Association Between Coping With Anger and Feelings of Depression Among Youths

Abstract

Objectives. I examined the association, among youths, between coping behavior when angry and depression.

Methods. Data were drawn from the Health Behavior in School-Aged Children in the United States survey (n=9938). Factor analyses and multivariate logistic regression analyses were used to determine the association between self-reported coping behavior when angry and depression. Gender-specific models were run.

Results. Factor analysis of 11 coping behaviors indicated a 4-factor solution: substance use, physical activity, emotional coping behavior, and aggressive behavior. Substance use, emotional coping, and aggressive behavior coping were associated with increased likelihood of depression, whereas physical activity was associated with decreased likelihood of depression. Male youths were more likely to engage in physical activity and were less likely to feel depressed.

Conclusions. These data provide preliminary evidence of a link between specific coping behavior when angry and the likelihood of depression among youths. Whether these associations may be useful in identifying youths at risk for depression cannot be determined from these data alone but may be an important area for future study.

Several recent studies have documented an association between adults’ coping styles and risks of depression and depressive symptoms.1^,2 Data suggest that emotion-focused coping is associated with increased odds of depression and that task-oriented coping behavior may be associated with a lower likelihood of depression.

The mechanism of action between coping styles and the risk of depression is not entirely clear, but several hypotheses have been put forward. It could be that specific modes of interpreting positive and negative events are differentially depressogenic and may therefore result in different lifestyles, which accommodate either the positive or the negative perspective.3 For instance, it may be that emotion-focused coping with regard to loss ultimately increases the likelihood of depressive feelings through social isolation because the individual is less emotionally equipped to reach out to old contacts or make new friends, and this often leads to social withdrawal, isolation, and depression.

It is also conceivable that there are neuroendocrine or neurobiological substrates associated with emotion-focused coping that increase the risk of depression through neuro-chemical changes or pathways.4 It is further possible that depression leads to the development of an emotion-oriented style of coping as a result of slowed thinking, leads to having a negative worldview, or leads to limits in cognitive or physical functioning associated with the depression itself.

Alternatively, it may be that a third outside variable, such as a personality factor, is associated with the co-occurrence of depression and specific coping behaviors. For instance, neuroticism may be associated with an increased likelihood of smoking cigarettes and depressive symptoms when distressed.5^,6

Data from clinical studies on adult inpatients and outpatients suggest that there are differences in coping behaviors between patients with and without depression.7^–9 Also, Jorm et al.10 examined the relation between depression and a wide variety of coping behaviors, finding that there are significant differences in types of behavior depending on the level of severity of depression. Previous studies have also shown that coping behaviors are associated with physical outcomes among those with medical illnesses as well,11^,12 yet these analyses have not specifically been extended to mental disorders.

There is a long-held belief that anger and depression are intricately linked and that anger that is excessive, unexpressed, or “turned inward” leads to depression.13^,14 Thus, an individual’s method of coping with anger may be related to the likelihood of that individual experiencing depression. If a person engages in behaviors that are linked to effectively managing and discharging angry feelings and increasing healthy behavior, that person’s likelihood of developing depression may be different (lower) from that of someone who copes with depression by engaging in behaviors that are harmful to that person’s health, self-destructive in terms of social relationships or economic well-being, or associated with an increased risk of depression.

Although previous investigations among adults in clinical and community samples suggest that coping styles may be related to depressive symptoms, several pertinent areas have remained relatively neglected. First, it is not known whether previous findings are generalizable to youths in the community. Second, previous data are from clinical samples; therefore, it is not possible to determine whether coping strategies are associated with depression among youths in the community or whether coping styles are associated with selection into treatment. Third, despite gender differences in coping behaviors and risks of depression, no study has examined the relation between gender, coping behavior when angry, and the likelihood of depression.

Against this background, the goal of the current study was to begin to fill this gap by examining the association between coping behavior when angry and the likelihood of depression among youths. First, I examined the relation between coping behavior when angry and depression. Second, I examined the association between coping behavior when angry and gender. Third, I determined the association between coping behavior when angry and feelings of depression by gender.

On the basis of previous findings, I hypothesized that activity-oriented coping behavior when angry would be associated with a significantly lower likelihood of feelings of depression among youths compared with that associated with emotion-oriented coping behavior when angry. I also predicted that activity-oriented coping would be more common among male youths than among female youths.

METHODS

Study Population

The National Institute of Child Health and Human Development supported a nationally representative survey of US youths in grades 6 through 10 during spring 1998. The survey, titled the Health Behavior of School-Aged Children (HBSC),15 was part of a collaborative, cross-national research project involving 30 countries and coordinated by the World Health Organization. The US sampling universe consisted of all public, Catholic, and other private school students in grades 6 through 10, or their equivalent, excluding schools with enrollments of fewer than 14 students.

The sample was a stratified 2-stage cluster of classes. The sample selection was stratified by racial/ethnic status to provide an oversample of Black and Hispanic students. The sample was also stratified by geographic region and counties’ metropolitan statistical area status (largest urban areas/not largest urban areas), with probability proportional to total enrollment in eligible grades of the primary sampling units. Sample size was calculated in order to provide adequate numbers for making comparisons and producing results for all US students in grades 6 through 10.

Sampling

The sampling plan was designed to support 2 overlapping studies with different sampling requirements. The “base study,” or HBSC study, employed methods that produced a self-weighting sample of students at each of 3 target age levels (11, 13, and 15 years).

The base study sample was designed to be equivalent to ±3% at a 95% confidence level (CI), which was established by the international HBSC commission. The full US study was designed to meet the additional goal of estimating African American and Hispanic characteristics within 5% at a 90% CI. These results are based on the full US study.

Response Rates

Because of the lack of state and local infrastructure to support the HBSC in the United States, a relatively low school participation rate was anticipated. In order to achieve the requisite number of participants by subpopulation, a low, conservative participation rate was assumed. There were 664 schools selected to participate in the HBSC survey. Of those 664, 386 schools agreed to participate, yielding a school participation rate of 58%. Within the 386 participating schools, 20 533 students were eligible for participation, and 17 000 participated, yielding a student response rate of 83%. These participation rates were sufficient to achieve the targeted precision levels and confidence intervals for the subpopulations of interest.

The school-based sample design used 1 class period for completion of the questionnaire. Responding students in sampled classes were excluded if they were out of the target range for grade or if their age was outside the 99th percentile for grade (n = 440 students), or if either grade or age were unknown (n = 39 students), yielding an analytic sample of 15 686 students.

Measures

Measures were obtained from a self-report questionnaire containing 102 questions about health behavior and relevant demographic variables. Items were based on both theoretical hypotheses related to the social context of adolescents and measurements that had been validated in other studies or previous World Health Organization–HBSC surveys. Measures were pretested.

Behavior When Angry

Respondents were asked a series of self-report questions regarding their behavior when angry. Respondents were asked, “What do you usually do when you get angry?” and then specific activities were queried, including, ”Find someone to talk to, drink alcohol, take drugs, stuff myself with food, listen to music, get into a physical fight, get into a verbal argument, go ride a bike, think about hurting myself on purpose, smoke a cigarette, exercise, pray, go for a walk, and cry.” Answers were yes or no. “Stuff myself with food,” “think about hurting myself,” and “cry” were omitted from the analyses in the current study because they may be symptoms of depression.

Feelings of Depression

Depression was assessed with a self-report item: “During the past 12 months, did you ever feel sad, blue, down, or depressed almost every day for TWO WEEKS OR MORE IN A ROW?” Answers were yes or no. Then we took those who responded affirmatively to this question, and also, to define the participants with feelings of depression, endorsed at least 4 of the following 10 depression symptoms: (1) irritable when depressed, (2) lost interest when depressed, (3) gained weight when depressed, (4) lost weight when depressed, (5) couldn’t concentrate when depressed, (6) couldn’t sleep when depressed, (7) slept a lot when depressed, (8) rotten person when depressed, (9) thought of hurting self when depressed, (10) thought of death when depressed. Participants who endorsed the depression self-report item, in addition to endorsing at least 4 out of 10 depression symptoms, were considered to have feelings of depression, for the purposes of this study.

Analytic Strategy

First, all 11 coping behaviors were entered into an exploratory factor analysis with the use of principal components analysis with Varimax rotation. Next, the association between each of these 4 factors and the likelihood of depression was examined with multivariate logistic regression analyses to produce odds ratios with 95% CIs.

Next, the same method was used to examine the relation between gender and specific coping behaviors. Third, independence-based F tests were used to determine the relation between gender, depressive feelings, and self-reported coping behaviors when angry.

Next, multivariate logistic regression analyses were used to determine the association between each coping factor and the likelihood of depression in the past year, after adjustment for differences in gender, race, parental education, and having a single parent. The sample was stratified by each grade assessed and among the whole sample, with adjustment for age.

The analyses were stratified in order to make apparent any changes in relation between coping behavior when angry, gender, and depressive symptoms between pre- and postpubertal youths. Additional gender-specific analyses were also run. SPSS for Windows (9.0) was used for all statistical analyses.16

RESULTS

Whole-Sample Results of Factor Analysis

An exploratory factor analysis with an eigenvalues-greater-than-1.0 criterion for extraction (Kaiser rule) indicated a 4-factor solution: substance use factors (alcohol use, drug use, cigarette smoking); physical activity factors (exercising, bike riding, taking a walk); emotional coping factors (talking to someone, praying, listening to music); and aggressive behavior factors (fighting, arguing) (Table 1 ▶).

TABLE 1—

Factor Analysis of Coping Behaviors When Angry Among Youths: United States, 1988

Factor	Eigenvalue
Substance use
Alcohol	1.1
Drugs	1.0
Cigarettes	0.4
Physical activity
Exercise	2.4
Biking	2.0
Walking	0.5
Emotional coping
Talking to someone	0.9
Praying	0.6
Listening to music	0.8
Aggressive behavior
Fighting with someone	0.7
Arguing	0.6

Factors 1, 3, and 4 were associated with significantly increased likelihood of depression (Table 2 ▶), whereas factor 2 was associated with a significantly decreased likelihood of depression. Specifically, factor 1 (substance use) (OR = 1.538; 95% CI = 1.536, 1.54]), factor 3 (emotional coping) (OR = 1.557; 95% CI = 1.536, 1.559), and factor 4 (aggressive-behavior coping) (OR = 1.579; 95% CI = 1.577, 1.581) were associated with increased depression, whereas factor 2 was associated with decreased likelihood of depression (OR = 0.928; 95% CI = 0.927, 0.929).

TABLE 2—

Association Between Coping Factors and Depression and Gender Among Youths: United States, 1988

	Factor 1: Substance Use, OR (95% CI)	Factor 2: Physical Activity, OR (95% CI)	Factor 3: Emotional Coping, OR (95% CI)	Factor 4: Aggressive Coping, OR (95% CI)
Association with depression	1.62 (1.618, 1.623)	0.985 (0.984, 0.987)	1.308 (1.306, 1.31)	1.486 (1.484, 1.489)
Association with female gender	1.039 (1.038, 1.04)	0.842 (0.841, 0.843)	2.131(2.128, 2.134)	0.862 (0.861, 0.863)

Note. OR = odds ratio; CI = confidence interval.

Association Between Gender and Feelings of Depression

Male youths were less likely to report feeling depressed in the past year than female youths (25.9% vs 42.0%, x = 279.6, df = 1, P < .0001) (data not shown). Female youths were more likely to report feelings of depression (OR = 2.0; 95% CI = 1.8, 2.2), even after adjustment for differences in sociodemographic characteristics.

Association Between Gender and Behavior When Angry

There were several significant differences between male and female youths in behavior when angry (Table 3 ▶). Female youths were more likely to pray, go for a walk, talk to someone, argue, and listen to music. Male youths, in contrast, were more likely to fight physically, bike, exercise, drink alcohol, and use drugs. There was no gender difference in cigarette smoking.

TABLE 3—

Association Between Gender and Behavior When Angry Among Youths: United States, 1988

Behavior When Angry	Male Youths, % (n = 4760)	Female Youths, % (n = 5178)	χ2, df = 1 (P)	AOR (95% CI)
Drink alcohol	10.1	8.0	13.2 (< .0001)	0.8 (0.6, 0.9)*
Take drugs	8.1	5.6	21.7 (< .0001)	0.7 (0.6, 0.8)*
Smoke cigarettes	14.7	13.9	1.1 (.3)	0.95 (0.83, 1.09)
Bike	35.0	27.3	65.3 (< .0001)	0.7 (0.66, 0.8)*
Take walk	42.9	54.6	129.0 (< .0001)	1.8 (1.6, 1.9)*
Exercise	39.1	31.4	61.7 (< .0001)	0.76 (0.7, 0.8)*
Talk to someone	37.5	61.1	529.9 (< .0001)	2.8 (2.5, 3.1)*
Listen to music	76.0	87.7	223.8 (< .0001)	2.2 (2.0, 2.5)*
Pray	32.8	49.9	284.8 (< .0001)	2.1 (1.9, 2.3)*
Fight physically	22.1	10.3	246.9 (< .0001)	0.4 (0.35, 0.5)*
Argue	46.5	49.6	9.2 (.002)	1.2 (1.1, 1.3)*

Note. AOR = adjusted odds ratio, adjusted for parental educational and single parent status; df = degrees of freedom; CI = confidence interval.

* P < .05.

Adjusted Association Between Coping Behaviors When Angry and Feelings of Depression

When specific coping behaviors and depression were compared, after adjustment for gender, parental education, and single-parent status, alcohol use, drug use, listening to music, fighting, arguing, smoking, praying, and taking a walk were associated with increased likelihood of depression (Table 4 ▶). In contrast, bike riding was associated with decreased likelihood of depression.

TABLE 4—

Association Between Depression and Behavior When Angry Among Youths: United States, 1988

Behavior When Angry	No Depression Past Year, % (n = 6449)	Depression Past Year, % (n = 3370)	χ21(P)	AOR (95% CI)
Drink alcohol	4.4	19.2	425.1 (< .0001)	5.1 (4.2, 6.1)*
Take drugs	3.3	14.5	276.4 (< .0001)	4.7 (3.8, 5.8)*
Smoke cigarettes	9.5	29.5	483.2 (< .0001)	4.1 (3.5, 4.8)*
Bike	34.5	27.6	78.7 (< .0001)	0.7 (0.6, 0.8)*
Take walk	46.7	56.8	62.8 (< .0001)	1.3 (1.2, 1.5) *
Exercise	34.6	37.6	1.5 (.2)	1.0 (0.9, 1.2)
Talk to someone	49.1	51.6	20.6 (.001)	0.97 (0.86, 1.1)
Listen to music	79.3	87.9	92.2 (< .0001)	1.7 (1.4, 2.0)*
Pray	37.8	46.7	62.7 (< .0001)	1.2 (1.1, 1.4)*
Fight physically	11.5	22.0	167.8, (< .0001)	2.9 (2.5, 3.4)*
Argue	43.1	67.7	457.5 (< .0001)	2.8 (2.5, 3.1)*

Note. AOR = adjusted odds ratio (for gender, parental educational, and single-parent status); CI = confidence interval.

* P < .05.

After additional adjustment for differences in sociodemographic characteristics (gender, age, race, single parent, parents’ education) in the final model, drinking alcohol (adjusted odds ratio [AOR] = 1.72; 95% CI = 1.34, 2.2), arguing (AOR = 1.32; 95% CI = 1.16, 1.51), and smoking cigarettes (AOR = 1.76; 95% CI = 1.45, 2.14) when angry were associated with increased likelihood of feelings of depression among the whole sample of youths.

In contrast, riding a bike (AOR=0.72; 95% CI=0.62, 0.84) and talking to someone (AOR= 0.84; 95% CI = 0.73, 0.95) were associated with a decreased prevalence of feelings of depression. Minority racial status (OR = 1.33; 95% CI = 1.17, 1.52) and having a single parent (OR = 1.24; 95% CI = 1.17, 1.52) were also independently associated with increased odds of feelings of depression. Higher maternal (OR = 0.81; 95% CI = 0.70, 0.93) and paternal education (OR = 0.86; 95% CI = 0.75, 0.98) were independently associated with decreased likelihood of depressive feelings.

DISCUSSION

These data suggest that specific coping behaviors (e.g., smoking, arguing, drinking alcohol) when angry are associated with feelings of depression among youths. Specifically, these data suggest that after adjustment for differences in sociodemographic differences and potentially confounding social factors, bike riding is associated with decreased frequency of depressed feelings among youths. Factor analysis provided empirical data to show the clustering of coping behaviors into 4 groups: substance abuse, physical activity, emotional coping, aggressive behavior coping. Our results also suggest that there are gender differences in these behaviors.

Whereas previous studies suggest that many of these behaviors (e.g., fighting, cigarette smoking) can be characteristic of depression in youths,17^–19 previous studies have not examined these behaviors in relation to anger and the likelihood of feelings of depression in a large, nonclinical sample. Building on the long-studied relation between anger and depression, these data provide initial evidence to suggest that the gender difference in depression may be contributed to, at least in some part, by differences in behaviors used to cope with feelings of anger.

The mechanism of the association between activity-oriented behavior and decreased likelihood of feelings of depression is not known. It may be that behaviors such as riding a bicycle decrease depressive feelings through neurochemical changes, which have been shown in laboratory studies of depression and exercise.17^–20 There are also community-based data from adults that corroborate this finding, showing an association between regular physical exercise and decreased likelihood of depressive symptoms.21^–23

It is also possible that feelings of enjoyment when riding a bicycle or the ability to remove oneself from a situation when angry (i.e., by bicycling) decreases the likelihood of feeling depressed. As the data do not describe the intensity/duration of bike riding, social/environmental effects are a plausible explanation as well. The same might apply for going on a walk.

Not having feelings of depression may also lead to an increased chance that a youth will choose to ride a bicycle, whereas children who are depressed may not choose this because of lethargy or anhedonia. Alternatively, there may be a third factor, either genetic or environmental, that is associated with a decreased likelihood of becoming depressed and an increased likelihood of engaging in active, rather than destructive, behaviors when feeling angry. This factor could be environmental/ learned (e.g., parents modeling behaviors) or individual (e.g., personality factors such as extraversion or conscientiousness).

Physical activity may be a way of releasing feelings of anger that decrease the likelihood that these feelings will become internalized and lead to depression, despite the differences in functions employed when riding a bicycle compared with talking to someone. Previous data show that specific coping behaviors are associated with improved outcomes of physical health problems among youths.12^,24^–26 These results are consistent with and extend those findings to include mental health among youths.

The reason for the association between specific types of substance use behaviors and increased likelihood of depressed feelings is also not available from these data. There are, however, several possibilities suggested by these results that could be explored further in future studies, ideally with longitudinal data.

It may be that engaging in specific behaviors when angry leads to situations that are depressogenic. For instance, cigarette smoking when angry could lead to anxiety disorders, such as panic attacks,27^,28 which are associated with an increased risk of depression.29^–31 Also, consuming alcohol and taking illicit drugs may lead to decreased school performance, which could result in feeling sad and depressed because of lack of academic success. The use of drugs and alcohol could also lead to depressive feelings through neurobiological pathways.32^,33

It is also conceivable that a third factor, either parental influence or personality factors, may increase the likelihood that an individual may both engage in these behaviors and become depressed. For instance, violent behavior when angry (e.g., physical fighting), which is also common among youths with conduct disorder or other behavioral problems,34 is also associated with an increased likelihood of depression.

Alternatively, it is also possible that being depressed, or having symptoms of depression, could have affected responses to the questionnaire by leading those who were depressed to endorse behaviors that are less active and more commonly associated with depression.

That gender differences in coping behavior when angry appear to moderate the relation between gender and the likelihood of feelings of depression among 6th- and 8th- but not 10th-grade students is a new finding among youths but is consistent with previous adult data. Previous literature has consistently shown that the gender difference in major depression emerges at the age of approximately 13 years.29^,30 The finding that coping behavior when angry explains the observed gender difference in feelings of depression among 6th-grade students, who have a mean age of 11 years, and 8th-grade students, who have a mean age of 13 years, is not necessarily surprising or inconsistent with previous literature.

Although this explanatory function did not appear among 10th-grade students, coping behaviors when angry appear to have some ability to moderate the gender differences in depressive feelings among youths, and this may be a worthwhile avenue for future study, especially in efforts to identify effective, non-invasive methods of preventive intervention. The reason for the gender differences in coping behavior when angry is not known. Although male youths were more likely to participate in several potentially less productive coping behaviors (e.g., drinking alcohol, taking drugs), they were also more likely to engage in physical activity, which is associated with decreased depressive feelings.

Limitations of this study are numerous and must be considered when interpreting these findings. First, the measure of depression is not a clinical assessment by a clinician of Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition35 major depression, and therefore future studies that replicate these findings with data that contain Diagnostic Interview Schedule for Children–diagnosed depression would be useful.

Second, the cross-sectional nature of the data does not allow causal assumptions regarding the sequential nature of the relation between coping behaviors when angry and the onset of feelings of sadness or depression. Studies that investigate these associations with longitudinal epidemiological data are needed next.

Third, the measure of these behaviors does not include any information on the frequency of the behavior; therefore, there is no measure of whether a dose–response group relation may exist between the extent to which an individual engages in each behavior and the likelihood of feeling depressed.

Fourth, although the response rate was 83%, which is relatively strong, there may be unidentified differences between responders and nonresponders in the association between coping behaviors when angry and depressive feelings.

Finally, there were no available data on the range of mental disorders; therefore, it is not possible to adjust for the potential confounding effect of mental disorders. Given the limitation of the data available to investigate this link in the current study, replication with longitudinal data that includes more sophisticated assessment of depression is needed.

If intervention programs could provide youths with new behaviors for coping with anger, and if these behaviors could be learned and practiced, these data may offer youths with alternate, healthier modes of coping with anger and other unpleasant feelings that will help them to maintain and protect mental health.