The study of mindfulness and well-being has aroused an increasing scholarly interest over the last two decades (Garland et al. 2017; Keng et al. 2011). Mindfulness is a mental state of awareness, a psychological process, and a set of practices that drive individuals to directly contact with the present moment (Chambers et al. 2015; Coffey et al. 2010; Hofmann et al. 2010). Mindfulness emphasizes on paying attention to the present moment, including inner experiences such as thoughts and emotions, through a non-judgmental attitude (Hill and Updegraff 2012; Kabat-Zinn 1994). It can be defined either as a disposition or a set of meditative/therapeutic technique that requires practitioners to be aware of the present moment and to observe and accept thoughts and emotions nonjudgmentally (Kabat-Zinn 1990). Mindfulness is associated with a number of adjustment outcomes, including better well-being, better focus and attention, and fewer symptoms of depression anxiety (e.g., Baer et al. 2012; Bowlin and Baer 2012; Bränström et al. 2011; Chambers et al. 2009; Chambers et al. 2008; Hollis-Walker and Colosimo 2011; Lyvers et al. 2014; Roemer et al. 2009).

The precise mechanisms underlying the effect of mindfulness have received recent theoretical attention (e.g., Garland et al. 2015; Lindsay and Creswell 2017; Teper et al. 2013). Importantly, scholars highlighted emotion regulation as one of the core correlates of dispositional mindfulness and mental health (Chambers et al. 2009; Desrosiers et al. 2013). Emotion regulation refers to the process of modulating one or more aspects of emotional experiences and expressions in response to environmental demands (Gross 1998; Cole et al. 2004). Teper et al. (2013) proposed a conceptual model of mindfulness and emotion regulation, in that mindfulness fosters emotion regulation through the underlying pathways of cognitive and emotional processing. According to the model, the cultivation of mindfulness is associated with an open acceptance of negative experiences, as opposed to critical judgment, denial, or thought and emotion suppression. Importantly, dispositional mindfulness and experiences of meditation foster an open acceptance of errors and their associated affective response (Roemer and Orsillo 2002). Such an open acceptance facilitates non-biased awareness and observation of the experience, which makes room for cue detection (Lindsay and Creswell 2017; Teper et al. 2013). At the same time, mindfulness is directly associated with non-biased awareness, which encapsulates continuous attention of the present moment and subtle emotional changes (Cardaciotto et al. 2008). The non-biased awareness may, in turn, facilitate acceptance through individuals’ attendance to and detection of affective cues (Teper et al. 2013; see also Lindsay and Creswell 2017, for a discussion of attention monitoring and acceptance as mechanisms underlying dispositional mindfulness and mindfulness training). Returning to the conceptual model proposed by Teper et al. (2013), open acceptance and non-biased awareness are integral to executive control, i.e., the ability to shift between tasks or attention, to monitor and update information, to inhibit dominant responses, and to be cognitively flexible (Miyake et al. 2000). Executive control may, in turn, facilitate emotion regulation and mental health by virtue of an increased ability to exercise self-control (Baumeister et al. 2007).

Numerous studies have lent support to the relations between mindfulness, emotional awareness, and emotional acceptance in community samples (e.g., Jha et al. 2007; Lattimore et al. 2011; Moore and Malinowski 2009; Peters et al. 2011; Thompson and Waltz 2008). Notably, individuals who received an 8-week mindfulness-based stress reduction course performed significantly better in a cognitive switching task than did the randomized control group, suggesting that mindfulness training is central to facilitating attentional processing (Jha et al. 2007). Likewise, Moore and Malinowski (2009) found that meditators participating in a 6-week mindfulness meditation course showed greater awareness and acceptance than did non-meditators. Aside from awareness and acceptance, mindfulness was also associated with executive control, which encompasses cognitive processes such as the inhibition of dominant impulses (Diamond 2013; Oberle et al. 2012; Peters et al. 2011). Notably, self-reported dispositional mindfulness was closely related to inhibitory control and impulsivity, uncontrolled eating, and alcohol consumption (e.g., Lattimore et al. 2011; Lyvers et al. 2014; Murphy and MacKillop 2012; Oberle et al. 2012; Peters et al. 2011). By acting mindlessly without full attention, individuals were more likely to experience impulse control difficulties (Peters et al. 2011). Altogether, extant findings highlighted the complex associations between mindfulness, awareness, non-judgmental acceptance, and inhibitory control related to mental health and psychological distress.

Emotion regulation has received increasing empirical attention as a process in further linking mindfulness and mental health (Jimenez et al. 2010; Mandal et al. 2017; Roemer et al. 2015). By facilitating nonjudgmental awareness and attention in the present moment, mindfulness improves individuals’ affective detection, thereby enhancing their inhibitory learning and emotion regulation (Roemer et al. 2015). Improved emotion regulation may, then, foster the reduction of the automatic self-referential processing related to emotional distress (Roemer et al. 2015). For example, in a cross-sectional study, Jimenez et al. (2010) found that the relation between college students’ self-reported dispositional mindfulness and depressive symptoms was mediated by self-acceptance and emotion regulation. Specifically, greater dispositional mindfulness was associated with greater self-acceptance and emotion regulation, both of which were linked to fewer depressive symptoms. In another self-reported study, college students’ emotion regulation mediated between dispositional mindfulness and mental health (Coffey et al. 2010).

Previous studies demonstrated the mediation effect of specific emotion regulation strategies, such as cognitive reappraisal, between dispositional mindfulness and mental health outcomes (e.g., Curtiss et al. 2017; Desrosiers et al. 2013; Garland et al. 2011). Notably, a study conducted by Garland et al. (2011) found that self-reported dispositional mindfulness and perceived stress were mediated by positive reappraisal in a community sample. In a randomized controlled trial of mindfulness meditation, Jain et al. (2007) also found that mindfulness meditation was associated with self-reports of reduced psychological distress through rumination. Similarly, Chambers et al. (2008) found that mindfulness mediation was associated with self-reports of reduced rumination, negative affect, and depressive symptoms. The significance of emotion regulation as a mediator between mindfulness and adjustment outcomes was also noted in clinical samples. For example, a cross-sectional study conducted with a clinical sample indicated that self-reported cognitive reappraisal mediated between dispositional mindfulness and depressive symptoms (Desrosiers et al. 2013). Specifically, mindfulness predicted better cognitive reappraisal, which in turn predicted fewer depressive symptoms. In another cross-sectional study, Curtiss et al. (2017) found that self-reported cognitive reappraisal and expressive suppression mediated between dispositional mindfulness and psychological distress symptoms. Altogether, findings accumulated to date suggested that emotion regulation serves as a core process between mindfulness and mental health.

Although mindfulness training was found to benefit Western community samples, the study of mindfulness in the Chinese context remains under-explored (Lau and Hue 2011). Among the handful of recent findings, Lau and Hue (2011) found that individuals from Hong Kong who received a 6-week mindfulness-based program revealed a significant increase in well-being and decrease in depressive symptoms and anxiety symptoms. In another study, Wong et al. (2016) found that an 8-week mindfulness-based cognitive therapy course predicted a reduction of anxiety in a Hong Kong clinical sample. Similar findings in symptom reduction were revealed in a Chinese clinical sample that utilized compassion–mindfulness therapy (Lo et al. 2015). Although cultural similarities were observed in the association between mindfulness and mental health outcomes, there exist cultural differences in emotion regulation and its association with mental health between Chinese and Westerns samples (e.g., Matsumoto et al. 2008; Soto et al. 2011; Tsai et al. 2017). For example, Soto et al. (2011) found that emotion suppression was associated with mental health outcomes among European Americans from the USA, but not among Chinese individuals from Hong Kong. In another study, R. Y. M. Cheung and Park (2010) found that while anger suppression was associated with depressive symptoms in both Asian and European American samples, its strength of association differed across cultures.

The fact that the Chinese culture emphasizes in-group harmony (Smith 2010; Leung et al. 2002) may enhance non-judgmental acceptance towards negative emotions experienced by the self (Hall et al. 2011), such that individuals are harmonious with others or with the environment (Hall et al. 2011). Nevertheless, as discussed earlier, emotional acceptance may not be the same as emotional expression, as Chinese individuals also have a greater tendency to suppress emotions than their Western counterparts (Matsumoto et al. 2008). Importantly, although the cultural emphasis on face concern (e.g., Ho 1976; Smith 2010) may enhance interpersonal harmony, it may undermine intrapersonal harmony. Collectivism may also bring about concerns over individuals’ relations with other people that may give rise to anxiety (Caldwell-Harris and Aycicegi 2006). Given the complexity of harmony in the Chinese context, understanding the relation between mindfulness, emotion regulation, and psychological symptoms is extremely important in providing culture specificity and generalizability of the conceptual model (Teper et al. 2013).

Grounded in a conceptual model of mindfulness and emotion regulation (Teper et al. 2013), the present study aims to test multiple mediating processes between mindfulness and outcome variables, including depression and anxiety symptoms, in a sample of Chinese emerging adults (see Fig. 1). Controlling for baseline measures at time 1, awareness and acceptance of negative emotions, impulse control difficulties, and limited access to emotion regulation strategies were hypothesized to mediate between mindfulness and psychological distress, as indexed by symptoms of depression and anxiety. Based on previous research showing gender differences in mindfulness, emotion regulation, and psychological distress (e.g., Soysa and Wilcomb 2015), the moderating role of gender was tested in a multi-group path model.

Fig. 1
figure 1

Conceptual model of processes between mindfulness and symptoms of depression and anxiety

Full size image

Method

Participants

Participants were 333 Chinese emerging adults in Hong Kong (male = 95; female = 238) recruited online and through mass e-mailing. Participants were between 18 and 26 years of age, with a mean age of 19.96 years at time 1 (SD = 1.69 years). The retention rate at each time point was fairly high. Specifically, 87.09% (n = 290) of the participants from time 1 (T1) were retained at time 2 (T2); 95.51% (n = 277) of the participants from T2 were retained at time 3 (T3); and 90.25% (n = 250) of the participants from T3 were retained at time 4 (T4).

Procedure

The study was approved by the Human Research Ethics Committee prior to its implementation. Informed consent was obtained prior to the administration of the questionnaire. Upon informed consent, participants completed baseline measures described below. They were invited to complete follow-up questionnaires for four times in two consecutive years, with each time point spanning 6 months apart. Participants received a supermarket coupon at each time point (totaling HK$250/US$32.05 for four time points) as a token of appreciation for their participation.

Measures

Mindfulness

The Chinese version of the 39-item Five Facet Mindfulness Questionnaire (FFMQ, Baer et al. 2006) was used to assess mindfulness. Sample items included, “I pay attention to sensations, such as the wind in my hair or sun on my face” and “I rush through activities without being really attentive to them.” Participants rated on a 5-point scale from 1 (never or very rarely true) to 5 (very often or always true). The item scores were averaged to form a composite score of mindfulness, with higher scores indicating greater mindfulness. The FFMQ has been validated previously in Chinese community samples (Deng et al. 2011; Hou et al. 2014). The measure had adequate internal consistency with Cronbach’s alpha = .83.

Awareness of Negative Emotions

The 6-item Emotional Awareness subscale of the Difficulties in Emotion Regulation Scale (DERS; Gratz and Roemer 2004) was used to measure participants’ awareness of emotions on a 5-point scale from 1 (almost never) to 5 (almost always). Following the back-translation procedures (Brislin 1970), the DERS was translated from English to Chinese by two independent research assistants. Discrepancies were resolved by the first author upon follow-up discussions. Sample items included, “When I’m upset, I believe that my feelings are valid and important” and “I care about what I am feeling.” The item scores were reverse-coded and averaged to form a composite score of awareness of negative emotions, with higher scores indicating greater awareness. The measure had been validated in a Mainland Chinese sample (Wang et al. 2007). Confirmatory factor analysis of the present Chinese measure at T1 yielded excellent fit to the data, χ2 (5) = 4.23, p = .52, CFI = 1.00, TLI = 1.00, RMSEA = .00, SRMR = .01. All of the factor loadings were significant at ps < .01. The measure also had adequate internal consistency with Cronbach’s alpha = .71 at T1 and .74 at T2.

Acceptance of Negative Emotions

The 6-item Emotional Acceptance subscale of DERS (Gratz and Roemer 2004) was used to measure participants’ awareness of emotions on a 5-point scale from 1 (almost never) to 5 (almost always). Following the back-translation procedures (Brislin 1970), the DERS was translated from English to Chinese by two independent research assistants. Discrepancies were resolved by the first author upon follow-up discussions. Sample items included, “When I’m upset, I feel ashamed with myself for feeling that way” and “When I’m upset, I become irritated with myself for feeling that way.” The item scores were reverse-coded and averaged, such that the composite score reflected acceptance of negative emotions, with higher scores indicating greater acceptance. Confirmatory factor analysis of the present Chinese measure at T1 yielded acceptable fit to the data, χ2 (5) = 17.43, p = .07, CFI = .98, TLI = .95, RMSEA = .09, SRMR = .03. All of the factor loadings were significant at p < .001. The measure also demonstrated adequate internal consistency with Cronbach’s alpha = .85 at T1 and .87 at T2.

Impulse Control Difficulties

The 6-item Impulse Control Difficulties subscale of DERS (Gratz and Roemer 2004) was used to measure participants’ awareness of emotions on a 5-point scale from 1 (almost never) to 5 (almost always). Following the back-translation procedures (Brislin 1970), the DERS was translated from English to Chinese by two independent research assistants. Discrepancies were resolved by the first author upon follow-up discussions. Sample items included “When I’m upset, I have difficulty controlling my behaviors” and “I experience my emotions as overwhelming and out of control.” The item scores were averaged to form a composite score of impulse control difficulties, with higher scores indicating greater difficulties. Confirmatory factor analysis of the present Chinese measure at T1 yielded excellent fit to the data, χ2 (5) = 9.21, p = .10, CFI = 1.00, TLI = .99, RMSEA = .05, SRMR = .02. All of the factor loadings were significant at p < .001. The measure demonstrated adequate internal consistency with Cronbach’s alpha = .86 at T1 and .85 at T3.

Limited Access to Emotion Regulation Strategies

The 8-item Limited Access to Emotion Regulation Strategies subscale of DERS (Gratz and Roemer 2004) was used to measure participants’ awareness of emotions on a 5-point scale from 1 (almost never) to 5 (almost always). Following the back-translation procedures (Brislin 1970), the DERS was translated from English to Chinese by two independent research assistants. Discrepancies were resolved by the first author upon follow-up discussions. Sample items included, “When I’m upset, I believe that there is nothing I can do to make myself feel better” and “When I’m upset, I believe that I will remain that way for a long time.” The item scores were averaged to form a composite score of limited emotion regulation strategies, with higher scores indicating greater limited strategies. Confirmatory factor analysis of the present Chinese measure at T1 yielded good fit to the data, χ2 (16) = 29.28, p = .02, CFI = .99, TLI = .98, RMSEA = .05, SRMR = .03. All of the factor loadings were significant at p < .001. The measure demonstrated adequate internal consistency with Cronbach’s alpha = .86 at T1 and .87 at T4.

Depressive Symptoms

The 20-item Center for Epidemiologic Studies Depression Scale (CES-D; Radloff 1977) was used to measure depressive symptoms in the past week on a 4-point scale from 1 (rarely or none of the time, less than 1 day) to 4 (most or all of the time, 5–7 days). Sample items included, “I felt depressed” and “I felt sad.” The item scores were summed to form a composite score of depressive symptoms, with higher scores indicating more symptoms. The measure has previously been validated in a Chinese community sample (C. K. Cheung and Bagley 1998). The measure demonstrated adequate internal consistency with Cronbach’s alpha = .91 at T1 and .84 at T4.

Anxiety Symptoms

The 7-item Generalized Anxiety Disorder-7 measure (GAD-7; Spitzer et al. 2006) was used to measure depressive symptoms in the past 2 weeks on a 4-point scale from 0 (not at all) to 3 (nearly every day). Sample items included “Not being able to stop or control worrying” and “Feeling nervous, anxious or on edge.” The item scores were averaged to form a composite score of anxiety symptoms, with higher scores indicating more symptoms. The measure has previously been validated in a Chinese sample of hospital outpatients (He et al. 2010) and a sample of Chinese individuals with epilepsy (Tong et al. 2016). Confirmatory factor analysis for the present community sample of emerging adults at T1 yielded good fit to the data, χ2 (10) = 26.62, p = .003, CFI = .99, TLI = .98, RMSEA = .07, SRMR = .02. All of the factor loadings were significant at p < .001. The measure demonstrated adequate internal consistency with Cronbach’s alpha = .93 at T1 and .94 at T4.

Data Analysis

Zero-order correlations, means, and standard deviations were conducted as preliminary analyses for all study variables. Multi-group path analysis was then used to test the central hypotheses for this study between genders. MPLUS, Version 7 (Muthén and Muthén 2012) was used to investigate the path model of awareness, attention, impulse control difficulties, and limited access to emotion regulation strategies between mindfulness and adjustment outcomes, including depressive and anxiety symptoms, over and above the effect of baseline measures. More specifically, T1 mindfulness was entered to predict T2 awareness and T2 attention. T2 awareness and T2 attention was entered to predict each other contemporaneously. T2 awareness and T2 attention were, then, entered to predict T3 impulse control difficulties. T3 impulse control difficulties were entered to further predict T4 limited access to emotion regulation strategies, which was, then, entered to predict T4 depressive and anxiety symptoms. All measures were controlled at T1 and direct effects were included between all predictors and outcome variables (e.g., T1 mindfulness and T4 depressive symptoms; T2 awareness and T4 anxiety symptoms). Maximum likelihood method was used to investigate the model fit to the observed matrices of variance and covariance. Full information maximum likelihood estimation was used to handle missingness. Bootstrapping was used to evaluate the mediation effects, as it could yield more accurate estimates of the indirect effect standard errors than other approaches to testing mediation (Shrout and Bolger 2002).

Results

Table 1 shows the zero-order correlations, means, and standard deviations for all variables under study.

Table 1 Zero-order correlations, means, and standard deviations of the study variables
Full size table

In the first step of the multi-group path analysis, all parameters were freely estimated between genders. In the next step, the unstandardized parameter estimates were constrained to be invariant between genders. A non-significant χ2 difference between the two models infers that the simpler model with constrained paths should be selected due to parsimony. In step 1, allowing all paths to differ between gender yielded good fit to the data, χ2(56) = 94.64, p = .001, CFI = .97, TLI = .93, RMSEA = .06, SRMR = .05. In step 2, gender similarities were demonstrated through a non-significant χ2 difference test between the models, after constraining all paths to be equal, ∆χ2(27) = 30.30, p = .30. Therefore, the findings demonstrated universality of the model between genders, with the constrained model yielding good fit to the data, χ2 (83) = 124.94, p = .002, CFI = .96, TLI = .95, RMSEA = .06, SRMR = .06.

Table 2 indicates the parameter estimates in the multi-group path model, with unstandardized Bs and SEs constrained to be equal between genders. Controlling for T1 awareness and acceptance of negative emotions, T1 mindfulness predicted T2 awareness of negative emotions (βmale = .17; βfemale = .19, ps < .01) and T2 acceptance of negative emotions (βmale = .13; βfemale = .14, ps < .05). However, T2 awareness and acceptance of negative emotions did not predict each other, ps > .05. T2 awareness and acceptance of negative emotions further predicted T3 impulse control difficulties (awareness: βmale and βfemale = − .14, respectively, ps < .01; acceptance: βmale = − .21; βfemale = − .23, respectively, ps < .001). T2 acceptance of negative emotions and T3 impulse control difficulties, in turn, predicted T4 limited access to emotion regulation strategies (acceptance: βmale and βfemale = − .30, respectively, ps < .001; impulse control: βmale = .19; βfemale = .18, ps < .01). Finally, T4 limited access to emotion regulation strategies predicted T4 outcomes, including depressive symptoms (βmale = .52; βfemale = .54, ps < .001) and anxiety symptoms (βmale and βfemale = .33, respectively, ps < .001). In addition, T4 depressive symptoms were negatively predicted by T2 awareness to negative emotions (βmale = − .16; βfemale = − .15, ps < .01). Autoregressive control variables at T1 were incorporated in the model for all study variables and the predictions were significant at ps < .001.

Table 2 Unstandardized parameter estimates, standard errors, and p values in the model
Full size table

The mediating roles of awareness, acceptance, and impulse control difficulties between mindfulness and limited access to emotion regulation strategies were tested via bootstrapping based on 1000 bootstrap samples with replacement. Prior to bootstrapping, the indirect effects (i.e., the product of regression coefficients) between mindfulness and limited access to emotion regulation strategies were significant (βmale and βfemale = − .06, respectively, ps < .01). The 95% confidence interval (CI) further indicated that the unstandardized indirect effect of mindfulness on limited access to emotion regulation strategies did not include a zero [CI: (− .11, − .02)], suggesting that baseline mindfulness was associated with a chain of mechanisms including greater awareness and acceptance of negative emotions at T2, lower impulse control difficulties at T3, and greater access to emotion regulation strategies at T4.

Next, the mediating roles of awareness, acceptance, impulse control difficulties, and limited access to emotion regulation strategies between mindfulness and mental health outcomes were tested via bootstrapping based on 1000 bootstrap samples with replacement. Prior to bootstrapping, the indirect effect between mindfulness and depressive symptoms was significant (βmale and βfemale = − .09, respectively, ps < .05). The 95% CI indicated that the unstandardized indirect effect of mindfulness on depressive symptoms did not include a zero [CI: (− .16, − .02)], suggesting mediation processes, i.e., baseline mindfulness was associated with greater awareness and acceptance of negative emotions at T2, which in turn was associated with lower impulse control difficulties at T3, and was further associated with greater access to emotion regulation strategies and fewer depressive symptoms at T4. Although a trend of indirect effect between mindfulness and anxiety symptoms was found (βmale and βfemale = − .05, respectively, ps = .08), the 95% CI indicated that the unstandardized indirect effect of mindfulness on anxiety symptoms did include a zero [CI: (− .11, .006)]. Consequently, the tested processes did not serve as mediators between mindfulness and anxiety symptoms.

Discussion

Building on a theoretical model of mindfulness and emotion regulation (Teper et al. 2013), this prospective study investigated the mediating roles of cognitive and emotional processes between mindfulness and psychological distress, including symptoms of depression and anxiety, in a Chinese emerging adult sample. Unique findings emerged to suggest that mindfulness was related to fewer symptoms of depression through a chain of underlying mechanisms, including greater awareness and acceptance of negative emotions, fewer impulse control difficulties, and more adequate use of emotion regulation strategies. Gender similarities were demonstrated to reflect homogenous pathways between genders in the path model. A marginal, though nonsignificant, trend was also indicated for the mediation effects between mindfulness and anxiety symptoms. These findings add to the accumulating evidence concerning the processes linking mindfulness and mental health (e.g., Bränström et al. 2011; Carsley et al. 2018; Chambers et al. 2009; Lyvers et al. 2014; Roemer et al. 2009). The findings also advance the extant literature by establishing the association between mindfulness and psychological distress using a longitudinal, process-oriented approach.

Unique to this study is the model substantiating the mediation findings over and above the effects of autoregressive control (see Fig. 2). By orienting fully to the present experience and paying attention “on purpose, in the present moment, and non-judgmentally” (Kabat-Zinn 1994; see also Bishop et al. 2004), mindfulness enhanced emerging adults’ awareness of their negative emotions. The non-judgmental emphasis of mindfulness was also related to unbiased acceptance of emotions associated with upsetting experiences, such that individuals were less preoccupied with avoiding, invalidating, or rejecting these emotions. On the contrary, when individuals engaged in a state of “autopilot” condition (Germer et al. 2016), they missed invaluable opportunities to observe, understand, and accept the emotions arising in the present moment, thereby jeopardizing their subsequent regulatory processes and mental health.

Fig. 2
figure 2

Mediation findings between mindfulness and symptoms of depression and anxiety. χ2(83) = 124.94, p = .001, CFI = .96, TLI = .95, RMSEA = .06, SRMR = .06. **p ≤ .001, ***p ≤ .001. Unstandardized parameter estimates and standard errors in parentheses are presented (see Table 2 for detailed findings). Direct effects are not included for the purpose of clarity

Full size image

Surprisingly, although the awareness and the acceptance of negative emotions were both associated with mindfulness, they were not correlated with each other within and between T1 and T2 (see Table 1). Likewise, the path model revealed that awareness and acceptance did not predict one another. These findings were intriguing, as they indicated that attentiveness to negative emotions was not sufficient to promote non-judgmental acceptance of the emotions (Cardaciotto et al. 2008). Likewise, accepting negative emotions without judgment or prejudice did not necessarily foster greater awareness of these emotions. Previous studies conducted in Western societies indicated that acceptance and awareness of negative emotions were weakly but significantly associated (e.g., Giromini et al. 2012; Gratz and Roemer 2004), whereas other studies indicated a negligible association (e.g., Weinberg and Klonsky 2009). The present findings suggested that acceptance and awareness were not significantly related to each other over time in the Chinese context. Given the inconsistent findings in the literature, future use of multiple methods and measures is needed to decipher how emotional acceptance and awareness are associated with each other (or not) in relation to dispositional mindfulness and mental health in diverse contexts.

Returning to Fig. 2, both awareness and acceptance of negative emotions predicted impulse control difficulties, which served as an index of executive control. Although the present research did not examine momentary changes between awareness, acceptance, and impulse control difficulties, thereby precluding us from revealing causal and moment-to-moment effects, we found that attuning to the present moment was longitudinally associated with a greater opportunity for individuals to deliberately observe, attend to, and accept their upsetting emotions non-judgmentally (Kabat-Zinn 1994). The unbiased observation and open acceptance were, then, linked to individuals’ efficiency in responding to affective cues (Roemer and Orsillo 2002; Teper et al. 2013) and a lower tendency to perceive that the negative emotions were overwhelming and hard to control. Altogether, these processes might lessen impulsive urges (see also Peters et al. 2011) associated with adjustment problems. With greater emotional acceptance and fewer impulse control difficulties, individuals were, in turn, more likely to gain access to emotion regulation strategies that were necessary to relieve their upsetting emotions and/or experiences. Altogether, the present findings supported Teper et al. (2013) conceptual framework, in that awareness, acceptance, and executive control served as important processes between mindfulness and emotion regulation. Regardless of gender, dwelling in the present moment purposefully and non-judgmentally predicted better emotion regulation over time. Extending Teper et al. framework, our findings also revealed that depressive symptoms and anxiety problems accrued as a result of mindlessness, impulse control difficulties, and emotion dysregulation. These findings echoed with previous evidence on the association between emerging adults’ emotion dysregulation and poorer mental health (e.g., R. Y. M. Cheung and Park 2010; R. Y. M. Cheung et al. 2018; Gross 1998; Hu et al. 2014; Klemanski et al. 2017; Rodriguez et al. 2016). By providing longitudinal support for the effect of mindfulness on mental health, the present findings serve important functions to inform prevention and intervention efforts in mental health promotion.

Limitations and Future Directions

The study has several limitations that merit consideration. First of all, the autoregressive control variables were limited to T1. In addition, the study consisted of four, instead of five, time points to demonstrate the temporal sequence among variables. According to Cole and Maxwell (2003), the rigor of longitudinal mediation models can be strengthened by including autoregressive control variables at all time points. Second, in this study, we did not measure cultural correlates of mindfulness, emotion regulation, and mental health. Given that previous research demonstrated cultural variations in the association between emotion regulation and mental health (e.g., Hu et al. 2014; Markus and Kitayama 1991; Matsumoto et al. 2008), future studies are necessary to determinate how the model is similar or different across cultures. Moreover, the emphasis of harmony in the Chinese culture (Smith 2010; Leung et al. 2002) may foster non-judgmental orientations toward negative emotions and experiences and expressive suppression at the same time. Given the relevance of culture, future work is necessary to examine the effect of cultural values on mindfulness, emotion regulation, and mental health.

In addition to culture, the gender imbalance of the present sample (male = 95; female = 238) limits generalizability of the findings, particularly to men. Consistent with previous research suggesting gender similarities of the relationship between mindfulness and mental health (e.g., Bränström et al. 2011; Carsley et al. 2018; Chambers et al. 2009; Lyvers et al. 2014; Roemer et al. 2009), future studies are necessary to examine the relation in a larger, gender-balanced sample. Fourth, the present study utilized self-report measures. Future studies may use a multi-informant and multi-method approach (e.g., physiological measures, observational measures, and questionnaires) to increase objectivity. Fifth, subscales of DERS were used as independent variables in the path model. Although previous research demonstrated differential predictability of DERS subscales on psychological outcomes (e.g., Bardeen and Stevens 2015; Bender et al. 2012; Bonn-Miller et al. 2008; Merwin et al. 2010), the subscales were generally correlated based on findings in validation studies (Bardeen et al. 2012; Weinberg and Klonsky 2009). In order to maximize independence among variables, future work should utilize different methods and multiple measures to assess each variable of interest.

Next, beyond the investigation of the purported model, it is possible that the predictions occur in a reverse order, such that psychological outcomes predict mindfulness via limited access to emotion regulation strategies, impulse control difficulties, and awareness and acceptance of negative emotions. It is also possible that multiple processes co-occur to predict well-being outcomes. Although theories are necessary to guide the hypothesized pathways, future research should collect more data to rule out reverse directionality of effects. Seventh, given the present sample of emerging adults, the degree to which the findings can be generalized to other developmental periods is uncertain, such as early adolescence or older adulthood. Also, as previous research suggested that community and clinical samples showed differences in the relationship between emotion regulation strategies and mental health (e.g., Garnefski et al. 2002; Roemer et al. 2015), future work is needed to compare and contrast the processes between community and clinical samples across developmental periods to demonstrate specificity.

Finally, the present study examined dysregulation of negative emotions as a mediating mechanism. Future studies may test competing theories to investigate how specific emotion regulation strategies (e.g., rumination, expressive suppression, and cognitive reappraisal) and specific emotions (e.g., happiness vs. anger) are associated with mindfulness and mental health. For example, Garland et al. (2015) conceptualized an upward regulation model, in that mindfulness generates positive emotion regulation that ultimately leads to meaningfulness in life, whereas Teper et al. (2013) focused on downward regulation of negative emotions via executive control using mindfulness. Testing competing models (e.g., upward vs. downward regulation of positive vs. negative emotions) can, ultimately, illuminate the determining processes between mindfulness and well-being. Notably, the up-regulation of positive emotions may serve as an important process to enhance mental health and alleviate psychological distress (e.g., R. Y. M. Cheung et al. 2018; Kiken et al. 2017).

Notwithstanding the above limitations, this study broadens the literature by providing new information about the longitudinal mediating mechanisms between mindfulness and mental health in a community context. These findings not only lend longitudinal support to the Teper et al. (2013) theoretical model, but also makes an initial effort to apply the model to an Eastern society.

This study calls attention to the relation between mindfulness and mental health through multiple underlying emotional processes, including emotional awareness and acceptance, impulse control difficulties, and emotion regulation. Grounded in a conceptual model of mindfulness and emotion regulation (Teper et al. 2013), the findings emphasized the unfolding cognitive and emotional benefits associated with mindfulness, regardless of gender. The study also underscores the longitudinal prediction of mindfulness on psychological distress in emerging adulthood, including depression and anxiety symptoms. Further to these findings, examination of competing theories of mindfulness, emotion regulation, and mental health in diverse contexts warrants future research. Psychological interventions, translational programs, and public health campaigns geared toward enhancing mindfulness in promoting mental health merits future investigation.