Does Experiential Avoidance Mediate the Relationship Between Gender Role Conflict and Psychological Distress?

Psychological inflexibility or experiential avoidance (EA) is an important construct in the understanding of psychological distress. Both EA and many forms of masculinity can be characterized by inflexibility in men's responses to negative intrapersonal experiences. The current cross-sectional, community-based study investigated whether experiential avoidance mediated the relationship between gender role conflict (GRC) and psychological distress (PD). A total of 120 men ( M = 35.63, SD = 12.22) completed an online questionnaire measuring key study variables. Results indicated that experiential avoidance significantly mediated the relationship between each of the four recognized patterns of GRC and PD. These findings suggest that EA may be a potential mechanism through which GRC is associated with PD.

Interest in the construct of psychological flexibility/inflexibility has grown in recent years. This is due to various factors such as the development of “third-wave” psychotherapeutic models, particularly Acceptance and Commitment Therapy (ACT). There has also been increased application of flexibility in well-established areas, such as coping (Cheng, Lau, & Chan, 2014). In addition, psychological flexibility has been a core component in long-standing domains of empirical research, such as repetitive though (particularly rumination), neuropsychological (executive) functioning, and emotion regulation (Kashdan & Rottenberg, 2010). Not surprisingly, there is significant variation in how this construct has been conceptualized and defined.

In the ACT model of psychopathology, psychological inflexibility or “experiential avoidance” is posited to be a key factor in the development of psychological distress. Experiential avoidance is hypothesized to involve attempts to control or alter uncomfortable intrapersonal experiences in a way that compromises personal values, or the pursuit of goals (Hayes, Wilson, Gifford, Follette, & Strosahl, 1996). In experiential avoidance, actions are guided not by values or goals, but by psychological reactions to these internal experiences (Bond et al., 2011). Several empirical studies have implicated experiential avoidance in various psychological problems across cultures (Blakey, Jacoby, Reuman, & Abramowitz, 2016; Nam, 2016; Valiente, Provencio, Espinosa, Duque, & Everts, 2015), and psychological inflexibility more broadly defined has been associated with depression and anxiety (Kashdan & Rottenberg, 2010).

Inflexibility/experiential avoidance may have particular relevance to men given the conceptual similarities with masculinity ideology, and hegemonic masculinity in particular. The latter is a framework that assumes masculinity is an idealized but pluralistic view of how men should feel, think, and act. This paradigm has emerged from the work of scholars such as Pleck (1995) and Connell (1995), among others. It is widely agreed that adherence to masculine ideology is associated with reduced psychological well-being (Cournoyer & Mahalik, 1995; Mahalik, Pierre, & Wan, 2006).

Masculine ideology can be viewed as relevant to experiential avoidance in that they can both be characterized by an inflexible approach to intrapersonal experiences. For example, hegemonic masculinity represents a highly prescribed (inflexible) form of “normative” masculine ideology characterized by avoidance of experience including affect (Levant, Hall, & Rankin, 2013). Similarly, experiential avoidance (as indicated above) encapsulates an inflexible way of responding to uncomfortable inner experiences. Consequently, examining the relationship between experiential avoidance and masculinities may be an important means by which we can improve our understanding of key men’s mental health issues such as low help-seeking rates, and over-representation in certain mental health issues such as suicide and substance use disorders (Vogel & Heath, 2016).

An important extension of this masculine ideology/gender role strain paradigm is gender role conflict (GRC) which has identified patterns of negative consequences resulting from masculine socialization. GRC has been defined as “…a psychological state in which socialized gender roles have negative consequences for the person or others…[and]…occurs when rigid, sexist, or restrictive gender roles result in restriction, devaluation, or violation of others or self” (O’Neil, 2008; p. 362). GRC is a particularly prominent construct in the masculinity literature, with several studies establishing a relationship between GRC and psychological distress (Magovcevic & Addis, 2005; Theodore & Lloyd, 2000; Wester, Christianson, Vogel, & Wei, 2007).

Four conflict “patterns” have been established empirically through the development of the Gender Role Conflict Scale (GRCS): (a) success, power, and competition; (b) restrictive emotionality; (c) restrictive affectionate behavior between men; and (d) conflicts between work and leisure. GRC can also be experienced in different contexts, such as the intrapersonal domain via negative cognition and affect resulting from devaluations, restrictions, and violations of gender roles (O’Neil, 2008). The concept of inflexibility is again apparent in GRC through this intrapersonal restriction, and inhibited display of emotion as characterized in conflict patterns two and three. Experiential avoidance may represent a maladaptive response to intrapersonally manifested GRC through further inflexibility thereby representing a potential mechanism through which GRC leads to psychological distress. A recent meta-analysis has suggested that coping flexibility has a significant effect on psychological well-being (Cheng et al., 2014). Factors, such as experiential avoidance, that restrict one’s coping repertoire and ability to respond to situation demands is therefore likely to lead to increasingly negative mental health outcomes.

There is little or no empirical research investigating a potential relationship between GRC and psychological flexibility and/or experiential avoidance. Despite this, GRC incorporates restricted emotionality (see below), and emotional suppression has also been identified as a component of experiential avoidance (Chawla & Ostafin, 2007). Wong and colleagues have also found that restricted emotionality is associated with unfavorable attitudes toward emotional expression (Wong, Pituch, & Rochlen, 2006). Furthermore, GRC has been linked to alexithymia (Guvensel, Dixon, Chang, & Dew, 2017). In sum, there appears to be a relationship between GRC and measures of unwillingness or inability to engage with internal emotional experiences.

To summarize, various forms of avoidance are key components of some forms of masculinity (particularly idealized or hegemonic manifestations) and GRC. A potentially important relationship may therefore exist between GRC and experiential avoidance. When avoidance is a key feature of a masculine script, it may represent a mechanism through which psychological distress results. Identifying mechanisms by which GRC results in psychological distress is of great interest because this information could be used to help address key men’s mental health issues, and help refine current psychotherapeutic approaches with men.

The Current Study

Previous research has not attempted to examine experiential avoidance as a mechanism through which GRC leads to psychological distress. In this way, the current study seeks to provide a novel and potentially useful contribution to our understanding of mental health difficulties in men. The current study sought to investigate the relationship of GRC to experiential avoidance and psychological distress, with the primary research question being “Does experiential avoidance mediate the relationship between GRC (as measured by each subscale of the GRCS), and psychological distress?” The hypothesized relationship between the three main study variables is presented in the conceptual diagram in Figure 1. The complete set of current study hypotheses is given below:

Figure 1.

Hypothesized relationships among study variables.

A positive relationship would be identified between GRC (as measured by each of the subscales on the GRCS) and psychological distress.

A positive relationship would be identified between GRCS subscales and experiential avoidance.

A positive relationship would be identified between experiential avoidance and psychological distress.

Experiential avoidance would mediate the relationship between GRCS subscales and psychological distress.

Method

Participants

Male participants aged 18 years and over in the UK were invited to complete an online questionnaire that included the key study variables, along with basic demographic and socioeconomic information. The current study was advertised using social media (Facebook and Twitter), email promotion to organizations interested specifically in men’s mental health, and poster advertisements in a university campus. In total, 120 men participated in the current study, with an average age of 35.63 years (SD = 12.22). An a priori sample size calculation was conducted using the guidelines of Fritz and MacKinnon (2007). Based on previous research findings, medium effect sizes were estimated for the α and β pathways of the indirect effect. This resulted in a sample size requirement of 115. Demographic data requested on the questionnaire included age, ethnicity, relationship and employment status. The ethnic composition of the sample was 48% (N = 58) White British/Caucasian, 18% (22) Continental European, 10% (13) White other, 4% (5) Caribbean, 3% (4) Irish, 3% (4) Continental African, 10% (13) “other” ethnicities. Relationship status identification was 42% (50) single, 31% (37) married/civil partnership, 14% (17) cohabitating, 5% (6) separated/divorced, and 8% (9) “other” status. Reported employment status was 55% (66) employed full time, 20% (24) student, 13% (16) employed part time, 6% (7) self-employed, and 6% (7) unemployed. Missing data analysis was completed using Little’s Missing Completly At Random (MCAR) test. Results of this test indicated a χ2 = 48.92 (df = 214; p = .247). The nonsignificant result suggests that data were missing at random.

Measures

In addition to the collection of basic demographic information (as reported above), the following measures were administered:

Gender role conflict

GRC was measured with the Gender Role Conflict Scale (GRCS; O’Neil, Helms, & Gable, 1986). The GRCS is a 37-item self-report measure of gender role conflict among men. Participants indicate the extent to which they agree with a series of statements on 6-point Likert scale (ranging from “strongly disagree” to “strongly agree”). Example items include “Affection with other men makes me tense” and “Being smarter or physically stronger than other men is important to me.” Higher scores indicate higher levels of gender conflict. The GRCS is comprised of four subscales (number of items per subscale indicated in parentheses); success, power, and competition (13); restrictive emotionality (10); restrictive affectionate behavior (8); and conflict between work and family relations (6). Support for the four-factor structure of the GRCS has been reported (e.g., Moradi, Tokar, Schaub, Jome, & Serna, 2000), and it has good internal consistency (α = .70 to .89) as reported across several studies (O’Neil, 2008). In the present study, the α coefficient for the overall scale was .89 and the α coefficients for the subscales were .82 (SPC), .68 (RE), .67 (RAB), and .81 (CWL).

Experiential avoidance

EA was measured with the Acceptance and Action Questionnaire-II (AAQ-II; Bond et al., 2011). The AAQ-II is a seven-item self-report instrument that measures psychological inflexibility/experiential avoidance. The AAQ-II comprises a series of statements that utilize a 7-point Likert response scale (ranging from “never true” to “always true”). Example items include “Worries get in the way of my success” and “I’m afraid of my feelings.” Higher scores indicate greater levels of inflexibility. Satisfactory psychometric properties have been reported by Bond et al. (2011) in the domains of structure, reliability, and validity. A single-factor structure was reported by these authors, along with good internal consistency (mean α = .84 across six samples) and test–retest reliability (α = .79 at 12 months) (Bond et al., 2011). The α coefficient for the AAQ-II in the current study was .93.

Psychological distress

This was measured with the Depression Anxiety Stress Scale (DASS; Lovibond & Lovibond, 1995). The DASS is a 42-item self-report instrument that measures depression, anxiety, and stress. Items are a series of statements that utilize a 4-point Likert scale (ranging from “does not apply to me at all” to “applied to me very much or most of the time”). Example items include “I felt sad and depressed,” “I was aware of dryness of my mouth,” and “I found it hard to wind down.” Higher scores indicate greater levels of psychological distress. Excellent internal consistency has been reported for the instrument overall (α = .97), and for each of the depression, anxiety, and stress subscales (α = .96, .92, and .95, respectively). Internal consistency for the DASS in the current study was comparably high (.97).

Procedure

A favorable ethical opinion for the research was sought and obtained from the University of Surrey’s University Ethics Committee. The above study measures were combined with demographic questions to form a larger questionnaire that was made available for completion online using Qualtrics survey design software (Qualtrics, 2005). After following the questionnaire link, prospective participants were first presented with an information sheet that provided information about the study purpose, along with anticipated risk and benefits of taking part. For those who continued, a consent form was then presented where proceeding to the start of questionnaire confirmed consent to taking part. Participants were free to withdraw from questionnaire completion at any time. Data from these individuals were removed from the dataset and not analyzed. For those who did complete the questionnaire, they completed the study measures as described above.

Statistical Analysis

Descriptive statistics and bivariate correlations were calculated for study variables. Qualitative descriptors for effect sizes were based on Cohen’s (1988) recommendations for the identification of small (.1 ≤ r < .3), medium (.3 ≤ r < .5), and large (r ≥ .5) effects. The mediation analysis in the current study was conducted using the PROCESS macro for SPSS (Hayes, 2013). Specifically, this was used to test for the presence of a mediating effect for experiential avoidance in the relationship between GRC and PD. The macro utilized bootstrapping to obtain regression coefficients for each path in the model, and parameter estimates of the mediation effect were obtained with 10,000 samples. The criterion used for determining significance of the mediation effect was the 95% bias-corrected confidence interval for the indirect effect not containing a value of 0. Bootstrapping is preferable to the causal steps method (Baron & Kenny, 1986) due to the latter being characterised by low power and nonquantitative derivation of an indirect effect estimate (Hayes, 2009).

Effect size estimates for the indirect effect were calculated in the current study, but the use of effect sizes in simple mediation analysis remains controversial. The P effect size was used in the current investigation following recent concerns raised about the use of k2 (Wen & Fan, 2015). The recommendations of Wen and Fan (2015) were followed in the current study by reporting P when the indirect and total effects are both in the same direction. This index represents the ratio of indirect to total effect.

Results

The parametric assumptions of normality, no multicollinearity, homogeneity of variance, and no outliers were assessed and met prior to data analysis. Demographic variables measured for this study were not significantly correlated with PD. Descriptive statistics for the measured study variables are presented in Table 1. The key study variables were related to each other in the expected directions as stated in Hypotheses 1–3; a positive association between GRC and EA (r = .39, p < .001), a positive association between GRC and PD (r = .42, p < .001), and a positive association between EA and PD (r = .75, p < .001).

Table 1.

Means, Standard Deviations, and Correlations Among Study Variables.

Variable	M	SD	1	2	3	4	5	6	7	8	9	10
1. Age	35.63	12.22
2. AAQ_Total	22.35	9.02	−.17
3. DASS_Total	28.44	22.90	−.09	.75**
4. DASS_Stress	12.49	8.77	−.15	.73**	.94**
5. DASS_Anxiety	6.31	6.57	−.15	.55**	.86**	.76**
6. DASS_Depression	10.02	9.79	−.03	.68**	.92**	.79**	.67**
7. GRCS_Total	30	8.5	.40**	.39**	.42**	.47**	.26**	.38**
8. GRCS_SPC	3.53	0.87	.44**	.22*	.24**	.30**	.12	.21*	.83**
9. GRCS_RE	3.36	0.74	.23**	.47**	.34**	.38**	.17	.33**	.76**	.39**
10. GRCS_RABBM	3.12	0.73	.20*	.22*	.31**	.34**	.22*	.28**	.72**	.38**	.58**
11. GRCS_CBWFR	3.63	1.12	.29**	.33*	.45**	.48**	.34**	.41**	.79**	.54**	.48**	.45**

Note. AAQ_Total = Acceptance and Action Questionnaire II total score; DASS_Total = Depression Anxiety Stress Scale total score; DASS_Stress = DASS stress subscale score; DASS_Anxiety = DASS anxiety subscale score; DASS_Depression = DASS depression subscale; GRCS_Total = Gender Role Conflict Scale total score; GRCS_SPC = GRCS success, power, and competition subscale score; GRCS_RE = GRCS restrictive emotionality subscale score; GRCS_RABBM = GRCS restrictive affectionate behavior between men subscale score; GRCS_CBWFR = GRCS conflict between work and family relations subscale score.

p = < .05. **p = < .01.

The results of the mediation/path analysis are summarized in Table 2. Consistent with Hypothesis 4, EA significantly mediated the relationship between GRC and PD (based on the criterion for mediation stated above). This was the case for each of the GRCS subscales; success, power, and competition (SPC) b = .302, BCa CI [0.015, 0.634], P = .62, restrictive emotionality (RE) b = 1.094, BCa CI [0.740, 1.499], P = 1.054, restrictive affectionate behavior between men (RAB) b = 0.610, BCa CI [0.128, 1.132], P = .51, and conflict between work and leisure (CWL) b = 0.758, BCa CI [0.346, 1.203], P = .49. Comparison of the effect size estimates P indicated that the mediation effect was largest for the restrictive emotionality subscale of the GRCS.

Table 2.

Psychological Flexibility as a Mediator in the Relationship Between GRC and Psychological Well-being (DASS Scores).

Predictor variable	R2	Path c (SE)	Path a (SE)	Path b (SE)	Path c’ (SE)	Path a × b (SE)	95% CI of path a × b	PM
GRCS_Total	.58***	.390*** (.079)	.141*** (.061)	1.77*** (.165)	.141* (.061)	.249 (.068)	[.132, .393]	.64
GRCS_SPC	.58***	.485** (.180)	.161* (.071)	1.871*** (.156)	.183 (.124)	.302 (.158)	[.015, .634]	.62
GRCS_RE	.56***	1.038*** (.266)	.564*** (.098)	1.93*** (.176)	-.056 (.212)	1.094 (.192)	[.740, 1.499]	1.054
GRCS_RAB	.59***	1.189*** (.339)	.332* (.136)	1.835*** (.154)	.579* (.235)	.610 (.257)	[.128, 1.132]	.51
GRCS_CWL	.62***	1.541*** (.277)	.439*** (.115)	1.725*** (.154)	.783*** (.205)	.758 (.219)	[.346, 1.203]	.49

Note. CIs (95%) calculated with 5,000 bootstrap samples. GRCS_Total = Gender Role Conflict Scale total score; GRCS_SPC = GRCS success, power, and competition subscale; GRCS_RE = GRCS restrictive emotionality; GRCS_RAB = GRCS restrictive affectionate behavior between men; GRCS_CWL = GRCS conflicts between work and leisure.

p < .05. **p < .01. ***p < .001.

Discussion

The current study investigated experiential avoidance (EA) as a potential mechanism through which GRC leads to psychological distress. It was proposed that GRC can lead to avoidance which, in turn, results in adverse consequences for well-being. Results supported the hypothesis that EA mediates the relationship between the four recognized patterns of GRC and psychological distress. These findings add to the body of literature that has identified several potential mediators of the link between GRC and PD. For example, Szymanski and Ikizler (2013) reported that internalized heterosexism mediated the relationship between GRC and depression among sexual minority men.

The current research responds to recent calls for more mediation studies in GRC to aid theory development (O’Neil, 2008). Current findings suggest the effect of GRC on psychological distress could operate, to some extent, through reduced coping flexibility and an over-reliance on an avoidant coping strategy, namely EA. Other research has indicated that EA is maladaptive (Blakey et al., 2016; Nam, 2016; Valiente et al., 2015), and Cheng et al. (2014) argue through their meta-analytic findings that coping flexibility is an important factor in psychological well-being. The current study appears to provide further support in this area. EA may be detrimental to mental health in men due to its representation as a maladaptive response to GRC.

Effect size analysis indicated that the mediation effect was largest for the restrictive emotionality pattern of GRC. Restrictive emotionality “…is defined as having restrictions and fears about expressing one’s feelings as well as restrictions in finding words to express basic emotions (O’Neil, 2008). It could be argued that restrictive emotionality is more conceptual similar to experiential avoidance compared with other GRC patterns, and this could account for the effect size results reported here. Speculating more broadly, GRC and EA may be differentiated in that restrictive emotionality is an initial psychological state, whereas EA represents a response to that state intended to modulate unpleasant internal experiences resulting from GRC. For instance, restrictive emotionality appears related to other variables, such as negative attitudes toward emotional expression and alexithymia (Berger, Levant, McMillan, Kelleher, & Sellers, 2005; O’Neil, 2008). This suggests both an unwillingness and inability to engage with affective experiences ultimately leads to experiential avoidance. Experiential avoidance might therefore be the result of skills deficits and attitudinal factors. These suggestions are largely hypothetical however. Further operationalization and empirical investigation of the two constructs may facilitate understanding of how they might be related and/or represent different stages of an underlying process.

Three other hypotheses were supported in the current study in that a positive association was reported between GRC and PD. Effect sizes were small to moderate in this area. These findings are largely consistent with other studies in this area (Magovcevic & Addis, 2005; Theodore & Lloyd, 2000; Wester et al., 2007). The hypothesized relationship between GRC and EA was also confirmed, as was the proposed relationship between EA and PD. Effect sizes for the latter relationship were large in the current study. The finding that EA is associated with PD is consistent with previous research (Blakey et al., 2016; Nam, 2016; Valiente et al., 2015).

Implications

Given the methodological (cross-sectional) design and novel nature of the current study, results of this research should be regarded tentatively. With this in mind, perhaps the most obvious implication of these findings concerns how we might conceptualize the way in which GRC relates to PD. Assuming that EA and a restricted coping response is critical in the development of psychological distress, this leads to a particular focus in addressing GRC therapeutically. Providing alternatives to restriction and avoidance of unpleasant GRC-related psychological processes (such as unpleasant affect) may be a key treatment focus in attempts to address stress, anxiety, and depressive symptoms.

Supporting men to develop more flexibility in their responses to GRC may represent another therapeutic target. Men face challenges regardless of how masculinity is exhibited on an individual level. For example, Evans and Frank (2003) discuss the issues faced by men working in the feminized profession of nursing. Thus, it may be useful and pragmatic to focus, not only on “how you define yourself as a man,” but also on how to adaptively respond to GRC. In this way, GRC can be an opportunity for positive development and “journeying” with gender roles in a way in which men can resolve GRC (O’Neil & Denke, 2016).

Restricted emotionality yielded the largest mediation effect. This suggests an important focus of therapeutic interventions may involve examining the development of this restriction, and how it leads to experiential avoidance. For example, an individual may restrict due to unhelpful assumptions about emotional expression (e.g., “my negative feelings will get out of control if I focus on them”). Testing out such unhelpful assumptions (e.g., cognitive therapy) or developing an ability to observe emotional experiences (e.g., mindfulness) may assist in the adaptive reevaluation of such assumptions and reduce experiential avoidance.

Recent positive conceptualizations of masculinities (e.g., Englar-Carlson & Kiselica, 2013) may be a useful means by which greater coping flexibility can be achieved by exploring a wider range of options when “doing masculinities.” For instance, “strength” could be illustrated by seeking professional psychological help despite criticism from others (and not just by avoiding negative emotional states). Research in other areas provides examples of men taking a flexible approach in coping with specific social roles (e.g., carer; Spendelow, Adam, & Fairhurst, 2016), mental health problems (e.g., depression; Spendelow, 2015), and resisting expectations of hegemonic masculinities (Way et al., 2014). This research can inform therapeutic interventions for GRC and the development of greater coping flexibility through the provision of examples and means by which men can adopt a more flexible and/or nonhegemonic embodiment of masculinities.

Therapeutic work focused in the above areas compliments professional opinions on how to work with men psychotherapeutically. For instance, a large survey of psychological practitioners revealed that therapeutic work should be characterized by supporting men to forge an individual gender role identity and ensuring that the impact of gender socialization on emotion is addressed appropriately (Mahalik, Good, Tager, Levant, & Mackowiak, 2012).

Limitations and Future Research

The major shortcomings of the current study design are the cross-sectional nature and relatively small sample size on which results are based. Being a novel and exploratory area of investigation, such a study design was appropriate to generate initial empirical data. Furthermore, potentially important sample characteristics (e.g., sexuality) were not measured, and there may be important moderator variables that affect the relationship between GRC, EA, and PD within certain groups of men. For example, there may be a relatively weak relationship between GRC and EA in groups of men where a wide range of (adaptive) coping strategies are practiced. This may be the case among groups of men where self-care is emphasized in their professional lives (e.g., allied health practitioners). It would be important to assess the presence of moderator variables (e.g., social support, age) which may reflect different forms of masculinity and consequently different relationships of GRC to EA and PD.

Another potential limitation involves the wording of the measures used in this study. Some items of the AAQ-II may run counter to certain masculine norms. For example, the item “Worries get in the way of my success” may result in an inaccurate response from a participant who is reluctant to admit he experiences anxiety, or that it affects performance in various domains. Additionally, some participants may have been reluctant to endorse items on the DASS (PD measure) such as “I felt scared without any good reason.” Again, this might violate some strongly held masculine norms (e.g., exhibiting strength or fearlessness).

Future research should employ a longitudinal design to provide a more methodological robust examining of the proposed relationship between GRC, experiential avoidance, and psychological distress presented in the current study. GRC has been reported to be comprised of four separate patterns: success, power, and competition; restrictive emotionality; restrictive affectionate behavior between men; and conflict between work and family relations. Future research employing larger samples and more sophisticated statistical models are needed to simultaneously examine the specific relationship between these factors, experiential avoidance and psychological distress. Such models would ideally take contextual factors into account, which have been recently incorporated into the GRC paradigm (O’Neil & Denke, 2016). Future research could also look at other psychological problems highly relevant to men’s mental health, such as anger and alcohol/drug dependence. Finally, it may be informative to explore the potential role of personality traits (e.g., openness) in the relationships between the variables examined in the current study.