Adolescent Humor and its Relationship to Coping, Defense Strategies, Psychological Distress, and Well-Being

Child Psychiatry Hum Dev (2007) 37:255 271 DOI 10.1007/s10578-006-0034-5 ORIGINAL PAPER Adolescent Humor and its Relationship to Coping, Defense Strategies, Psychological Distress, and Well-Being Sarah J. Erickson Æ Sarah W. Feldstein Published online: 14 November 2006 Ó Springer Science+Business Media, LLC 2006 Abstract Objective This study investigated the psychometric properties of the Humor Styles Questionnaire (HSQ) in measuring adolescent humor, including the relationship between humor and coping style, defense style, depressive symptoms, and adjustment in a non-clinical sample of adolescents. Method Humor, coping, defense strategies, depressive symptoms, and adjustment were investigated in 94 adolescents aged 12 15. Results The HSQ demonstrated adequate internal consistency. Inter-scale correlational patterns and scale means were similar for adolescents and adults. Convergent validity for the HSQ was supported by its relationship to: an adolescentstandardized humor defense scale; coping and defense strategies; and depressive symptoms and adjustment. Clinical utility of the HSQ was demonstrated by its unique contribution in predicting both depressive symptoms and adjustment above and beyond contributions from coping and defense composites. Discussion The HSQ appears to be a psychometrically sound and clinically useful measure to assess humor dimensions in adolescents. Keywords Adolescence Æ Humor Æ Coping Æ Defenses Æ Depressive symptoms Æ Adjustment Æ Positive psychology Introduction Adolescence in mainstream American culture requires adjustment to new biological, social, and academic challenges [1, 2]. Although conceptualizations of adjustment are S. J. Erickson (&) Psychology Department, Logan Hall, University of New Mexico, Logan Hall, University of New Mexico, Albuquerque, NM 87131, USA e-mail: erickson@unm.edu S. W. Feldstein University of New Mexico, Albuquerque, NM 87131, USA

256 Child Psychiatry Hum Dev (2007) 37:255 271 culturally contextual [3], because adolescence often entails the evaluation and synthesis of childhood identifications in order to construct a viable pathway toward adulthood [4], this process is often stressful [5, 6]. Due to the stress inherent in this period, adolescence is an opportune time to investigate the nature of adolescent humor, how humor is related to moderators of stress such as coping and defense strategies, and the role humor may have in predicting mood disturbance and psychological well-being. In the absence of a psychometrically sound measure of adolescent humor in the literature, the aim of this study was to investigate the psychometric properties of an adult-standardized humor scale with an adolescent sample. Multidimensional Nature of Humor Construct Humor generally refers to a stable personality trait or individual differences variable [7] and is conceptualized as a multifaceted concept involving cognitive, emotional, behavioral, psychophysiological, and social components [8]. According to Martin et al. [9], sense of humor is a comprehensive and multidimensional construct comprised of a number of modestly related factors, including but not limited to a cognitive ability, an aesthetic response, a habitual behavior, an emotion-related temperament trait, an attitude, and a coping strategy or defense mechanism. Although the vast majority of humor research has been conducted with adults, research on children s humor provides evidence for the cognitive, social, defensive, and physiological functions of humor [10]. This literature suggests that children with a good sense of humor are more likely to be assertive in social and academic settings [11, 12], and more likely to exhibit social and academic competence [13] as well as communicative competence [14]. Regarding the more developed empirical literature with adults, it appears that not all types of humor support positive adjustment [9, 15, 16]. In fact, recent theoretical models of humor elucidate what constitutes humor and under what conditions humor may or may not benefit physical or psychological health [15 17]. A leading contemporary multidimensional model of humor that addresses both the positive and negative components of humor has identified empirically four humor dimensions, two of which are positively associated and two of which are negatively associated with psychosocial well-being in adults [9]. The four dimensions are relatively independent and are distinguished by their location along two continua: humor that is relatively benign and benevolent versus potentially detrimental or injurious to oneself or others; and humor that functions to enhance oneself versus one s relationships with others. The first two types of benign or positive humor are what have been traditionally measured in humor research. First, affiliative humor involves the tendency to joke around with others, tell amusing stories, and laugh with others in an effort to enhance relationships. This form of humor is positively associated with cheerfulness, high self-esteem, and psychological well-being and negatively correlated with depression and anxiety [9]. The second form of positive, self-enhancing humor, includes perspective-taking humor, the tendency to maintain a humorous outlook on life, and the use of humor in emotion regulation and coping. Like affiliative humor, self-enhancing humor is positively associated with cheerfulness, high self-esteem, and psychological well-being and negatively associated with depression and anxiety [9]. Because of their associations with well-being, their role as buffers against distress, their adaptive, self-organizing nature, and their embeddedness in social contexts,

Child Psychiatry Hum Dev (2007) 37:255 271 257 these positive forms of humor may be considered positive individual traits, even strengths and virtues, within a positive psychology framework [18, 19]. In fact, within this paradigm, humor has been classified as a character strength within the transcendence virtue [20] by meeting criteria including ubiquity across cultures, contribution to personal fulfillment, moral value, elevation of others, and traitlike stability. On the other end of the adjustment correlate spectrum are the humor dimensions that have not been addressed in past research. Aggressive humor includes efforts to enhance oneself at the expense of relationships with others and includes sarcasm, teasing, and the use of humor to criticize or manipulate others. Aggressive humor is positively associated with hostility and aggression [9]. At the expense of oneself versus relationships, self-defeating humor includes using humor in an excessively self-disparaging and ingratiating way. Self-defeating humor has been positively correlated with depression, anxiety, hostility, aggression, psychiatric symptoms and negatively correlated with self-esteem and psychological well-being [9]. Humor and Coping Humor has often been conceptualized as a form of coping [21 24] in that it involves a multidimensional, transactional process concerning how people handle stress [25 27]. One of the few adolescent-specific coping models with empirical support is the two-tiered model of approach versus avoidance coping [25, 26, 28, 29]. Approach coping is defined as an individual s attempts to directly alter a stressor or the impact of a stressor, and has been associated with positive adjustment [25, 26, 28] and lower levels of depressive symptoms [5] in adolescents. Although not orthogonal to approach coping, avoidance coping encompasses those responses which orient one away from contending with the stressor or related thoughts and emotions. Adolescents who employ avoidance coping as their only copig strategy have demonstrated less adaptive adjustment [25, 26, 30]. There are several theories as to how positive forms of humor may function as intrapersonal adaptive coping strategies [21 24]. For example, positive humor may serve as a coping strategy by moderating the effects of stress through more positive appraisals and more realistic cognitive processing of environmental information [31]. In this way humor may involve a cognitive affective shift or restructuring of threatening situations [24, 32]. In these theories, positive humor may be conceptualized as a form of approach coping. Although not directly addressed in the literature, we would expect that negative forms of humor would be associated with avoidance coping. In evaluating humor and coping with younger samples, the limited empirical research in this area suggests that the use of humor as a coping strategy may increase with age and differ by gender. For example, Fuhr [33] found a significant increase in the use of humor in uncertain situations between the ages of 11 and 12. This parallels the coping literature wherein age impacts available resources, breadth, capacity, and flexibility to alter or ameliorate stressful situations [34]. In terms of gender differences in approach and avoidance coping, several studies have found that girls use significantly more approach coping strategies, and boys rely more consistently upon avoidance coping techniques [25, 26, 28]. Furthermore, a study with adolescents

258 Child Psychiatry Hum Dev (2007) 37:255 271 found that boys used aggressive and sexually related humor strategies more, whereas girls reported that humor had a greater ability to cheer them up [33]. Humor and Defense Strategies Humor may also be conceptualized as a defense [9]. Consistent across theories of defense, defenses operate unconsciously through the alteration of one s perception of internal and external reality in order to reduce the experience of distress [27, 35 38]. Humor as a defense allows individuals to face challenging or threatening situations without being overwhelmed by negative emotion [39]. Defenses may be classified into two categories, mature and immature. Mature defenses synthesize and attenuate stress, maintaining the ego and reality in the process, whereas immature defenses deny and distort conflict [38]. Within a defense paradigm humor is often conceptualized as a mature defense [38, 40]. In relation to Martin et al. s [9] multidimensional model of humor that includes dimensions formerly not included in humor or defense measures, the positive dimensions of humor would be considered mature defenses whereas the more negative dimensions of humor would be considered immature defenses. Although the categories of mature and immature defense do not change with age and developmental period, each stage of development includes a greater reliance on different strategies within these categories [40 42]. In particular, there is a predictably higher use of immature defenses with younger ages [40, 42]. Gender differences have been identified in the organization of defenses, the frequency of use of individual defense strategies, and the association of defenses to psychological symptoms [37, 40, 42]. However, at the composite level, gender differences have been non-significant [40, 42]. Humor and Depressive Symptoms Throughout the coping literature in adolescent samples, there is an association between poor coping, poor adjustment, and depressive symptoms [5, 28, 43]. To the extent that humor is conceptualized as a form of coping or defense, one would expect positive humor to be associated with adjustment and negative humor to be associated with poor adjustment, including depressive symptoms. Toward that end, Freiheit et al. [44] found that when both depressed and non-depressed adolescents used humor to deal with stressful situations they reported significantly lower levels of depressive symptoms. Furthermore, in studies of humor as a coping strategy with adults, humor coping has been associated with lower levels of depressive symptoms [32, 39, 45 47]. Humor and Adjustment Adolescent adjustment may be defined in cross-situational and comprehensive terms: facility in interpersonal relations, relationship to parents, self-esteem and selfreliance [48]. Research has consistently shown a relationship between children s adjustment and social competence [11 13, 49] while children who experience difficulties in peer and social interactions face increased risk of negative self-perceptions

Child Psychiatry Hum Dev (2007) 37:255 271 259 [49, 50]. To the extent that positive forms of humor reflect social competence [9], and more generally, positive, adaptive individual traits within social contexts [18], these positive forms of humor are hypothesized to relate positively to adjustment. In contrast, similar to the adult literature [9], the negative forms of humor are hypothesized to relate inversely to adjustment. Current Study The purpose of this study was to investigate the psychometric properties of the Humor Styles Questionnaire (HSQ) in measuring adolescent humor, and to gain a greater understanding of the relationship between humor and coping style, defense style, depressive symptoms, and adjustment in a non-clinical sample of adolescents. Preliminary hypotheses related to age and gender effects. Because there is little consistent support for age-related differences for children between ages 12 15 in either humor, coping [34], defense [40, 42], depressive symptoms [51], or adjustment [48], we hypothesized that we would not find age effects when the age range was limited to 12 15. Based on the literature related to gender differences, we hypothesized that adolescent girls would employ more approach coping methods whereas adolescent boys would endorse more avoidance coping methods and more aggressive and self-defeating humor. Primary aims of this study related to investigating the psychometric properties of the HSQ with an adolescent sample. Specifically, we addressed questions relating to internal item consistency, comparison with adult sample responses, and convergent validity. We evaluated convergent validity by comparing HSQ scale scores with a humor subscale standardized on adolescents; by investigating the relationship between humor and coping and defense composites; and by comparing correlations between HSQ scale scores and measures of adolescent depressive symptoms and personal adjustment with similar correlations with adults. Finally, in order to address the clinical utility of the HSQ in adolescents, we investigated whether the HSQ scales added any unique prediction of depressive symptom or personal adjustment variance above and beyond coping and defense predictors. Method Participants Through the YMCA and Parks and Recreation Day Camp programs in one southwestern city, 113 adolescents were recruited to participate in this study. Nineteen of these potential participants failed to provide adequate data to be included, decreasing our total sample to 94. Our participation rate was therefore 83.2%. Of the total sample, 57 (60.6%) were female and 37 (39.4%) were male. The mean age of the participants was 12.97 years (SD =.94) (age 12: 40.4%, age 13: 27.7%, age 14: 26.6%, age 15: 5.3%) and, on average, study participants had completed seventh grade (SD = 1.05). A majority of the sample was Caucasian (n = 69, 73.4%), followed by Hispanic (n = 11, 11.7%), African American (n = 6, 6.4%), Hispanic Caucasian (n = 3, 3.2%), Asian-American (n = 3, 3.2%), and Native American (n = 2, 2.1%).

260 Child Psychiatry Hum Dev (2007) 37:255 271 Procedures Approval was obtained from the sponsoring University s Institutional Review Board. Prior to data collection, parents and adolescents were provided information about the details of participation. Interested parents and adolescents provided written consent and assent prior to participation. Questionnaires were administered in small groups (n = 2 5). The investigator read each question and response option aloud, while the adolescents independently completed answer forms. After completion of the forms, the investigator debriefed the groups by explaining the purpose of the study. As compensation for their time and effort, participating adolescents received a pack of gum and entry into a lottery whereby 1 per 15 participants received a $25 gift certificate to a local shopping mall. Measures Humor Styles Questionnaire [9] A review of the humor literature revealed no psychometrically sound humor measures standardized on adolescents. Thus, the HSQ was selected in consultation with its author [52]. All items were reviewed and deemed potentially appropriate for adolescents. The purpose of the HSQ is to determine how one generally uses humor. The subscales of this measure were derived through Jackson s [53] construct-based approach [9]. The final set of items were assessed with a principal components analysis with Varimax rotation, yielding a reliable self-report measure of four humor dimensions (internal consistency coefficients:.77.81; test retest correlations:.80.85; correlations between scales: <.13.36) [9]. The HSQ is a 32-item self-report measure with response options on a seven-point Likert scale. Four final scores are derived: (1) affiliative humor (items include: I enjoy making people laugh. ), (2) self-enhancing humor (items include: If I am feeling upset or unhappy I usually try to think of something funny about the situation to make myself feel better. ), (3) aggressive humor (items include: If I don t like someone, I often use humor or teasing to put them down. ), and (4) selfdefeating humor (items include: I will often get carried away in putting myself down if it makes my family or friends laugh. ). Coping Response Inventory-Youth [26] The Coping Response Inventory-Youth (CRI-Y) is a 48-item self-report measure requiring the respondent to identify the most important problem faced within the last year, the valence and pervasiveness of the problem, and the ways in which the problem was addressed. Response options are on a four-point Likert scale. The measure yields two composite scores: approach coping and avoidance coping. Internal consistency coefficient estimates range from.68 to.72 for approach coping and from.55 to.72 for avoidance coping [26]. Internal consistency coefficients for our sample were.85 and.82 for approach and avoidance coping, respectively.

Child Psychiatry Hum Dev (2007) 37:255 271 261 Response Evaluation Measure [40] The response evaluation measure (REM-71) is a 71-item self-report measure of defense reactions standardized with adolescents [40, 54]. Participants indicate their endorsement of each item on a nine-point Likert scale. The instrument includes 21 types of defenses by which individuals may distort unwelcome realities [40]. The measure yields two composite scores: mature and immature defenses. Internal consistency coefficient estimates were.84 and.68 for mature and immature composites, respectively [40]. Internal consistency coefficients for our sample were.80 and.81 for mature and immature defenses, respectively. Of particular interest in this study, the humor scale includes four items reflecting how the individual deals with emotional conflict or external stressors by emphasizing the amusing or ironic aspects of the conflict or stressor. Items include I laugh at myself pretty easily and I usually see the funny side to the problems I have. Internal item consistency of the humor scale for our sample was.69. Children s Depression Inventory-Short Form [55] The Children s Depression Inventory-Short Form (CDI-S) is a 10-item abbreviated version of the most widely used 27-item self-report measure assessing cognitive, affective, and behavioral depressive symptoms in children and adolescents. The CDI-S is considered a brief screening and research tool, rather than a clinical assessment tool [55]. Each item on the CDI-S is scored on a three-point scale. For children aged 8 17, this measure has strong internal consistency, good test retest reliability, and good convergent validity with other self-report measures in clinical and non-clinical samples. The CDI-S correlates r =.89 with the full form and its internal consistency coefficient estimates range from.80 to 94 [55]. The internal consistency coefficient for our sample was.68. Behavior Assessment System for Children: Self-report of Personality, Adolescent Form [48] The Behavior Assessment System for Children: Self-report of Personality, Adolescent Form (BASC-SRP-A) personal adjustment composite is a measure of adaptive functioning [48]. The BASC-SRP-A includes 186 dichotomous (true/false) items designed to measure adaptive and problem behaviors for adolescents aged 12 18 [48]. The measure yields five composite scores: school maladjustment, clinical maladjustment, other problems, an emotional symptoms index, and personal adjustment. The personal adjustment composite score was selected for use in this study. The BASC-SRP-A has demonstrated high levels of internal consistency and moderate test retest reliability, and an internal consistency coefficient estimate of.80 [48]. The internal consistency coefficient for our sample was.84. Statistical Analyses Analyses first examined potential mediators requiring consideration, followed by the HSQ psychometric analyses. To provide an omnibus test of age with all variables of interest, we conducted a single step regression wherein age was regressed on humor

262 Child Psychiatry Hum Dev (2007) 37:255 271 scales (affiliative, self-enhancing, self-defeating, and aggressive), coping dimensions (approach and avoidance), defense composites (mature and immature), depressive symptoms (CDI-S), and personal adjustment (BASC). Acknowledging the resultant overfitting of the model and inflated model estimates, the regression equation was significant (adjusted R 2 =.09, p <.05), but no variable was uniquely predictive after correction for type I error. In order to explore gender differences we conducted a profile analysis by gender on the four humor subscales, and a second profile analysis on the two coping and two defense composites with Wilk s Lambda tests. Our sample evidenced no significant differences between genders on the four humor scales or on the coping and defense composites. Thus, we pooled subjects across age and gender for subsequent analyses. The psychometric properties of the HSQ with an adolescent sample were assessed in several ways. First, inter-item consistencies of the four HSQ scales were obtained with Cronbach alphas. We then investigated inter-scale correlations with our adolescent sample and compared them with adult inter-scale correlations using Fisher z-transformations. Next, mean differences between the adult and adolescent samples on the four humor scales were compared using effect size estimates adjusted for small sample biases [56]. Convergent validity was determined by investigating relationships between the four scales of the HSQ with the humor subscale of the REM-71; with coping and defense composites; and with measures of adolescent depressive symptoms (CDI) and personal adjustment (BASC Personal Adjustment Scale). Because the CDI was positively skewed and non-normally distributed, and the BASC Personal Adjustment Scale was negatively skewed and non-normally distributed, we utilized median splits and treated both as dichotomous variables. Specifically, we correlated the four HSQ scales with the REM-71 humor subscale, the CRI-Y and REM-71 composites, the CDI, and the BASC Personal Adjustment Scale. We utilized Pearson bivariate correlations, adjusted for multiple tests, for the first three sets of correlations and point biserial correlation coefficients for dichotomous variables for the latter two set of correlations. Because correlation coefficients are standardized scores, free from scale of measurement, we then contrasted these latter two sets of correlations with similar ones using different measures obtained with adults. Fisher z-tests were used for these contrasts. Finally, two hierarchical logistic regressions were conducted to assess how the four humor scales accounted for adolescent depressive symptoms and personal adjustment after accounting for coping and defense measures. Both logistic regressions consisted of two steps: in step one measures of approach and avoidance coping and mature and immature defenses were entered. The four humor scales were then entered in the second step, with incremental variance accounted for in the criterion measure used to assess the relative collective importance of the four humor scales. Results Inter-item Consistency Cronbach alpha was used to assess the extent of inter-item consistency of the HSQ applied to an adolescent sample. Table 1 includes a comparison of the internal

Child Psychiatry Hum Dev (2007) 37:255 271 263 Table 1 Internal consistency of the Humor Styles Questionnaire (HSQ) scales with an adolescent sample (N = 94) HSQ scale Cronbach alpha Reported by Martin et al. (2003) 1. Affiliative humor.80.81 2. Self-enhancing humor.81.79 3. Self-defeating humor.80.65 4. Aggressive humor.77.58 For this sample consistency coefficients from this sample and those reported with an adult sample [9]. Inter-scale correlations Inter-correlations among the four scales were compared to those reported by Martin et al. s [9] for an adult sample using Fisher z-transformations (Table 2). A similar magnitude and direction of scale correlations was observed for adolescents and adults on three of the six correlations: between the affiliative and self-enhancing, self-defeating and self-enhancing, and aggressive and self-defeating scales. In contrast, three of the six correlations were statistically different for adolescents compared with adults. For the adolescent sample, the affiliative and aggressive humor dimensions were unrelated, whereas they were positively and significantly related for adults (z = 2.53, p <.05). In addition, for the adolescents only, affiliative and self-defeating forms of humor were significantly and negatively associated while these scales were unrelated for the adult sample (z = 2.32, p <.05). Finally, the contrast of correlations between self-enhancing and aggressive humor dimensions for adolescents and adults was significant, but the correlations for both adolescents and adults were non-significant. Normative comparison Table 3 displays the four scale means for both Martin et al. s [9] adult sample and the adolescent sample in this study. Effect size estimates adjusted for small sample biases [56] are provided to indicate the magnitude of mean differences between the Table 2 Scale inter-correlations in the Humor Styles Questionnaire: a comparison of adolescent and adult responses Affiliative Self-enhancing Self-defeating Aggressive Affiliative.49***.20 a, *.02 a Self-enhancing.35***.09.19 a Self-defeating.05 b.07.24* Aggressive.25 b, ***.04 b.23*** Correlations for adolescents are above the diagonal, for adults below the diagonal. Adolescent N = 94; adult N = 1,195. Superscript letters indicate statistical difference between adolescent and adult correlation (p <.05) * p <.05 ** p <.01 *** p <.001

264 Child Psychiatry Hum Dev (2007) 37:255 271 Table 3 Humor Styles Questionnaire scale means (SD) for adult and adolescent samples Affiliative Self-enhancing Self-defeating Aggressive Adult sample 46.4 (7.17) 37.3 (8.33) 25.9 (9.22) 28.5 (8.79) Adolescent sample 45.96 (8.56) 37.59 (9.32) 25.21 (7.82) 28.38 (8.11) Effect size (d).06.03.08.01 Significance.57.75.48.90 Adult sample = Martin et al. [9] sample two sample responses. As shown, absolute mean differences between the adult and adolescent samples appear to reflect sampling error alone (all ps >.05), with the largest effect size estimate (d =.08, self-defeating scale) generally considered a small effect or mean difference [57]. Convergent validity Convergent validity was determined by investigating relationships between the four scales of the HSQ with the humor subscale of the REM-71; with coping and defense composites; and with measures of adolescent depressive symptoms (CDI) and personal adjustment (BASC Personal Adjustment Scale). First, the item consistency of the four REM-71 humor items comprising the humor subscale was determined to be adequate (Cronbach alpha =.69). Because the REM-71 humor scale was normally distributed, we employed Pearson correlations with this scale and the four HSQ scales (Table 4). In order to determine the relationship between HSQ humor scales and coping and defense composites, we conducted bivariate Pearson correlations (Table 4). Correlations revealed significant relationships in the predicted direction: affiliative humor was negatively associated with the immature defense composite and positively associated with the mature defense composite; self-enhancing humor was positively associated with approach coping and the mature defense composite; aggressive humor was positively associated with the immature defense composite and negatively associated with approach coping and the mature defense composite; and selfdefeating humor was positively associated with avoidance coping and the immature composite. In sum, humor dimensions were significantly associated with defense Table 4 Bivariate correlation matrix with Humor Styles Questionnaire dimensions, coping composites, and defense composites (pooled within-groups matrices correlations) Affiliative humor Self-enhancing humor Aggressive humor Self-defeating humor REM-71 humor.44**.49**.07.25* Approach coping.17.28**.22*.01 Avoidance coping.00.05.05.27** Immature defense.24*.05.21*.45*** Mature defense.25*.60***.27**.08 * p <.05 (with Bonferonni correction) ** p <.01 (with Bonferonni correction) *** p <.001 (with Bonferonni correction)

Child Psychiatry Hum Dev (2007) 37:255 271 265 composites in six out of a possible eight correlations, and with coping composites in three out of a possible eight correlations. We then correlated the four HSQ scales with the CDI and BASC Personal Adjustment Scale. Preliminary data analyses indicated that the CDI measure was significantly and positively skewed and the personal adjustment measure was significantly and negatively skewed. Neither distribution contained statistical outliers ( ± 3SD). Our adolescent sample scored well within the non-clinical range on both measures, with no respondents reporting clinically elevated scores. Specifically, regarding the CDI-S, the majority of the sample had scores equal to or below 2, placing them in the 50 59% of age-matched peers. Our sample s mean BASC adjustment score was 166.88, placing the average adolescent s score in the 15 percentile of age-matched peers. Within the adolescent sample, point biserial correlations between the four HSQ humor scales and measures of depressive symptoms and personal adjustment followed predictable patterns (Table 5). Specifically, affiliative and self-enhancing humor were significantly and negatively correlated with depressive symptoms and significantly and positively correlated with personal adjustment. In contrast and as expected, self-defeating humor evidenced the opposite pattern, and aggressive humor was negatively and significantly correlated with adjustment. Pearson correlations with similar constructs but different measures of depressive symptoms and psychological well-being were reported by Martin et al. [9] and are included in Table 5 with the point biserial correlations used with our adolescent sample. Seven of the eight contrasts between Martin et al. [9] and the adolescent sample correlations with convergent measures were not significant, suggesting that the nature of responses to the humor scales and the relationship of these responses to measures of depressive symptoms and well-being were relatively similar for adults and adolescents. In particular, the magnitude of correlations between aggressive humor and depressive symptoms was significantly different for adolescents and adults. However, the absolute magnitude of these two correlations was small, Table 5 Humor Styles Questionnaire (HSQ) convergent validity analyses: adult and adolescent samples HSQ scales Depressive symptoms C a Well-being/adjustment C a Adult Adolescent Adult Adolescent Affiliative.22***.23**.23.26*.28***.43 Self-enhancing.33***.28**.31.46***.33**.41 Self-defeating.24***.24**.50.24*.29***.32 Aggressive.08.17.01.11.17*.36 Adult depressive symptoms were based on the Center for Epidemiological Studies-Depression Scale; adult well-being from the Ryff Well Being Scale. Adolescent depressive symptoms were based on the Children s Depression Inventory; adolescent adjustment from the Behavioral Assessment for Children Personal Adjustment scale. Adult correlations utilized Pearson bivariate correlations; adolescent correlations utilized point biserial correlation coefficients * p <.05 ** p <.01 *** p <.001 a Correlations were contrasted using Fisher z-tests

266 Child Psychiatry Hum Dev (2007) 37:255 271 suggesting that neither adolescent nor adult aggressive humor is related to depressive symptoms. Predicting Depressive Symptoms and Personal Adjustment: Contribution of Humor Two hierarchical logistic regressions were done to assess how the four humor scales accounted for adolescent depressive symptoms and personal adjustment after accounting for coping and defense measures. Both regressions consisted of two steps: in step one measures of approach and avoidance coping and mature and immature defenses were entered. The four humor scales were then entered in the second step, with incremental variance accounted for in the criterion measure used to assess the relative collective importance of the four humor scales. With adolescent depressive symptoms as the dichotomized dependent measure (median split), the four humor scales accounted for a significant increment in variance (P 2 (4) = 10.25, p <.04) beyond that accounted for in step 1 (P 2 (4) = 20.13, p <.001). Inspection of beta weights associated with the four humor measures in step 2 indicated that the self-defeating (Beta =.08, p <.04) scale accounted for the effect in the predicted direction. In the second regression, personal adjustment was the dependent measure (dichotomized using a median split) and in step 2 the four scales again accounted for a significant increment in variance (P 2 (4) = 12.28, p <.02) beyond that accounted for in step 1 (P 2 (4) = 24.45, p <.0001). In this regression equation, however, no humor scale uniquely contributed to the prediction of the dichotomoized personal adjustment measure. Discussion Given that stress is ubiquitous in adolescence, there appear to be multiple methods adolescents may employ to manage stress, including humor, coping efforts, and defense strategies. There is limited empirical literature related to adolescents use of humor and no standardized humor measures for this population. Thus, this study explored the psychometric properties of the HSQ with adolescents and subsequently, the associations between humor and coping and defense strategies, as well as the relative contribution of humor in predicting depressive symptoms and adjustment during a period of vast developmental transitions. Regarding primary aims of the study, the HSQ appears to have adequate internal consistency with our adolescent sample. The coefficients for affiliative and selfenhancing humor scales are acceptable and comparable to those reported by Martin et al. [9] with an adult sample. Our coefficients for the more negative dimensions of humor, however, were substantially weaker and lower than those reported by Martin et al. [9]. Specifically, whereas the self-defeating scale evidenced low but not unacceptable internal consistency, the aggressive scale demonstrated relative instability. This finding of limited internal consistency for the two negative humor dimensions may suggest that these forms of humor emerge as coherent constructs at a later developmental stage. Future research with a larger sample, divided into younger and

Child Psychiatry Hum Dev (2007) 37:255 271 267 older adolescents, would allow for a direct test of age on the internal consistency, and thereby stability, of these two scales. Overall, means for the four humor scales were not significantly different for adolescents versus adults [9], suggesting that adolescents reported similar endorsement levels to adults. In addition, the pattern of inter-correlations between HSQ subscales for adolescents was similar to those for adults [9]. Specifically, the magnitude and direction of scale inter-correlations were similar for three of the six intercorrelations. In particular, a similar magnitude and direction of scale correlations, all in the expected direction, were observed for adolescents and adults on the affiliative and self-enhancing, self-defeating and self-enhancing, and aggressive and selfdefeating scales. Specifically, the associations between the two adaptive forms of humor and the two maladaptive forms of humor were both significantly and positively related. In addition, although the associations between self-enhancing and aggressive humor were significantly different for adolescents and adults, both correlations were non-significant, suggesting that for both groups, these two humor dimensions are largely unrelated. In contrast, for the adolescent sample the affiliative and self-defeating forms of humor were significantly and negatively associated while these scales were largely unrelated for the adult sample. This result may be explained by the notion that both forms of humor, although one is benign and the other is at one s expense, are used to enhance one s relationship with others. Our finding that adolescents efforts in relationship enhancement are correlated underscores the importance of relationship-building and maintenance during this period [50, 58 60]. In addition, for the adolescent sample the affiliative and aggressive humor dimensions were unrelated, whereas they were positively and significantly related for adults. Because affiliative humor serves to enhance one s relationships with others in a benign way and aggressive humor serves to enhance oneself at the expense of others, it is perhaps more surprising that they were positively correlated for adults [9] than that they were unassociated for adolescents. Regarding additional psychometric results, the HSQ evidenced strong convergent validity. First, affiliative, self-enhancing, and self-defeating humor dimensions were all significantly associated with the humor subscale of the REM-71, a measure standardized with adolescent samples. Not surprisingly, the two positive humor dimensions evidenced the strongest associations. The finding that aggressive humor was not significantly related to the REM-71 humor subscale may be explained by the relative instability of the aggressive humor scale with our adolescent sample and the fact that most humor measures, including the REM-71 subscale, do not purport to measure this less adaptive form of humor [40]. To further investigate convergent validity evidence, the relationship between HSQ humor scales and coping and defense composites was investigated. This study suggests that the area of overlap between humor and coping and defense strategies is partial and in the predicted direction: affiliative humor was negatively associated with the immature defense composite and positively associated with the mature defense composite; self-enhancing humor was positively associated with approach coping and the mature defense composite; aggressive humor was positively associated with the immature defense composite and negatively associated with approach coping and the mature defense composite; and self-defeating humor was positively associated with avoidance coping and the immature composite. These findings are congruent with the literature insofar as less adaptive strategies tend to cluster [3, 61].

268 Child Psychiatry Hum Dev (2007) 37:255 271 These results further suggest that although humor appears to be moderately associated with both defense and coping composites, the HSQ scales were more aligned with defense operations than more conscious coping efforts. In sum, six of a possible eight correlations were significant between HSQ scales and defense composites whereas three of a possible eight correlations were significant between HSQ scales and coping composites. The single strongest association was between selfenhancing humor and the mature defense composite, followed by self-defeating humor and the immature defense composite. This overall pattern of associations is what one would expect from the coping and defense literature [26, 38, 40]. Our finding that adolescents HSQ scale scores were similarly and predictably associated with measures of depressive symptoms and personal adjustment compared with what one would expect from the adolescent literature and compared with the adult sample [9] provide additional convergent validity support. In particular, affiliative and self-enhancing humor were negatively correlated with depressive symptoms and positively correlated with personal adjustment, while self-defeating and aggressive humor evidenced the opposite pattern. These findings are congruent with the adolescent literature [33, 44, 48]. Seven of the eight contrasts between Martin et al. [9] adult sample and the adolescent sample correlations with different but convergent measures were not significant, suggesting that the nature of responses to the humor scales and the relationship of these responses to measures of depressive symptoms and well-being were relatively similar for adults and adolescents. Although the magnitude of correlations between aggressive humor and depressive symptoms was significantly different for adolescents and adults, the two correlations were both non-significant, suggesting that neither adolescent nor adult aggressive humor is related to depressive symptoms. Results of the two sets of hierarchical logistic regressions suggest that the HSQ humor dimensions are highly relevant to adolescent psychosocial functioning in content and scope. These results, along with the results of the correlational analyses, further suggest that the theoretical domains measured by the HSQ overlap moderately with coping and defending, but largely tap distinct arenas of functioning. Humor dimensions appear to yield significantly different information regarding an adolescent s overall functioning than do conscious coping strategies and unconscious defense strategies. In part, because negative dimensions of humor have not traditionally been included in measures of coping and defense, these dimensions may largely represent the unique variance explained by the humor scales in predicting both depressive symptoms and adjustment. However, this point is speculative given the limited stability of these negative humor dimensions in our adolescent sample. In fact, in order to address validity questions with adolescents, future research is warranted to identify a more stable way in which to assess these negative humor dimensions, and to associate them with developmental trends. Of particular interest, self-defeating humor was uniquely predictive of depressive symptoms above and beyond coping and defense contributions. Because selfdefeating humor involves enhancing one s relationship with others to the detriment of the self [9], these results suggest that this type of humor may be particularly salient during adolescence. One of the primary developmental tasks of adolescence is to individuate from one s parents and to move toward the development of a strong peer support network [50, 58 60, 62]. As Martin et al. [9] contend, self-defeating humor may be one of the tools an individual may employ in his or her efforts to create and sustain friendships. The internal consistency of the self-defeating humor

Child Psychiatry Hum Dev (2007) 37:255 271 269 dimension is only modest, however, calling into question the relative stability of this dimension with adolescents. Finally, because this study is cross-sectional, an adolescent s adjustment and level of depressive symptoms could also influence his or her choice and implementation of various methods with which to contend with normative stressors. Contrary to the literature, we found no significant gender differences between humor [9, 33] and coping composites [25, 26, 28]. Our finding of no gender effect for defense composites was similar to previous research with adolescents [61]. Our sample had a high proportion of younger participants and perhaps gender differences for some of these constructs increase with pubertal change and development. This study included a cross-sectional, self-report design and utilized a humor scale developed and refined with a large adult sample. The small sample size precluded cross-validation of the humor factor structure and confirmatory factor analysis with this adolescent sample. The small sample size also precluded moderator analyses (with gender or ethnic group, for example) and the cross-sectional design precluded causal analyses. Although the sample of participants represented those who frequent the summer programs within a southwestern city s metropolitan area, a sample with greater diversity in age, ethnicity, and socio-economic background would yield more externally valid results. In addition, method variance limitations were posed by the use of all self-report measures. Toward this end, obtaining informant information or employing naturalistic or laboratory observations of relevant behavior may be useful. Summary This study addressed how humor was associated with coping, defense, distress, and adjustment in a normative, non-clinical adolescent sample. It is important to consider that because adolescents are developing their strategies for how to manage various stressors, it is developmentally appropriate for early-mid adolescents to employ a variety of strategies in order to manage a host of stressors. The results of this study suggest that humor, as measured by the HSQ, is a psychometrically sound construct for adolescents, overlapping moderately with defense and coping strategies, but including a significant amount of complementarity. Because humor predicted depressive symptoms and adjustment above and beyond coping and defense strategies, the role of humor in adolescents may warrant further investigation. Furthermore, to the extent that positive dimensions of humor represent the transcendence virtue [20], this study contributes to the emerging positive psychology literature. References 1. Cicchetti D, Rogosch FA (2002) A developmental psychopathological perspective on adolescence. J Consult Clin Psychol 70(1):6 20 2. Erikson EH (1968) Identity: youth and crisis. W.W. Norton, New York 3. Masten AS (2001) Ordinary magic: resilience processes in development. Am Psychol 56(3):227 238 4. Marcia JE (1980) Identity in adolescence. In: Adelson J (ed) Handbook of adolescent psychology. Wiley, New York 5. Dumont M, Provost MA (1999) Resilience in adolescents: protective role of social support, coping strategies, self-esteem, and social activities on experience of stress and depression. J Youth Adolesc 28(3):343 363

270 Child Psychiatry Hum Dev (2007) 37:255 271 6. Igra V, Irwin CE (1996) Theories of adolescent risk-taking behavior. In: DiClemente RJ, Hansen WB, Ponton LE (eds) Handbook of adolescent risk behavior. Plenum, New York, pp 35 51 7. Ruch W (ed) (1998) The sense of humor: explorations of a personality characteristic. Mouton de Gruyter, New York 8. Martin RA (2000) Humor. In: Kazdin AE (ed) Encyclopedia of psychology, vol 4. American Psychological Association, Washington/Oxford University Press, New York, pp 202 204 9. Martin RA, Puhlik-Doris P, Larsen W, Gray J, Weir K (2003) Individual differences in uses of humor and their relation to psychological well-being; development of the Humor Styles Questionnaire. J Res Pers 37:48 75 10. McGhee PE, Chapman AJ (1980) Children s humour. Wiley, Chichester, England 11. Bell NJ, McGhee PE, Duffey NS (1986) Interpersonal competence, social assertiveness and the development of humor. Br J Dev Psychol 4:51 55 12. McGhee PE (1980) Development of the sense of humour in childhood: a longitudinal study. In: McGhee PE, Chapman AJ (eds) Children s humour. Wiley, Chichester, England, pp 213 236 13. Masten AS (1986) Humor and competence in school-aged children. Child Dev 57:461 473 14. Carson DK, Skarpness LR, Schultz NW, McGhee PE (1986) Temperament and communicative competence as predictors of young children s humor. Merrill Palmer Q 32:415 426 15. Kuiper NA, Grimshaw M, Leite C, Kirsh G (2004) Humor is not always the best medicine: specific components of sense of humor and psychological well-being. Humor 17(1/2):135 168 16. Martin RA (2001) Humor, laughter, and physical health: methodological issues and research findings. Psychol Bull 127(4):504 519 17. Kuiper NA, Olinger LJ (1998) Humor and mental health. In: Friedman H (ed) Encyclopedia of mental health, vol 2. Academic, San Diego, pp 445 457 18. Seligman MEP, Csikszentmihalyi M (2000) Positive psychology: an introduction. Am Psychol 55(1):5 14 19. Seligman MEP, Steen TA, Park N, Peterson C (2005) Positive psychology progress: empirical validation of interventions. Am Psychol 60(5):410 421 20. Peterson C, Seligman MEP (2004) Character strengths and virtues: a handbook and classification. American Psychological Association, Washington 21. Abel M (2002) Humor, stress, and coping strategies. Humor 15(4):365 381 22. Henman LD (2001) Humor as a coping mechanism: lessons from POWs. Humor 14(1):83 94 23. Lefcourt HM, Martin RA (1986) Humor and life stress: antidote to adversity. Springer, Berlin Heidelberg, New York 24. Martin RA, Kuiper NA, Olinger LJ, Dance KA (1993) Humor, coping with stress, self-concept, and psychological well-being. Humor 6:89 104 25. Causey DL, Dubow EF (1992) Development of a self-report coping measure for elementary school children. J Clin Child Psychol 21(1):47 59 26. Ebata AT, Moos RH (1991) Coping and adjustment in distressed and healthy adolescents. J Appl Dev Psychol 12:33 54 27. Lazarus RS (2000) Toward better research on stress and coping. Am Psychol 55(6):665 673 28. Herman-Stahl MA, Stemmler M, Petersen AC (1995) Approach and avoidant coping: implications for adolescent mental health. J Youth Adolesc 24(6):649 665 29. Phipps S, Srivastava DK (1997) Repressive adaptation in children with cancer. Health Psychol 16(6):521 528 30. Recklitis CJ, Noam GG (1999) Clinical and developmental perspectives on adolescent coping. Child Psychiatry Hum Dev 30(2):87 101 31. Kuiper N, Martin R (1998) Laughter and stress in daily life: relation to positive and negative affect. Motiv Emot 22:133 153 32. Nezu AM, Nezu CM, Blissett SE (1988) Sense of humor as a moderator of the relation between stressful events and psychological distress: a prospective analysis. J Pers Soc Psychol 54(3):520 525 33. Fuhr M (2002) Coping humor in early adolescence. Humor 15:283 304 34. Compas BE, Connor-Smith JK, Saltzman H, Thomsen AH, Wadsworth ME (2001) Coping with stress during childhood and adolescence: problems, progress, and potential in theory and research. Psychol Bull 127(1):87 127 35. Cramer P (2000) Defense mechanisms in psychology today: further processes for adaptation. Am Psychol 55:637 646 36. Erickson SJ, Feldman SS, Steiner H (1996) Defense mechanisms and adjustment in normal adolescents. Am J Psychiatry 153(6):826 828