An examination of the relationships among emotion management, interpersonal hassles, and depressive symptomatology

Abstract

This study extended previous research by testing three models predicting depressive symptoms from emotion management (EM) and interpersonal hassles in a sample with a wide range of depression scores. Adults (n = 218) from sources including a depression internet site completed measures of interpersonal hassles, depression, and three aspects of EM (attention to, clarity of, and repair of emotions). Regression analyses supported a model in which lower clarity and repair scores and greater frequency of interpersonal hassles each contributed directly to depression scores. While lower attention to emotions was associated with fewer interpersonal hassles, it did not correlate significantly with depression scores. Moderation analyses did not support a model of EM components acting on depression by buffering effects of interpersonal hassles. A model of depression and hassles predicting EM difficulties was also considered; when both hassles and depression scores were entered to predict EM, only depressive symptoms uniquely predicted clarity and repair.

Key words:

• depression
• emotion management
• emotional intelligence
• interpersonal hassles
• stress
• trait meta‐mood scale

Recent years have seen increasing interest in individual difference constructs related to managing emotions, including the tendency to attend to, understand, and repair one's own emotions (Salovey, Mayer, Goldman, Turvey, & Palfai, 1995). Various conceptualisations have been postulated of these and related constructs within both the EM and emotional intelligence (EI) literature, with emerging evidence that an ability to manage one's emotions is predictive of health outcomes. The ability to process emotions constructively can potentially assist people in perceiving and responding to stress and help promote physical, social, and mental health, including a lessening of depressive symptomatology (Ciarrochi, Chan, Caputi, & Roberts, 2001). The current research aims to examine several possible models of the interrelationships among EM, daily interpersonal hassles, and depressive symptoms.

Emotion management (EM) and depression

The EI literature describes at least three abilities related to emotion processing: attention to, clarity of, and repair of emotions. These abilities are theoretically hierarchically and sequentially organised (e.g., Martinez‐Pons, 1997) in that the constructive management of emotions involves first noticing and attending to how one feels (attention), then discriminating between emotions and clearly identifying and labelling the current emotional state (clarity), and finally taking active steps to regulate negative moods through processes such as focusing on positive or constructive thoughts or activities and reminding oneself of pleasant experiences (repair). Consistent with the theory, Palmer, Gignac, Bates, and Stough (2003) report that the relationship between attention and repair was mediated by clarity in their sample. Theoretically, abilities higher on the hierarchy (particularly repair), would be expected to be most proximal to, and the strongest predictors of, well‐being. Furthermore, deficits in EM abilities might be expected to predict depressive symptoms if one views depression as at least partly involving difficulty in managing and regulating mood states.

Empirically, in studies of university students, individuals reporting greater clarity and repair of emotions do report fewer depressive symptoms, although attention to emotions has not been linked to depressive symptoms (Salovey, Stroud, Woolery, & Epel, 2002; Salovey et al., 1995). Together, these constructs have been found to predict 15% of the variance in depression symptomatology (Extremera & Fernández‐Berrocal, 2006). Similarly, studies of alexithymia, which involves difficulty identifying and expressing emotions (Taylor, 2001), support a link between EM and depression (Honkalampi, Hintikka, Laukkanen, Lehtonen, & Viinamäki, 2000).

Interpersonal hassles and depression

Stress has been theorised as a major antecedent of depression with empirical support for this link (Lazarus & Folkman, 1984; Rabkin & Struening, 1976). Although major life events are linked to depression (Zhang, Shi, & Wang, 2005), daily hassles also predict greater depression (Ciarrochi, Deane, & Anderson, 2002; Ravindran, Griffiths, Waddell, & Anisman, 1995), and may be better predictors than major events. For example, Weinberger, Hiner, and Tierney (1987) found daily stressors correlated with depression more strongly (r = 0.53) than did major events (r = 0.29).

Daily hassles are common, minor disruptive or frustrating events that occur in everyday life (Maybery & Graham, 2001). Frequency of hassles, rather than their intensity, correlates with poor mental health (Kanner, Coyne, Schaefer, & Lazarus, 1981), suggesting that it is the repeated experience of stressors that cumulatively creates mood or emotional disturbances.

Hassles can be categorised into non‐interpersonal hassles such as traffic jams or forgetting information, and interpersonal hassles such as disagreements or conflict (Maybery, 2003). Consistent with an interpersonal model of depression, interpersonal hassles may be expected to be more distressing than non‐interpersonal hassles. In support of this proposition, when examining interpersonal hassles, work demands, financial problems, and transport hassles as predictors, Bolger, De Longis, Kessler, and Schilling (1989) found interpersonal hassles contributed more than three quarters of the variance in negative mood. The authors' recommendation to reorient stress research to interpersonal hassles has been supported by studies showing that focusing on interpersonal hassles improves predictions of distress and well‐being (Maybery, 2003; Maybery & Graham, 2001) and that many depressed individuals report experiencing interpersonal rejection (Marcus & Nardone, 1992) and dysfunctional relationships (Salovey et al., 2002). However, no research to date has examined the direct relationship between interpersonal hassles and depression.

EM and interpersonal interactions and hassles

Various studies have suggested that greater ability to manage emotions is associated with fewer interpersonal hassles and difficult interactions. In a recent meta‐analysis (Schutte, Malouff, Thorsteinsson, Bhullar, & Rooke, 2007), greater ability (objectively measured) EI was associated with better social skills, relationships, marital satisfaction (Schutte et al., 2001), and relationship quality (Ciarrochi, Chan, & Caputi, 2000). In addition, self‐perceived abilities to be clear about and manage emotions have been linked to managing disagreements better (Fitness & Curtis, 2005) and greater interpersonal satisfaction (Salovey et al., 2002). Accordingly, Ciarrochi et al. (2002) suggested that differences in EM abilities may contribute to understanding the link between relationship difficulties and poor mental health.

Models of the relationships among EM abilities, hassles, and depression

While interrelationships among EM ability, experience of hassles, and depressive symptoms have been demonstrated, the specific mechanisms explaining these associations are not clear. Several possible models of these relationships were considered in the current study. The first model suggests that the relationship between interpersonal stress and depression changes as a function of emotional competencies. This model suggests that an ability to understand and manage emotions may buffer the effects of hassles, so that individuals better able to manage their emotions are less prone to become depressed when faced with frequent hassles (Ciarrochi et al., 2002). On the basis of the hierarchical model of EM, repair and, to a lesser extent, clarity, would be expected to moderate the relationship between hassles and depression. Theoretically, when faced with frequent hassles that could generate stress and negative mood states, those who can regulate their emotions through maintaining a more positive and constructive outlook, rather than ruminating (Salovey et al., 1995) or suppressing thoughts (Rude & McCarthy, 2003), should more successfully avoid depressive responses. Two studies were identified that examined a moderation model of this sort. One study (Ciarrochi et al., 2002) found that the relationship between hassles and suicidal ideation was lower for individuals who reported better ability to manage others' emotions (a social skills‐related measure). However, this moderation effect was not found when the outcome measure was depression. These findings suggest that self‐related EM may not buffer effects of hassles (although managing own emotions was not tested). Day, Therrien, and Carrol's (2005) findings also did not support moderation using another EI measure. These two studies examining moderation effects included only university samples, which would limit the level of depression in the sample. Therefore, the current study examined a broader sample targeted to include greater variation in participants and depression levels.

A second proposed model was an independent effects model in which EM and hassles each are associated independently with greater depressive symptoms. Rather than EM assisting individuals in managing hassles, it is possible that EM operates to generally maintain positive moods and that regardless of EM skills, frequent interpersonal hassles lead to more depressive symptoms. If this model is supported, EM and hassles will each predict unique variance in depression. This model is consistent with the Ciarrochi et al. (2002) and Day et al. (2005) findings that hassles correlated with perceived EI and depression or well‐being.

A third possible model is based on the proposition that individuals with less ability to be clear about and repair emotions are more likely to have or perceive interpersonal hassles. This model posits that EM deficits lead to less satisfactory interpersonal interactions and more perceived interpersonal hassles, which in turn enhance depressive symptoms. Consistent with this mediation model, individuals with lower perceived EI scores have been found to be less agreeable (Schutte et al., 2001) and less prone to prosocial behaviours (Hodgson & Wertheim, 2007), which could result in more actual or perceived negative interactions.

Thus far, we have viewed depressive symptoms as outcome variables; however, it is possible that depression makes affect regulation more difficult. For example, emotion repair processes involve focusing on pleasures in life, and positive, optimistic thoughts, which are likely to be more difficult for individuals with depression. Furthermore, there is evidence that depression generates interpersonal difficulties and stress (Hammen, 2006), which could make managing emotions more difficult. While the current study is cross‐sectional and cannot determine direction of effect, we also examined the combined contributions of interpersonal hassles and depressive symptoms in predicting repair and clarity of emotions, and whether interpersonal hassles mediate the relationship between depressive symptoms and EM.

Measuring EM

The EI literature has involved development of both ability and perceived EI measures. Several authors (Lopes, Salovey, & Straus, 2003; Roberts, Zeidner, & Matthews, 2001) have suggested that ability measures assess optimal performance (e.g., perceptual skills and knowledge about emotions) rather than typical performance. In the current study, typical performance (what individuals do in practice) was considered more relevant than optimal performance. Therefore, a perceived EI measure was used, since it assesses perceptions of how well in daily life individuals attend to, are clear about, and repair their emotions.

AIMS

In summary, the main aim of this research was to extend past findings by examining the relationships among interpersonal stress; perceived ability to understand and manage emotions; and depression in a sample including a wide range of depression levels. The study assessed whether three aspects of EM—attention, clarity, and repair of emotions—as well as hassles, were associated with depressive symptoms. We aimed, in a sample with a high proportion of depressed individuals, to examine which of three possible models predicting depressive symptoms from EM and interpersonal hassles would be best supported. Given previous findings that interpersonal hassles are more related to negative mood than other hassles, the role of interpersonal hassles (e.g., disagreements, conflict) was the focus. Finally, we considered a possible model of depressive symptoms and hassles as predictors of EM.

METHOD

Participants

Participants were 218 adults aged between 18 and and 69 years (M = 33.02; standard deviation (SD) = 12.64), comprising 149 females (68%) and 69 males (32%). In the sample, 15 (7%) were educated to Year 10, 81 (38%) to Year 12, 75 (35%) had an undergraduate qualification, and 43 (20%) a postgraduate qualification. Eighty‐three (39%) self‐reported a mental illness.

Materials

A questionnaire assessed age, gender, education, and past and present mental health problems (e.g., ‘Have you had any mental health problems in the past?’ followed by a request to describe it), and the following scales.

Centre for Epidemiological Studies Depression Scale (CES‐D; Radloff, 1977)

The CES‐D assesses frequency of symptoms associated with depression, rating feelings and behaviours in the previous week from none of the time or rarely (less than 1 day) (1) to most or all of the time (5–7 days) (4). High internal consistency among items (Cronbach's α = 0.84 to −0.90), and satisfactory test–retest reliability and correlations with clinician ratings have been reported (Radloff, 1977). The present study Cronbach's α = 0.95. The CES‐D has been consistent with clinical structured interviews from the DSM‐IV in identifying cases of major depressive disorder (Caracciolo & Giaquinto, 2002).

Interpersonal Hassles Scale (IPHS, Maybery, 2003)

The IPHS assesses frequency of interpersonal problems experienced during the previous month, rated from did not happen (0) to 10 or more times (5) in the last month. Higher scores indicate more hassles. The IPHS has demonstrated construct validity with negative affect (r = 0.53, p < .05) and another hassles scale (r = 0.60, p < .01; Maybery & Graham, 2001). Previous literature reports high internal consistency (α = 0.89–0.98; Maybery, 2003). The current sample Cronbach's α = 0.79.

Trait Meta‐mood Scale (TMMS, Salovey et al., 1995)

The TMMS assesses how individuals perceive their abilities on three subscales: attention to emotions, clarity of emotions, and repair of emotions. The attention subscale examines how much attention individuals devote to emotions (e.g., ‘I pay a lot of attention to my emotions’), the clarity subscale assesses perceptions of correctly identifying and understanding emotions (e.g., ‘I am usually very clear about my feelings’), and the repair subscale assesses perceived ability to repair negative emotions and maintain positive emotions (e.g., ‘I try to think good thoughts no matter how badly I feel’). Individuals rate items from strongly disagree (1) to strongly agree (5). Attention, clarity, and repair total scores are calculated. Higher scores indicate higher self‐perceived abilities. The TMMS has shown adequate internal consistency, construct validity, and factorial reliability (Palmer et al., 2003; Salovey et al., 1995), and scores have correlated well with significant other ratings in an Australian community sample (Hodgson & Wertheim, 2007). Current study subscale Cronbach's alphas ranged from 0.82 to 0.90.

Procedure

Following ethics committee approval, participants responded to advertisements or researcher invitations to complete a study on EI, depression, and interpersonal hassles for which they could obtain a personal EI report. Twenty‐six percent (n = 56) responded to advertisements on a depression resource website (http://www.depressioNet.org.au), 11% (n = 22) answered advertisements in a student newsletter, local newspaper, or mental health organisation newsletter, and 65% (n = 133) responded to word‐of‐mouth or forwarded email from other participants. Interested individuals were sent information indicating the study explored relationships among EI, depression, relationship styles, and interpersonal hassles. Participants were told that requesting a SpeedSurvey online questionnaire link would indicate informed consent. Participants answered the background questionnaire followed by the CES‐D, IPHS, and TMMS (a further online questionnaire was completed later, to be reported in another article). Participants completing the full set of questionnaires could receive a summary report of their EI‐related scores.

RESULTS

Preliminary analyses

One case with >5% missing data was dropped; for lower percentage missing data (20 cases), mean score substitution was used. Three univariate outliers were replaced with missing data. Two cases with multivariate outliers (Mahalanobis' distances critical χ² = 18.47, p < .001) were excluded.

Scores on the CES‐D ranged from 0 to 54 (M = 18.7; SD = 14.5). Of the sample, 49% (n = 106) reached Radloff's (1977) arbitrary cut‐off score of 16 or more for possible depression, and 29% (n = 62) met the more stringent criterion for clinical depression of 27 proposed by Schulberg et al. (1985) and Costello and Devins (1989) for DSM‐III diagnosis. Eighty‐three (39%) self‐reported current mental health problems.

Relationships among variables

Table-1 displays correlations among variables. Interpersonal hassles and CES‐D scores were positively skewed; square root transformations (IPHS^SQ, CES‐D^SQ) were applied for correlations and regressions. Greater depressive symptom scores were associated with more interpersonal hassles reported, and lower scores on ability to understand and repair emotions, but not with attention to emotions. Reports of greater interpersonal hassles were associated with lower TMMS scores (attention to, clarity of, and repair of emotions). The TMMS subscales significantly correlated with each other (Table-1). Older participants reported significantly greater clarity of emotions. Females reported more attention to emotions and more depressive symptoms than males. Education level was correlated significantly with CES‐D^SQ (r = −0.22, p = .001) and clarity (r = −0.27, p < .001), but no other variables.

Table 1 Descriptive statistics and Pearson's correlation coefficients among variables

Measure	1	2	3	4	5	6	7	8
1. Age	–
2. Gender	0.03	–
3. Education	0.22**	0.02	–
4. CES‐D^SQ	−0.10	−0.20**	−0.22**	–
5. IPHS^SQ	−0.09	−0.11	−0.09	0.23**	–
6. Attention	0.00	−0.13	0.12	−0.10	−0.16*	–
7. Clarity	0.14*	0.09	0.27***	−0.57***	−0.18*	0.36***	–
8. Repair	0.01	0.06	0.09	−0.68***	−0.16*	0.14*	0.56***	–
Mean	33.02	–	–	18.65	13.88	50.57	38.82	21.48
SD	12.64	–	–	14.52	9.21	7.40	9.08	5.76

n = 213. CES‐D^SQ, Centre for Epidemiological Studies Depression Scale after square root transformation to address positive skewness; IPHS^SQ, Interpersonal Hassles Scale after square root transformation to address positive skewness; SD, standard deviation.

* p < .05,

** p < .01,

*** p < .001.

Examination of the direct, independent effects, and mediation models

A hierarchical regression analysis examined contributions of interpersonal hassles and the three EM subscales in predicting depression scores (CES‐D^SQ) after controlling for demographic variables (see Table-2). At Step 1, age, education, and gender significantly predicted CES‐D^SQ, F(3, 206) = 6.67, R² = 0.088, Adj R² = 0.075, p < .001. At Step 2, the three TMMS subscales accounted for an additional 45.8% of the variance of depression scores, FΔ (3, 203) = 68.20, R²Δ = 0.458, Adj R² = 0.53, p = .002, with clarity and repair subscales each explaining significant variance. At Step 3, IPHS^SQ explained significant additional variance in depression scores, FΔ (1, 202) = 4.80, R²Δ = 0.011, p = .03, with the final model being significant, F(7, 202) = 36.22, R² = 0.56, Adj R² = 0.54, p < .001. The regression weights for repair and clarity were not significantly different in Step 2 versus Step 3, suggesting no mediation effect (repair β = −0.53 vs −0.51; clarity β = −0.26 vs −0.25).

Table 2 Final step of hierarchical regression analysis of interpersonal hassles and perceived emotional intelligence on depression

Regression step	Variables	B	SE B	β	t
Step 1	Age	−0.01	0.01	−0.06	−0.81
	Sex	−0.67	0.24	−0.18	−2.77**
	Education	−0.40	0.13	−0.21	−3.06**
Step 2	Age	−0.00	0.01	−0.03	−0.65
	Sex	−0.48	0.17	−0.13	−2.77**
	Education	−0.17	0.10	−0.09	−1.77
	Attention	0.01	0.01	0.05	1.04
	Clarity	−0.05	0.01	−0.26	−4.01***
	Repair	−0.15	0.02	−0.53	−9.09**
Step 3	Age	0.00	0.00	−0.03	−0.51
	Sex	−0.44	0.17	−0.22	−2.53**
	Education	−0.16	0.10	−0.09	−1.71
	Attention	0.02	0.01	0.07	1.32
	Clarity	−0.05	0.01	−0.25	−3.98**
	Repair	−0.15	0.02	−0.51	−8.98***
	IPHS^SQ	0.15	0.07	0.11	2.19*
	Constant	8.28	0.77

N = 214. Dependent variable = Centre for Epidemiological Studies Depression Scale after square root transformation; IPHS^SQ = Interpersonal Hassles Scale after square root transformation. *p < .05, **p < .01, ***p < .001. Step 1 F(3, 206) = 6.67, R² = 0.088, p < .001; Step 2 FΔ (3, 203) = 68.20, R²Δ = 0.458, Adj R² = 0.53, p = .002; Step 3 FΔ (1, 202) = 4.80, R²Δ = 0.011, p = .03.

Moderation analyses

Hierarchical regression analyses were conducted to investigate whether the relationship between interpersonal hassles and depression was stronger for people with lower scores on TMMS subscales. Interaction terms based on z‐scores were constructed. After entering age, gender, education, IPHS^SQ, and attention, the attention–IPHS^SQ interaction term did not contribute further significant variance (p > .05). The parallel regressions for clarity and repair also yielded no significant interaction effects.

Depressive symptoms and hassles predicting EM

In two standard regressions, age, gender, and education were controlled for in Step 1, and IPHS^SQ was entered in Step 2, followed by CES‐D^SQ predicting repair and then clarity. At Step 1 no demographic variables predicted repair (p > .17). At Step 2 IPHS^SQ predicted repair, FΔ (1, 206) = 4.12, p = .04, R²Δ = 0.019, but at Step 3 CES‐D repair was the only significant predictor, FΔ (1, 205) = 173.13, p < .0005, R²Δ = 0.443, CES‐D^SQt = −13.16, p < .0005, IPHS^SQt = −0.02, p = .98, with CES‐D^SQ fully mediating the relationship. Turning to clarity of emotions, greater education was the only demographic to significantly predict clarity (p < .0005). At Step 2, IPHS^SQ only had a tendency towards significance (p = .08, R²Δ = 0.014), with CES‐D^SQ again significant at Step 3, FΔ (1-205) = 82.57, p < .0005, R²Δ = 0.258, CES‐D^SQt = −9.09, p < .0005, IPHS^SQt = −0.22, p = .82. As attention did not correlate with CES‐D, it was not tested.

DISCUSSION

The present research examined the relationships among interpersonal hassles; the EM processes of attention, clarity, and repair; and depression, in a mixed community sample of adults with a wide range of depression scores. Three possible models predicting depression and a secondary model predicting EM were examined.

Consistent with theories in which interpersonal hassles are viewed as stressful, negative experiences which can decrease mood, individuals experiencing more interpersonal hassles were more likely to report depressive symptoms. This study was one of the few (Bolger et al., 1989) to examine a discrete measure of interpersonal hassles.

Findings supported the idea that greater perceived abilities to identify and distinguish between emotional states and to repair and regulate negative mood states are associated with lower depression levels. The findings of repair being the strongest, and attention the least strong, predictor are consistent with the hierarchical, sequential EM model in which attention to emotions is a foundation process, with clarity about and ability to actually repair negative emotions being progressively more proximal to, and important in, determining mental health outcomes. The findings support research (Salovey et al., 1995, 2002) suggesting that individuals who better understand emotions may experience a greater rebound from negative mood and think more clearly under acute stress situations (Gohm & Clore, 2002). Those with higher skills in repair of emotions appear to be able to interrupt negative moods, prolong positive emotional states, and maintain an optimistic outlook (Salovey et al., 1995), which appears to foster lower depression.

Of the three models examining the possible effect of interpersonal hassles and perceived EI on depression, the only model supported was an independent effects model. Emotion repair, clarity of emotions, and frequency of interpersonal hassles, each explained unique variance in depression scores. In contrast, although better emotion processing was associated with fewer interpersonal hassles, there was no evidence that the role EM plays in depressive symptoms is as a buffer in the relationship between interpersonal hassles and depression. This finding is consistent with the two previous studies that examined this model (Ciarrochi et al., 2002; Day et al., 2005). According to the findings, poor EM also does not appear to increase depression through promoting more interpersonal hassles. Instead, lesser ability to repair emotions and, to a lesser extent, more hassles may independently foster depression.

The sorts of EM skills assessed in this study support the importance of being able to identify and label emotions and mood‐related states, as a foundation for regulating them by focusing on positive experiences, retaining positive cognitive frames for the future, and focusing on constructive thoughts and activities. These skills are associated with therapeutic approaches such as cognitive behaviour therapy, EM modules in dialectical behaviour therapy, or self‐regulation approaches based on learned helplessness and learned optimism theories of depression. Our findings point to the possibility of also teaching these skills to potentially assist in preventing depressive symptoms.

The three main models tested assumed a causal effect of emotion repair and hassles on depression; however, given the correlational nature of the study, depression could instead be a causal factor for hassles and poor emotion regulation. When EM abilities were examined as outcome variables, depression significantly predicted both reported clarity and repair abilities, with no evidence that interpersonal hassles mediated these effects. Future prospective and intervention studies are needed to further assess direction of effects, examining whether teaching depressed individuals better skills for managing their emotions reduces depression levels and/or whether as depression lifts (e.g., through medication), perceived clarity about feelings and mood regulation spontaneously improve.

Methodological considerations and future directions

The present study's strengths included being one of the first to examine the direct relationship between interpersonal hassles and depression. Second, the sample was drawn from a variety of sources enabling representation across the full spectrum of depression scores on the CES‐D. The use of a sample where more than one quarter had CES‐D scores meeting a stringent threshold for likely clinical depression further supports that it is not only to subclinical or university groups that the findings apply. Indeed, the three EM scales together explained an additional 46% of the variance in depression scores after controlling for demographic variables, which is much higher than previously reported in a university sample (e.g., 15%, Extremera & Fernández‐Berrocal, 2006).

It should also be noted that the measures in this study were self‐reports, which can be influenced by response sets. Therefore, future research should broaden methods to include standardised diagnostic procedures, interviews, reports of significant others, and constructs assessed in ability EI measures. Other factors, such as coping skills (Zhang et al., 2005), would also be worth examining as moderator variables in the relationship between hassles and depression.

In conclusion, this study supported the conceptualisation of components of managing emotions being related to depression and interpersonal hassles. Results provide support for a model in which three key factors, including clarity about one's emotions, ability to repair one's emotions, and the experience of fewer interpersonal hassles, each contribute unique variance to reported depression levels. An alternative model in which depressive symptoms interfere with the ability to clearly identify and repair emotions was also considered, suggesting that prospective research should clarify direction of effects. The study has extended prior research by demonstrating these relationships in broader sample than previously examined and by further clarifying the specific relationships among these variables.