intake meeting and were not informed about which night was randomized to which treatment condition. Recruitment and enrollment took place in July 2021 through February 2022. Data collection was completed in August 2022.

Transparency and Openness

The parent study aims/hypotheses were preregistered (NCT04927286), and the aims/hypotheses and analytic framework for the EMA sub-study were established a priori (F31AT011642), but not formally pre-specified in the trial registration. Of note, the pre-specified model did not converge, which prompted the need for an alternative approach that would allow for us to achieve our pre-specified aims without model convergence issues; therefore, we revised our analytic plan to utilize generalized linear mixed models. The updated data analysis plan is presented below. Within this paper, we also report on how determinations were made regarding sample size, data exclusions, manipulations, and measures used, and follow the APA Journal Article Reporting Standards. Data and analysis code are freely available (https://github.com/ahead-lab-research/EMA-MBI). Analyses were conducted in R Studio (R Core Team, 2021) using the “glmmTMB,” “lme4,” and “EMMeans” packages.

Participants

Participants were 81 adolescents (10-18 years old; M_age=13.75 years, SD=2.17 years) who were referred to a community-based mentoring program for exhibiting indicators of risk (e.g., Department of Human Services [DHS]/juvenile-justice involvement, behavioral/emotional problems) and most, but not all, adolescents were facing chronic stressors. Adolescents were also participating in a larger randomized controlled trial (R. Lucas-Thompson et al., 2024). Inclusion criteria were enrollment in the mentoring program, being between the ages of 10-18 years old, and being English-speaking because the mentoring program is only delivered in English (Figure 1). Slightly fewer girls/females (37%, n=30) than boys/males participated in the study; 7% identified as another gender (n=6). Adolescents identified as non-Hispanic White (n=46, 57%), Hispanic/Latino (n=19; 24%), Native American (n=6; 7%), Asian/Pacific Islander or Black/African American (n=4; 5%), or more than race (n=6; 7%). Although adolescents resided within a Northwestern state of the United States, the racial and ethnic composition was more diverse than the population of adolescents within this region (United States Census Bureau, 2022). More specifically, the typical racial/ethnic make-up of this region is 78% Non-Hispanic White, 12% Hispanic/Latino, 4% Asian/Pacific Islander and 3% Black/African American and the median household income is approximately $79,000 (United States Census Bureau, 2022). Approximately 23% of adolescents also reported going up in a rural community, but the mentoring program was delivered within an urban area. Four adolescents (5%) withdrew from the study before receiving the intervention and one withdrew at post-intervention. There were no significant differences in age, income, race, or gender identity between those who completed the study (n=76) and those who withdrew (n=5; p-values>.16).

Figure 1. CONSORT Flow Diagram.

This sample demonstrated significant socio-emotional and behavioral vulnerabilities, indicative of exposure to chronic stressors (Romas & Sharma, 2022). Of those parents/guardians (n=76) who provided information on a baseline risk assessment (Herrera et al., 2013) measuring environmental (e.g., economic adversity) and individual (e.g., problem behaviors) risk factors, which are strongly associated with chronic stressors (Romas & Sharma, 2022), the majority (n=73; 96%) reported at least one indicator of risk. Five parents/caregivers declined baseline survey completion, which included the baseline risk assessment, but they did consent to their children's participation. On average, parents reported experiencing four out of ten (SD=2.42; Range=0-10) socio-emotional/behavioral risk factors. More specially, many indicated that within the past 12 months, their family had experienced difficulties paying bills (n=41; 54%). Similarly, 50% of families (n=37) reported making less than $20,000 a year. The median household income ranged from $40,000 to $59,999 and 49% (n=52) of parents/guardians reported that they did not receive a college degree. A majority (n=44; 56%) of parents/guardians also reported that their child had been diagnosed with a mental health diagnosis such as anxiety or depression. In addition, 55% of parents/guardians placed themselves on the bottom five rungs of the ladder on the MacArthur Scale of Subjective Social Status (Adler et al., 2000), which indicated that they believed that they were worse off compared to other people in the United States. Only 16% of parents/guardians (n=11) placed themselves on the top three rungs, eight to ten.

Procedures

The mentoring program, Campus Connections, occurred for 12 weeks between February 2021 to May 2021 and again from September 2021 to December 2021. During each 12-week period, program activities were offered three nights per week, with different adolescents served on each of the program. One night of the week was randomized to receive the MBI embedded within the mentoring program (n=38; MBI+mentoring) and the other night(s) received mentoring-alone (n=43). Simple randomization by night took place using a freely available list randomizer (i.e., random.org). The list randomizer generates randomness via atmospheric noise and produces a list in random order. Notably, data suggest that there were no systematic differences in age, race/ethnicity, gender, income or risk between different nights of the program p-values>.08. Participants randomized to MBI+mentoring received nine, 30-minute sessions of Learning 2 Breathe (L2B; Broderick, 2013) during weeks two through ten. Pre-intervention data collection occurred on mentoring program week one, mid-intervention data collection occurred on mentoring program week five and post-intervention data collection occurred on mentoring program week 11. Graduation from the mentoring program, an evening session when caregivers are invited to celebrate the accomplishments of the youth in the program, occurred on week 12. L2B is an evidence-based adolescent, group-based MBI intended to enhance emotion regulation with interactive and experiential activities, discussion, and guided mindfulness practices focused on body/thought/emotional awareness, attention to the body, thoughts and emotions, and tenderness or non-judgment. All adolescents in both intervention arms completed mentor–mentee activities to build positive relationships, support academic success, and explore prosocial interests (for a full description, see Weiler et al., 2013). Those in mentoring-alone received an extra 30 minutes of these activities as opposed to the 30-minute L2B program. One cohort of adolescents (n=44) participated in activities online while the other cohort (n=37) participated face-to-face. This was an unplanned adjustment due to the COVID-19 global pandemic that forced many interventions online. The average number of attended sessions was 7.61 (85%) out of a total of 9 sessions, which suggests that there were adequate rates of attendance. Attendance did not vary by cohort (p=.73) nor by arm (p=.64). Participants and research staff were not masked to condition arm.

EMA Protocol.

Before program participation, participants who agreed to take part in the EMA protocol received training in how to answer EMA questions. Participants then completed one week of baseline EMA, which included three signals per day. Depending upon how early the participant started and ended school, weekday signals were delivered at semi-random times between 3:00pm-9:00pm or between 7:00am-8:15am and between 4:15pm-9:00pm. On the weekends, signals were delivered at semi-random times between 9:00am-9:00pm. Once the participant received a notification to complete a survey, they had 30 minutes to respond before the survey expired. Adolescents received $1 per survey and a $5 bonus if they answered at least 76% of all possible surveys (i.e., 16 out of 21). This same EMA procedure was utilized again at mid-intervention and post-intervention. Each EMA survey was comprised of the same 13 questions and was estimated to take five minutes or less to complete. EMA procedures were aligned with prior EMA research for youth (Heron et al., 2017; Wen et al., 2017).

Measures

Life stressor exposure.

In line with Hankin and colleagues (2005), participants responded to the question, “In the last hour, has at least one negative event occurred?” Given that life stressors may be indicative of chronic stressors and/or daily life hassles, we refer to this construct as life stressors throughout the manuscript. Examples of negative events (e.g., getting a bad grade on a test) were provided during the baseline visit. This variable was measured dichotomously as presence (1) versus absence (0).

Mindfulness (vs. Mindlessness).

Two key dimensions of mindfulness were assessed: attention/awareness and non-judgment. To assess for mindful attention/awareness, participants completed the five-item Mindful Attention and Awareness Scale (Brown & Ryan, 2003). Participants rated the extent to which they were currently experiencing a statement reflective of low levels of mindfulness, which we will refer to as mindlessness (e.g., “I am preoccupied with the past or future”; Brown & Ryan, 2003) on a 7-point Likert-type scale from 1 (not at all) to 7 (very much). An average score was calculated (Ω_b=.94; Ω_w=.81). To assess for mindful non-judgment, participants completed one item from the Self-Compassion Scale for Children-Short Form (Sutton et al., 2018). Participants rated how much the statement, “I feel disapproving and judgmental of the things I don’t like about myself” currently applied to them on a 7-point Likert-type scale from 1 (not at all) to 7 (very much). Lower scores indicate greater mindful attention and mindful non-judgment; higher scores indicate what we will refer to as mindless attention and mindless non-judgment.

Emotion regulation difficulties.

Participants completed four items from the State-Difficulties in Emotion Regulation Scale (S-DERS; Lavender et al., 2017), a state-oriented measure of emotion regulation difficulties based on the original trait-oriented DERS (Gratz & Roemer, 2004). Items with the highest factor loadings on each subscale (i.e., non-acceptance, awareness, modulate, and clarity) from the S-DERS were selected (Lavender et al., 2017). However, the awareness item had a low factor loading (.06) and, therefore, was not included in analysis. Participants responded to how much a statement (e.g., “I feel embarrassed for feeling how I feel”) currently applied to their emotions on a Likert-type scale of 1 (not at all) to 7 (completely). An average score was calculated. Higher scores indicate greater difficulties with emotion regulation (Ω_b=.94; Ω_w=.70).

Data Analysis

Prior to analyses, variables were checked for non-normality and all variables were found to be significantly and positively skewed. Generalized linear mixed models (GLMM) with a gamma distribution and log link function were used to account for non-normality (Bolker et al., 2009). The gamma distribution can most appropriately handle positively (versus negatively) skewed distributions (Magnusson et al., 2017); therefore, mindfulness items were not reverse scored and mindlessness language is used within the results to aid with interpretability of estimates (i.e., higher scores indicate greater mindlessness and lower scores indicate greater mindfulness; Brown & Ryan, 2003). Interaction terms were also created between intervention arm (i.e., MBI+mentoring vs. mentoring-alone) and assessment interval (dummy coded to make comparisons between baseline, mid-intervention, and post-intervention), which will be referred to as ‘burst’ throughout because this was a measurement burst design study (Cho et al., 2019).

First, intra-class correlations (ICCs) were determined using intercept-only models to investigate the nested structure of the data. ICCs describe the proportion of variance in a variable that can be explained by the grouping variable, in this case, day, burst, and participant were all considered as grouping variables. The ICCs were all above .05 and the assumption of independence was likely violated (Kreft & De Leeuw, 1998). As such, a three-level GLMM model was used to test study hypotheses with random intercepts and slopes. Moment was specified as level 1, day as level 2, and participant as level 3. Burst was not included as a level because data were separated by burst to calculate the random effect of life stressors predicting mindlessness and emotion regulation difficulties at the concurrent, momentary-level. Next, the random effects from each burst were extracted from these models and entered into a two-level model as the dependent variable. More specifically, we tested if the interaction term of arm X burst predicted the random slope of life stressors predicting mindlessness and emotion regulation difficulties using a two-level linear mixed effect model with moment as level one and participant as level two (Supplemental Figure 1). Models with mindlessness and emotion regulation were conducted separately and run twice to make all pairwise comparisons between baseline, mid-intervention, and post-intervention. Maximum likelihood estimation was used to within all analyses; model equations can be found in the supplemental materials. Based on the power tables outlined by Arend and Schafer (2019) for multilevel models and the number of between- and within-person units in the current study, we had adequate power to detect small, medium, and large effects (see Table 7 in Arend and Shafer, 2019 for more information). To probe significant interaction terms, estimated marginal means (EMMs) for each model were plotted and pairwise comparisons between all contrasts were calculated using t-tests (Kassambara, 2019). Notably, given the log link function, all estimates in the tables and figures are on the log scale, but for added interpretability, the estimated marginal mean have been converted to percentages in the results section by converting the log coefficient to its natural form, subtracting by one and then multiplying by 100 to get the percent change. Effect sizes for fixed effects were calculated following Aiken and West (1991) recommendations and were considered small at 0.02, medium at 0.15, and large at 0.35 (Cohen, 1988).

Results

Eighty-one participants provided a total of 3,178 total data points across the three bursts. The average number of surveys (out of 21) completed was 14.64 at baseline, 14.96 at mid-intervention, and 14.28 at post-intervention, which indicates that adolescents completed approximately 68%-71% of surveys at each night of the three seven-day bursts. Number of completed EMA surveys across the bursts was not significantly associated with age, race/ethnicity, income, or gender identity (p-values>.23). The overall percent of missing data across key study variables was 0.43%. More specifically, the maximum number of items we could have collected from participants was 19,068 based on the 3,178 completed EMA survey multiplied by the number of variables in each survey. Of these possible 19,068 possible data points, there were 82 missing data points. Mindful attention had 1.5% missing data (n=48), mindful non-judgment had .81% missing data (n=26), emotion regulation had .19% missing (n=6), and stressful life events had .06% missingness (n=2). Given that the missing data are well below 5%, missingness can be considered ignorable in study analyses (Jakobsen et al., 2017).

The main effects or the relationships between life stressors, mindfulness, emotion regulation, arm and burst (without interaction terms) are in Supplemental Tables 1-4. Main effects were all in expected directions. In addition, there were no significant differences in mindfulness, emotion regulation, or life stressors between the two study arms at baseline (p-values>.08).

Life Stressor Exposure and Mindlessness

Mindlessness (Attention).

The random slope of life stressor exposure (absence vs presence) predicting the attention dimension of mindlessness (i.e., mindless attention) was significantly and negatively associated with the interaction term of arm X burst when baseline was compared to post-intervention (Table 1). This inverse association was also significant when mid-intervention was compared to post-intervention. The association between the random slope and the interaction term was not significant when comparing baseline to mid-intervention.

Table 1.

Linear Mixed Effects of Arm X Burst predicting Random Slopes of Life Stressors on Mindfulness and Emotion Regulation Difficulties

Baseline as Reference Group
	Mindless Attention			Mindless Non-Judgment			Emotion Regulation Difficulties
	Estimate	SE	p-value	Estimate	SE	p-value	Estimate	SE	p-value
Fixed effects
Intercept	−0.02	0.02	0.30	0.00	0.01	0.63	0.00	0.02	0.95
Arm	0.03	0.02	0.20	0.00	0.01	0.99	0.00	0.02	0.98
Mid-intervention	0.02	0.00	<.001	0.00	0.00	0.58	−0.01	0.01	0.20
Post-intervention	0.04	0.00	<.001	−0.01	0.00	0.19	0.02	0.01	<.001
Arm X Mid-intervention	−0.02	0.01	0.02	−0.01	0.01	0.18	0.01	0.01	0.56
Arm X Post-intervention	−0.05	0.01	<.001	−0.03	0.01	<.001	−0.04	0.01	<.001
Random effects (variances)
Intercept of subject	0.01	--	--	0.05	--	--	0.10	--	--
Residual	0.07	--	--	0.07	--	--	0.10	--	--
Effect size (Cohen’s f2)	.006			.005			.005
Mid-Intervention as Reference Group
	Mindless Attention			Mindless Non-Judgment			Emotion Regulation Difficulties
	Estimate	SE	p-value	Estimate	SE	p-value	Estimate	SE	p-value
Fixed effects
Intercept	0.00	0.02	0.98	0.01	0.01	0.45	−0.01	0.02	0.70
Arm	0.02	0.02	0.51	−0.01	0.01	0.48	−0.01	0.02	0.82
Baseline	−0.02	0.00	<.001	0.00	0.00	0.58	0.01	0.01	0.20
Post-intervention	0.02	0.00	<.001	−0.01	0.00	0.07	0.03	0.01	<.001
Arm X Baseline	0.02	0.01	0.02	0.01	0.01	0.18	0.00	0.01	0.56
Arm X Post-intervention	−0.04	0.01	<.001	−0.02	0.01	0.01	−0.04	0.01	<.001
Random effects (variances)
Intercept of subject	0.01	--	--	0.05	--	--	0.10	--	--
Residual	0.01	--	--	0.07	--	--	0.10	--	--
Effect size (Cohen's f2)	.006			.005			.005

Notes. All estimates are on the log scale to allow for ease of interpretation with regards to the direction of each effect

To probe significant interaction terms, EMMs revealed that life stressor exposure was more strongly associated with greater mindless attention over time for those in mentoring-alone (Table 2). For those in MBI+mentoring, associations between life stressor exposure and mindless attention were positive, but significantly smaller over time, which suggests that those who received MBI+mentoring may have been more protected from the negative effects of life stressors on mindless attention at post-intervention. More specifically, at baseline, the presence of at least one life stressor, compared to the absence of life stressors, was associated with 1.41% greater mindless attention for those in MBI+mentoring and a 1.78% less mindless attention for those in mentoring-alone (Figure 2). At mid-intervention, the presence (vs. absence) of a life stressor was associated with 1.61% greater mindless attention for those in MBI+mentoring and 0.10% greater mindless attention for those in mentoring-alone. At post-intervention, the presence (vs. absence) of life stressors was associated with 0.06% greater mindless attention for those in MBI+mentoring and a 2.12% greater mindless attention for those in mentoring-alone. Pairwise comparisons revealed that EMMs for those in MBI+mentoring were not significantly different from EMMs for those in mentoring-alone at baseline, mid-intervention, or post-intervention (p-values>.75). Instead, EMMs were significantly different within intervention arms over time.

Table 2.

Pairwise Comparisons of Estimated Marginal Means

	Random Slope for Mindless Attention			Random Slope for Mindless Non-Judgment			Random Slope for Emotion Regulation Difficulties
Comparison	Estimate	SE	p-value	Estimate	SE	p-value	Estimate	SE	p-value
Baseline Mentoring-Alone vs. Baseline MBI+mentoring	−0.03	0.03	0.79	0.00	0.01	1.00	0.00	0.02	1.00
Baseline Mentoring-Alone vs. Mid-Intervention Mentoring-Alone	−0.02	0.00	<.001	0.00	0.00	0.99	0.01	0.01	0.79
Baseline Mentoring-Alone vs. Mid-Intervention MBI+mentoring	−0.03	0.03	0.76	0.01	0.01	1.00	0.01	0.02	0.99
Baseline Mentoring-Alone vs. Post-Intervention Mentoring-Alone	−0.04	0.00	<.001	0.01	0.00	0.78	−0.02	0.01	0.01
Baseline Mentoring-Alone vs. Post-Intervention MBI+mentoring	−0.02	0.03	0.98	0.03	0.01	0.08	0.02	0.02	0.94
Baseline MBI+mentoring vs. Mid-Intervention Mentoring-Alone	0.01	0.03	0.99	0.00	0.01	1.00	0.01	0.02	1.00
Baseline MBI+mentoring vs. Mid-Intervention MBI+mentoring	0.00	0.00	1.00	0.01	0.00	0.79	0.01	0.01	0.37
Baseline MBI+mentoring vs. Post-Intervention Mentoring-Alone	−0.01	0.03	1.00	0.01	0.01	1.00	−0.02	0.02	0.96
Baseline MBI+mentoring vs. Post-Intervention MBI+mentoring	0.01	0.00	0.06	0.03	0.00	<.001	0.02	0.01	0.01
Mid-Intervention Mentoring-Alone vs. Mid-Intervention MBI+mentoring	−0.02	0.03	0.99	0.01	0.01	0.98	0.01	0.02	1.00
Mid-Intervention Mentoring vs. Post-Intervention Mentoring-Alone	−0.02	0.00	<.001	0.01	0.00	0.44	−0.03	0.01	<.001
Mid-Intervention Mentoring-Alone vs. Post-Intervention MBI+mentoring	0.00	0.03	1.00	0.04	0.01	0.05	0.01	0.02	0.99
Mid-Intervention MBI+mentoring vs. Post-Intervention Mentoring-Alone	−0.01	0.03	1.00	0.00	0.01	1.00	−0.03	0.02	0.76
Mid-Intervention MBI+mentoring vs. Post-Intervention MBI+mentoring	0.02	0.00	0.02	0.03	0.00	<.001	0.01	0.01	0.73
Post-Intervention Mentoring-Alone vs. Post-Intervention MBI+mentoring	0.02	0.03	0.96	0.03	0.01	0.25	0.04	0.02	0.52

Note. All estimates are on the log scale to allow for ease of interpretation with regards to the direction of each effect. Pairwise comparisons were calculated with t-tests. Estimated marginal means represent the random slope of life stressors predicting mindfulness and emotion regulation.

Figure 2. Estimated Marginal Means (EMMs) of Random Slopes.

Note. All estimates in the tables are in the log scale to allow for ease of interpretation with regards to the direction of each effect. Text in the results section has been converted to percentages to further aid with interpretation. Arm0 represents Mentoring-alone; Arm1 represents MBI+mentoring; burst1 represents baseline; burst2 represents mid-intervention; burst3 represents post-intervention. Figure 2a represents the estimated marginal means (EMMs) of life stressors predicting mindless attention across burst and arm; 2b represents the EMMs of life stressors predicting mindless non-judgment across burst and arm; 2c represents the EMMs of life stressors predicting DERS (difficulties with emotion regulation) across burst and arm.

Mindlessness (Non-judgment).

The random slope of the presence vs. absence of life stressors predicting the non-judgment dimension of mindlessness (i.e., mindless non-judgment) was negatively associated with the interaction term of arm X burst when baseline was compared to post-intervention. All other associations were not significant.

EMMs revealed that at baseline, the association of life stressors with mindless non-judgment was similar for those in MBI+mentoring and for those in mentoring-alone. At baseline, the presence (vs. absence) of at least one life stressor was associated with 0.4% greater mindless non-judgment for those in MBI+mentoring and mentoring-alone. At post-intervention, however, the presence (vs. absence) of at least one life stressor was associated with 2.65% less mindless non-judgment for those in MBI+mentoring and 0.20% less mentoring-alone.

Pairwise comparisons revealed that EMMs were significantly lower at post-intervention when compared to both baseline and mid-intervention for those in MBI+mentoring. All other comparisons were non-significant. These patterns suggest that most significant change in the random effect of life stressor exposure on mindless non-judgment occurred for those in MBI+mentoring at post-intervention.

Life Stressor Exposure and Emotion Regulation Difficulties

The random slope of the association between life stressors and emotion regulation difficulties was negatively associated with the interaction term of arm X burst when baseline was compared to post-intervention and when mid-intervention was compared to post-intervention. These associations were not significant when baseline was compared to mid-intervention.

To probe significant interaction terms, EMMs revealed that at baseline, the effect of life stressor exposure on difficulties with emotion regulation was similar for those in MBI+mentoring and for those in mentoring-alone. The presence of at least one life stressor was associated with 0.04% greater emotion regulation difficulties for those in MBI+mentoring and 0.10% greater emotion regulation difficulties for those in mentoring-alone at baseline when compared to moments without any reported life stressors. At mid-intervention, both arms experienced a reduction in the association of life stressor exposure with emotion regulation difficulties. The presence (vs. absence) of at least one life stressor was associated with 1.20% lower emotion regulation difficulties for those in MBI+mentoring and 0.60% lower emotion regulation difficulties for those in mentoring-alone at mid-intervention. At post-intervention, these patterns diverged as the presence (vs. absence) of life stressors was associated with 2.08% lower difficulties with emotion regulation for those in MBI+mentoring and 2.12% greater difficulties with emotion regulation for those in mentoring-alone.

Pairwise comparisons revealed that for those in mentoring-alone, life stressor exposure was more strongly associated with greater emotion regulation difficulties at post-intervention compared to both baseline and mid-intervention. For those in MBI+mentoring, life stressor exposure was more strongly associated with lower emotion regulation difficulties at post-intervention compared to baseline (but not mid-intervention). These findings suggest that an exacerbated relationship between life stressor exposure and emotion regulation difficulties over time for those in mentoring-alone, whereas for those in MBI+mentoring, the link between life stressor exposure and emotion regulation difficulties ameliorated over time.

Discussion

The goal of the current study was to understand how mindfulness training may alter the momentary relationships of life stressor exposure with mindfulness and emotion regulation difficulties within the context of a randomized controlled trial for adolescents exposed to chronic stressors. Existing evidence supports the notion that, in the absence of formal mindfulness training, adolescents may display difficulties in remaining mindful and emotionally regulated following stressful experiences (Lucas-Thompson et al., 2021; Miller et al., 2024). Results from this study suggest that compared to an active control, mindfulness training delivered within a community-based mentoring program does appear to ameliorate the adverse effects of life stressor exposure on mindfulness and difficulties with emotion regulation, particularly after adolescents receive the full intervention (i.e., at post-intervention). These findings are meaningful as they demonstrate the possible positive and protective effects of mindfulness training for youth facing high levels of chronic stressors and social adversity.

As supported by previous studies (Bai et al., 2020; Lucas-Thompson et al., 2021; Moore et al., 2016) and in line with current hypotheses, there were significant changes in the association of life stressor exposure with on mindfulness (vs. mindlessness) and emotion regulation difficulties from baseline to post-intervention for those who received mindfulness training within a mentoring program compared to adolescents who received mentoring-alone. For those who received distinctive mindfulness training, the association between life stressor exposure and lower mindful attention (or greater mindless attention) was significantly smaller at post-intervention compared to baseline. In addition, for those who received mindfulness training, the presence of at least one life stressor was associated with greater mindful non-judgment (or lower mindless non-judgment) and lower emotion regulation difficulties when post-intervention was compared to baseline. In other words, adolescents who received mindfulness training were more capable of exercising mindful non-judgment and emotion regulation capacities in the face of momentary stressors at post-intervention. For those who received mentoring-alone, the opposite effect was observed: life stressor exposure was more strongly associated with lower mindful attention and greater emotion regulation difficulties at post-intervention when compared to baseline. From mid-to-post-intervention, there were also noted increases in emotion regulation difficulties during life stressors within the mentoring-alone control group, which was not previously hypothesized. These findings may be explained in part by the stress sensitization model (Stroud et al., 2020), which posits that individuals exposed to multiple adversities and chronic stressors can become increasingly sensitive to the effects of stress over time. Given that most adolescents within the current sample had been exposed to chronic stressors and the fact that the control group did not receive mindfulness training, adolescents within the mentoring-alone condition may have experienced a more pronounced effect of stress on emotion regulation over time. However, replication of these results with additional samples of adolescents may be warranted to provide further evidence for this conclusion. Taken together, these results tentatively suggest that mindfulness training may help to protect adolescents from the adverse momentary effects of life stressors within their daily lives.

Notably, many study effects were relatively small. Although small effects may translate to limited clinical significance (Kazdin, 1999), it is important to acknowledge that these are momentary-level data and it is possible that small momentary effects of life stressors on mindfulness and emotion regulation may accumulate over time and explain a considerable proportion of the variance in mental health problems. This potential explanation is partially supported by research suggesting that cumulative stress, or stressful events experienced over time and across multiple domains, can accumulate and lead to greater mental and physical health concerns (Bøe et al., 2018; Haight et al., 2023). In addition, momentary-level reports of emotion regulation have been more closely associated with anxiety and depression diagnoses than even daily-level reports (Shin et al., 2022), which further highlights the potential for momentary-level processes to give rise to mental health challenges. Similarly, the small reductions in the associations of life stressors with mindlessness and emotion regulation difficulties experienced by adolescents who received mindfulness training may compound and give rise to improvements in mental health. Alternatively, it is also possible that these changes do not accumulate and remain small in nature. To understand the nature of small effects more definitively within momentary-level data, it may be helpful for future research to investigate if cumulative momentary effects of life stressors on mindfulness and emotion regulation help to explain the development and maintenance of mental health symptoms. In addition, it will be important to explore if small momentary changes experienced after an MBI accumulate and produce sustained positive changes in overall mental and physical health.

Contrary to hypotheses, significant between-intervention arm differences in associations of life stressors with mindfulness and emotion regulation difficulties were not observed from baseline to mid-intervention (measured after L2B session four/ week five of the mentoring program). Instead, changes were observed at post-intervention. As the cognitive capacities that support mindfulness and emotion regulation are still maturing in adolescence (Dahl, 2004; Steinberg, 2014), adolescents may require the entire mindfulness training dosage, which totaled ~4.5 hours spread over several months, to bolster functional mindfulness (i.e., mindfulness applied within everyday life) and emotion regulation capabilities. At the mid-intervention assessment, adolescents had only received ~2 hours of mindfulness training in the context of the mentoring program. This explanation is in line with research that found a potential dose-response relationship of a 6-week adolescent MBI and mindfulness (R. G. Lucas-Thompson et al., 2023). Twelve adolescents who received an MBI experienced stronger mindfulness stress buffering effects at the daily level during the second half of the MBI as opposed to at the mid-way point (Lucas-Thompson et al., 2023). Another aspect that merits consideration is that, given the scaffolded nature of the L2B program (Broderick, 2013), adolescents did not receive explicit training in practicing mindfulness and effective emotion regulation during moments of stress until L2B session six (i.e., week seven of the mentoring program). Positive effects of the MBI on mindfulness and emotion regulation may not have been observed at mid-intervention because adolescents had not yet received this explicit training. However, it remains unclear just how many and which components of MBI training is needed to produce momentary-level changes in mindfulness and emotion regulation. As such, explicit tests of MBI dosage effects on momentary mindfulness and emotion regulation are warranted to clarify the dose-response relationships between MBIs and these momentary-level processes.

Taken together, these findings have applied and theoretical significance. Results demonstrate that mindfulness training, delivered in the context of a community-based mentoring program for adolescents facing chronic stressors, reduces the momentary deleterious effects of life stressor exposure on mindfulness and emotion regulation. These results are in line with existing literature suggesting that adolescent MBIs increases mindfulness and emotion regulation capabilities (Miller et al., 2021; Rawlett & Scrandis, 2015), and they also extend this literature in several important ways. First, these findings highlight how skills and concepts taught during an MBI can be functionally applied within stressful circumstances among adolescents exposed to a high degree of chronic stressors. Second, these results demonstrate that the benefits of mindfulness training can be observed in real-world contexts (i.e., outside of a research laboratory). Third, findings from this study provide preliminary evidence that an MBI can impact theorized mechanisms of change, including state mindfulness and emotion regulation (Chiesa & Serretti, 2009; Lindsay & Creswell, 2017), at the momentary level. This result is important because most data on mechanisms of change for adolescent MBIs are derived from studies using between-subjects and cross-sectional designs that are subject to limited ecological validity (Galla, 2016). As such, results from this study contribute important knowledge on the active components that underlie mindfulness training and suggest that such desired changes in mindfulness/emotion regulation could be explanatory in benefiting later health outcomes following MBI. Going forward, it will be helpful to test to what extent improvements in momentary mindfulness and emotion regulation give rise to changes in mental and physical health outcomes.

Limitations

There are several important limitations to consider when interpreting these results. First, the EMA protocol in this study was designed to obtain data at three bursts (e.g., baseline, mid-intervention, and post-intervention) throughout the study. Although this design is commonly used within developmental research to reduce participant burden (Cho et al., 2019), there may be important processes to capture between bursts that were not captured within this study. Second, the study included a relatively small sample of adolescents. Although we had adequate power to detect small, medium and large effects given the intensive, repeated measurement design (Arend & Schäfer, 2019), it will be important to determine if results can be replicated in a larger sample to further evaluate the benefits of mindfulness training for adolescents within moments of stress. Third, this study did not include any measurement of strengths or resilience. It is essential to note that even when individuals experience a high degree of chronic stressors, they may also experience resilience or adaptive functioning after experiencing adversity; resilience may ultimately protect them from experiencing mental health problems (Malhi et al., 2019). In order to more comprehensively characterize the impact of chronic stressors and explore the benefits of mindfulness training on resilience, it will be important to include measurements of strengths and resilience within future research. Fourth, our rates of missing data were relatively low, but strategies for handling missingness in multilevel, repeated measurements within generalized mixed-effects models are still developing (Enders, 2023). As additional strategies for handling missingness in multilevel, repeated measurements become available, the hypotheses may benefit from additional testing. Moreover, this study did not include longer periods of follow-up. Although research suggests that there are long-term benefits of MBIs for mental health (e.g., Solhaug et al., 2019), and in some cases, even stronger effects at longer-term follow-up (Bögels et al., 2021), it is currently unknown how long the positive effects observed within this study may last. Future investigations would benefit from including longer periods of follow-up with a larger sample of adolescents.

Conclusions

This study contributes meaningful information suggesting the positive and potential buffering effects of mindfulness training for adolescents’ ability to be mindful and adaptively regulate their emotions within moments of stress. Results suggest that mindfulness training targets key proposed mechanisms of change, mindfulness and emotion regulation, by ameliorating problematic links between life stressor exposure, lower mindfulness and greater emotion regulation difficulties. Mindfulness training may instead support adolescents in practicing mindful non-judgment and regulating emotions during momentary stressors. Going forward, it may be helpful to explore if small momentary-level changes experienced after a mindfulness intervention can accumulate and produce more positive changes in overall mental and physical health. It will also be helpful to investigate dosage effects of adolescent MBIs to understand exactly when MBIs may benefit momentary levels of mindfulness and emotion regulation. Collectively, the findings of this study provide meaningful evidence for how and why MBIs can be beneficial for health and well-being among adolescents who may be particularly vulnerable to the impacts of life stressors.

Public Health Significance Statement:

Adolescents exposed to chronic stressors are at high risk for developing mental health problems, in part, due to impairments in emotion regulation that can happen after experiencing stressors. Results from this study revealed that mindfulness training may help to protect adolescents from the negative impacts of everyday life stressors on momentary mindfulness and emotion regulation difficulties in their daily lives.