The Association Between Sense of Humor and Trauma-Related Mental Health Outcomes: Two Exploratory Studies

ABSTRACT Two studies (n = 73, n = 132) explored the association between sense of humor and trauma-related well-being outcomes. It was found that sense of humor was not associated with reports of posttraumatic growth as measured by the Posttraumatic Growth Inventory (PTGI). Self-enhancing humor was positively associated with positive changes as measured by the CiOQ-P. Benign humor styles were associated negatively with emotion regulation difficulties and negative changes (CiOQ-N). Self-defeating humor was associated positively with negative changes, avoidant states, and emotion regulation difficulties. The results suggest that self-enhancing humor could be helpful in order to cope with trauma.

For example, Edwards and Martin (2010) and Jovanovic (2011) found no significant association between injurious humor styles and satisfaction with life. Several studies found a negative correlation between self-defeating humor, self-esteem (e.g., Edwards & Martin, 2010;Martin et al., 2003) and psychological well-being (e.g., Erickson & Feldstein, 2007;Martin et al., 2003;Páez et al., 2013). Edwards and Martin (2010) found a negative association between self-defeating humor and optimism while Martin and colleagues did not observe such an association. Aggressive humor was found to be correlated negatively with happiness (e.g., Jovanovic, 2011;Martin et al., 2003) and not to be correlated with psychological well-being (e.g., Erickson & Feldstein, 2007;Martin et al., 2003;Páez et al., 2013).
Although humor and its correlates have been increasingly studied, it is still not entirely clear whether a sense of humor could be a protective factor against the negative psychological impact of trauma. It is, furthermore, not well-known whether sense of humor is associated with posttraumatic growth. Several studies investigated the association between humor and posttraumatic growth (Cofini et al., 2014;Park, Cohen, & Murch, 1996;Peterson, Park, Pole, Andrea, & Seligman, 2008;Schroevers & Teo, 2008;Scrignaro, Barni, & Magrin, 2010). However, the results are somewhat mixed and none of the authors have used a standardized measure of humor, all assessing humor as part of a higher construct such as cognitive coping in "Coping Orientation to Problems Experienced" (COPE) and character strength in the "Values in Action-Inventory of Strength" (VIA-IS). To the best of our knowledge, the association between posttraumatic growth and sense of humor has never been investigated with a measurement that is only constructed to measure humor. The following studies will investigate the association between humor styles, posttraumatic growth, negative change following adversity, posttraumatic stress, and problems with regulating emotions.

Study one
Our first study examined the association among four humor styles, posttraumatic growth, and negative change following adversity. We hypothesized that benign humor is associated positively with posttraumatic growth and that injurious humor is associated negatively with posttraumatic growth. Two posttraumatic growth measurements were used. Although the Posttraumatic Growth Inventory (PTGI) is widely used and acknowledged, it could be prone to positive bias since it only measures positive change following adversity (Park & Lechner, 2006). Therefore, a second questionnaire that also accounts for negative change following trauma has been used. Furthermore, we expected a negative relationship between benign humor styles and negative changes following adversity. We assumed the reverse relationship for injurious humor.

Procedures and participants
Participants were recruited via a closed Internet survey. An Internet link was posted to six departments at a large English University. To increase the number of respondents, flyers were distributed and the survey link was sent to students who received feedback for another study conducted earlier. In total, around 2,900 students and members of the university received the survey invitation. Of those, 73 participants were included into the data analysis. Given the sample size and an alpha of 0.05, the power to find a medium effect (r ¼ 0.30 and rho ¼ 0.30; see Cohen, 1992) was 0.74 and 0.73 respectively. The response rate was around 2.51%. The study got ethical approval by the Ethics Board at the university.

Trauma
The study used an adapted form of the Traumatic Events Questionnaire by Vrana and Lauterbach (1994). The nature of the traumatic event was assessed by 12 items instead of 11 items. One item was included to enable the participants to negate the experience of a trauma. By adding this item, it was attempted to reduce a positive bias. Moreover, the added item served as a kind of filter to include only participants who had experienced a trauma. Three questions of the TEQ were used to evaluate on a 7-point Likert scale how traumatic the event has been (trauma severity). The average of all items was used to gain a total score. One item measured on a 7-point Likert scale how the trauma still affects the respective person (current distress). One variable measured how much time has been elapsed since the traumatic event.

Humor
The Humor Style Questionnaire (HSQ) by Martin et al. (2003) was used to measure sense of humor. The HSQ is a 32 item self-report measurement assessing sense of humor as a personality trait by a 7-point Likert Scale. The HSQ consists of four dimensions. Two factors are considered as rather psychologically benign while two factors are considered as rather psychologically injurious. The two adaptive factors are affiliative humor and self-enhancing humor. Affiliative humor is a social form of humor used in order to delight everyone (Martin, 2003). Self-enhancing humor does not necessarily fulfill a social function but may be better described as an internal process (Martin, 2003). This form of humor is based on the ability to change perspective and to gain pleasure out of absurdities when under stress (Martin, 2003). The two rather psychologically injurious humor styles are aggressive humor and self-defeating humor. Aggressive humor is directed toward the social environment in order to put other people down and to show one's own superiority . Self-defeating humor, in contrast, aims against oneself in order to gain acceptance from others or to avoid unpleasant feelings . Each humor style scale consists of 8 items. The range of each factor lies between 8 and 56.

Posttraumatic growth
Two measurements were used to measure posttraumatic growth. The first measurement used was the Posttraumatic Growth Inventory (PTGI) developed by Tedeschi and Calhoun (1996). The PTGI consists of 21 items, each of which is answered on a 6-point Likert scale. The PTGI can be used to assess five dimensions: new possibilities, personal strength, spiritual change, relating to others, and appreciation of life. In the current study, as we had no specific predictions regarding these subscales, only the total scale of the PTGI was be used. The second measurement used to measure posttraumatic growth was the positive scale of the "Changes in Outlook Questionnaire" (CiOQ-P) developed by Joseph, Williams, and Yule (1993). The CiOQ-P consists of 11 positive change items, each of which is answered on a 6-point Likert scale. The range of the positive change scale lies between 11 and 66 (Linley, Joseph, Cooper, Harris, & Meyer, 2003).

Negative change following adversity or trauma
The negative scale (CiOQ-N) of the CiOQ was used to measure negative change following adversity. The CiOQ-N consists of 15 negative change items, each of which is answered on a 6-point Likert scale. The range of the scale lies between 15 and 90 (Linley et al., 2003). The negative change scale and the positive change scale as well cover questions about future perspectives, finding meaning, acceptance of what has happened, valuing of life and relationships, and changes in attitudes.

Results and discussion
The sample consisted of 73 participants; the most frequently reported trauma was receiving news about the death or injury of a friend or relative (see Table 1). Except for the aggressive humor subscale, all scales were found to possess at least satisfactory internal consistency reliability (see Table 2).
All variables used were z-transformed. The variable negative change following adversity was not normally distributed. Hence, a nonparametric test was used for this variable. The correlations are shown in Table 3. Because the nature of the trauma varied to a significant degree, it could not be excluded that the results were influenced by the impact of the trauma. Therefore, we conducted partial correlations controlling for trauma severity and current distress. Generally, both variables did not affect the observed associations and they influenced no significant correlations.
Humor was not associated with posttraumatic growth measured by the PTGI. With the CiOQ, self-enhancing humor was associated positively with   positive changes. Self-enhancing humor was correlated negatively with negative changes, while self-defeating humor was positively associated with negative changes.
The results show that self-enhancing humor was associated with higher scores on positive changes, and lower scores on negative changes as measured by the CiOQ, but not with the PTGI. It may be that these different measurement tools assess different aspects of posttraumatic growth (Joseph & Linley, 2008;Park & Lechner, 2006). In summary, the results indicate that especially intrapsychic forms of humor may be associated with well-being outcomes (Cann et al., 2010). Whether this could also be the case for other traumarelated well-being outcomes will be tested within the next study.

Study two
We extended the perspective to other negative well-being measurements, namely, posttraumatic stress and problems regulating emotions. The latter outcome is not necessarily trauma-related. However, the regulation of emotions may often be problematic following a traumatic event (Wild & Paivio, 2003). Including such a measurement was thought to be helpful in order to account for such problems. We hypothesized that benign humor is associated positively with posttraumatic growth and that injurious humor is associated negatively with posttraumatic growth. We expected a negative relationship among benign humor styles, posttraumatic stress, and problems regulating emotions. We assumed the reverse relationship for injurious humor.

Procedures and participants
Participants were recruited conveniently via a pen-and-paper survey and through snowball sampling on the Internet. Around 482 subjects were approached. Of those, 132 participants experienced a trauma, were willing to fill in the survey, and gave their informed consent. Hence, the response rate was around 27%. The study got ethical approval by the Ethics Board at the University.

Measures
The questionnaires assessing humor, posttraumatic growth, and trauma were essentially the same as in study 1. However, because this study was part of a wider research project on reports of posttraumatic growth, the CiOQ has not been used. In addition, the trauma questionnaire allowed participants to indicate more than one trauma.
Posttraumatic stress. The Impact of Event Scale (IES) (Horowitz, Wilner, & Alvarez, 1979) measured posttraumatic stress. This self-report measurement consists of 15 items, each of which is answered on a 4-point Likert scale (0 ¼ not at all, 1 ¼ rarely, 3 ¼ sometimes, and 5 ¼ often). The IES consists of two subscales: intrusion and avoidance. The intrusion subscale consists of seven items and scores can range from 0 to 35. The avoidance subscale consists of eight items and scores can range from 0 to 40.
Emotion regulation difficulties. The Difficulties in Emotion Regulation Scale (DERS; Gratz and Roemer, 2004) was used to assess problems in regulating emotions. This 36-item scale was used because it builds on the theory that emotional regulation requires the awareness of emotions (Gratz & Roemer, 2004;Neumann, van Lier, Gratz & Koot, 2010;Wild & Paivio, 2003). The DERS can be used to assess six dimensions; however, in the current study only the total scale of emotion regulation difficulties was used. The score of the total scale of emotion-regulating difficulties ranges from 36 to 180.

Results and discussion
The sample consisted of 132 participants; the most frequently reported trauma was receiving news about the death or injury of a friend or relative (see Table 4). Except for the aggressive humor subscale, all scales were found to possess at least satisfactory internal consistency reliability (see Table 5).
All variables used were z-transformed. The correlations between variables are shown in Table 6. Given the sample size of the analyses (n ¼ 109) and an alpha of 0.05, the power to find a medium effect (r ¼ 0.30; see Cohen, 1992) was 0.89. We conducted partial correlations controlling for number of traumas, trauma severity, and current distress. Number of traumas and trauma severity did not significantly affect the observed associations. Some potential spurious effects were discovered for the association between affiliative humor, self-enhancing humor, intrusion, and avoidance when controlling for current distress.
Humor was not associated with posttraumatic growth as measured by the PTGI. However, benign humor was associated negatively with problems regulating emotions. Self-defeating humor was correlated positively with the avoidance subscale of the IES and problems regulating emotions.

General discussion
The primary aim of the paper was to test how sense of humor relates to several trauma-related well-being outcomes. Contrary to our hypotheses, the results of study 1 and study 2 found no significant association between any humor style and posttraumatic growth as measured by the PTGI. However, self-enhancing humor was associated positively with positive changes when measured by the CiOQ-P. These results support the assumption that different measurement tools may assess different aspects of posttraumatic growth (Joseph & Linley, 2008;Park & Lechner, 2006).  Self-enhancing humor may be linked to those dimensions of posttraumatic growth that are measured by the CiOQ-P, which may be tapping into more existential concerns than those of the PTGI.
The findings of the studies support our initial expectation that benign humor styles are correlated negatively with psychological difficulties. These results are in line with previous research that suggest a negative association between sense of humor and psychological problems (e.g., Besser et al., 2015;Edwards & Martin, 2010;Erickson & Feldstein, 2007;Sliter et al., 2013). Therefore, benign humor could potentially be an important factor that facilitates coping. However, it could also be the case that people who are less troubled by their traumatic experience are more likely to use humor. In our studies, especially self-enhancing humor was negatively associated with psychological problems. In line with Cann et al. (2010), this result could indicate that self-enhancing humor in particular could be useful in order to cope with adversity. While affiliative humor improves social interactions, self-enhancing humor may be an internal resource to approach and process adversity (Martin, 2003). However, a prospective study design would be needed to investigate this possibility in more detail.
As expected, self-defeating humor was correlated positively with negative changes in outlook following adversity, avoidant states, and problems regulating emotions. These results are in line with research conducted by Edwards and Martin (2010) and Erickson and Feldstein (2007), who found a positive association between self-defeating humor and depressive symptoms. In contrast to self-defeating humor that may be a way to escape or deny uncomfortable feelings, aggressive humor may be rather an instrument to gain power within a social context . Hence, aggressive humor may not be directly associated with well-being outcomes. In line with this theoretical assumption, aggressive humor was not significantly associated with any trauma-related well-being outcome.
The studies have several limitations. First, the sample size of study 1 was rather low. Therefore, associations below a value of r ¼ .23 could not become statistically significant. Second, it was not possible to indicate multiple traumas. Within study 2, participants were initially asked to state only the most traumatic event they have experienced. However, several participants stated more than one event and were unable to decide which of those was most traumatic. These circumstances led to the decision to allow for multiple answers. Because of this procedure, it is likely that the true number of experienced traumas has been underestimated. However, the number of traumas had no significant influence on the observed associations. The sample was not representative for both studies. Moreover, both studies could have been prone to sampling bias. This may be especially the case for study 2, where two different survey modes were used. However, correlations are less prone to selection bias (Heiervang & Goodman, 2011;Loewenthal, 1996).
In addition, both studies were correlational and, therefore, causal inferences cannot be drawn. Finally, the response rate for both studies was low.
In conclusion, the results suggest that especially intrapsychic forms of humor may be related to negative as well as positive well-being outcomes. Therefore, these results indicate the need for prospective studies investigating the potential buffering role of humor following trauma. This may especially hold true for intrapsychic forms of humor.

Notes on contributors
Michaela Boerner completed her PhD on posttraumatic growth in the School of Education at the University of Nottingham (UK).
Stephen Joseph is the cluster convenor for counseling and psychotherapy teaching and training in the School of Education at the University of Nottingham (UK).
David Murphy is currently the Program Leader for the MA Person-Centered Experiential Psychotherapy and Counseling course in the School of Education at the University of Nottingham (UK).