Does Humor Influence the Stigma of Mental Illnesses?

Abstract

Public stigma is a barrier for people with mental illness. Humor may have the potential to decrease stigmatizing attitudes in the context of disclosure. Participants completed measures on stigmatizing attitudes and humor style and were then randomized to one of three conditions (self-disclosure comedy sketch, the same comedy sketch with no disclosure, and a control comedy sketch). After reviewing the comedy sketch, participants repeated the attitude measures and provided perceptions of the comic. Humor styles and perceptions significantly interacted with condition to reduce stigma. Perceptions of the self-disclosed comic were associated with reduced stigma. People exhibiting affiliative humor style (i.e., they enjoy making others laugh) were shown to have significantly greater stigma changes in the disclosed condition compared to the non-disclosed and control conditions. Affiliative humor endorsers also interacted with the non-disclosed condition suggesting that mental health comedy might generally reduce stigma in people who use humor to improve relationships.

Public stigma – that is, the population's endorsement of prejudice about mental illness – is a barrier to the personal goals of many people with psychiatric illness including opportunities related to good jobs, satisfactory living arrangements, and quality health care (Sartorius et al., 2012). Public stigma has been defined as stereotypes about mental illness that lead to discrimination (Corrigan, 2005). One stereotype is blame: the belief that people “choose” to have a mental illness. According to attribution theory, people who are perceived as responsible for a condition (such as their mental illness) are disparaged and shunned (Corrigan, 2000; Weiner, 1985). Given stigma's pernicious effects, advocacy groups have developed a variety of anti-sigma programs meant to reduce its egregious impact. Results of a recent meta-analysis identified 71 empirical studies, with 13 using randomized controlled trials that showed contact-based approaches (i.e., where people with lived experiences share stories of recovery with the public) yielded the best effect sizes (Corrigan et al., 2012). One group, Stand Up for Mental Health, uses comedy to enhance its impact on stigma change. Stand Up for Mental Health, founded in 2004, is an organization of Canadians with mental illness who use comedy as a way to build confidence and fight stigma.

Humor might seem to be an unlikely aid in reducing stigma. Research fairly consistently shows humor and comedy are often associated with endorsing negative stereotypes about social groups (Ford & Ferguson, 2004; Maio, Olson, & Bush, 1997; Thomas & Esses, 2004) with the relationship between jokes and sexism especially well-documented (Ford, Boxer, Armstrong, & Edel, 2008; Greenwood & Isbell, 2002). The connection between humor and prejudice has been extended to people with disabilities in general (Anderson, 1988) and those with mental illness (Németh, 2009). However, humor does not always seem to exacerbate stigma; for example, a randomized controlled trial on the effects of Starved, a 2005 situation comedy about four people with eating disorders, showed no worsening in stereotypes (Katterman & Klump, 2010; Shaeffer, 2005). Theorists have wondered whether humor can be used to assuage the shame of stigma related to health conditions (Ruiz-Mellott, 2004), an assumption supported in a qualitative study with more than 250 service providers for people with HIV-AIDS showing gallows humor as an effective approach for dealing with the related self-stigma (Makoae et al., 2008). Social scientists have opined that some television shows have lessened the stigma of homosexuality citing Will and Grace, a 2001 to 2005 comedy about gay lawyer, Will Truman (Castiglia & Reed, 2004), and the 1997 comedy, Ellen, where character Ellen Morgan disclosed she was a lesbian shortly after DeGeneres publically came out (Ryan & Boxer, 1998). The last example is especially relevant to the goals of this study; how does humorous discussion of one's own “stigmatized condition” affect public stigma? We were unable to find any research that tested this kind of disclosure in the empirical literature.

In this study, we examine the impact of Stand Up for Mental Health founder David Granirer's two-minute YouTube routine where he discusses experiences with his illness and the public's reaction to the introduction of a local mental health clinic. We sought to distinguish the effects of his first person story from the humor per se. Hence, people were randomized to Mr. Granirer's routine with or without self-disclosure; i.e., knowing the comic did or did not have experience with mental illness himself. We hypothesize the self-disclosing sketch will significantly reduce mental illness stigma in viewers, consistent with the assumptions about TV shows discussing gay and lesbian subject matter referenced above. However, stripping out disclosures will translate the piece to humor alone, which we think will undermine any benefits. Authenticity of the comic, operationalized as perceiving a comic as knowing about what he speaks because of his lived experience, will likely be an important moderator in anti-stigma effects. We hypothesize that comics viewed as having a mental illness will seem more authentic and hence yield bigger reductions in stigma. Similarly, we expect sketches where the comic is viewed as funnier, more similar to the viewer, and more sincere to yield greater impact for the self-disclosure condition.

Research suggests comedy's effects are impacted by humor style, defined as individual differences in the way people react to humor that vary across several domains (Hodson, Rush, & MacInnis, 2010a). For example, those who endorse an aggressive style (“I often use humor or teasing to put others down.”) may perceive slapstick humor as funny and support notions of social dominance (“It is a good thing that some groups are at the top and others at the bottom.” [Hodson Rush, & MacInnis, 2010b]). We hypothesize that effects of comedy sketches will interact with a person's individual humor styles. Two styles are incorporated into this study. (1) Aggressive style, those higher in aggressive style will show less benefit from the authentic comic. (2) The opposite of aggressive, an affiliative style (“I enjoy making others laugh.”) will lead the viewer to empathizing with the comedian causing greater reduction in stigma.

Methods

Research participants completed pre-tests of stigmatizing attitudes and humor style and were then randomized to one of three conditions: a comedy sketch where the comic discloses his mental illness, the same sketch where all allusion to the comic's mental illness is omitted, and a control comedy sketch with no reference to mental illness. Research participants then repeated the measures of stigmatizing attitudes and reported their perceptions of the comic. Recruitment was conducted via advertising on the Chicago, Illinois Volunteers section of Craigslist (www.craigslist.org/chicago/volunteers). Participants accessed the study via hyperlink to a secure online survey site called Qualtrics (www.qualtrics.com). The only exclusion criteria was age; only adult participants (age 18 years or older) were permitted to participate in the study. Participants were informed this was a study about humor and mental illness and were reimbursed with a $5.00 Amazon gift card for completing research tasks. An overall concern of online research like this is keeping the task relatively brief so subjects do not prematurely terminate. Hence, short versions of tests and single item proxies of our constructs were selected; the total task took about 15 minutes in all.

To assess public stigma in terms of blame and attributions, we used the short form of the Attribution Questionnaire (AQ); in the measure, participants are presented a short vignette about Harry, a 35-year-old person with schizophrenia who lives alone and works full time as a clerk at a legal firm. In the full AQ (Corrigan et al., 2003; Reinke et al., 2004), participants answer 27-items (e.g., “I think that it was Harry's own fault that he is in the present condition.”) on a 9-point agreement scale (9=very much). The AQ-27 has strong reliability, construct validity, and sensitivity to change after anti-stigma program participation. Given that the original AQ is long for many outcome studies, a short form with nine-items was tested and found to have similar strong reliability, validity, and sensitivity to change (Corrigan et al., 2004; Corrigan et al., 2013). A total score from the AQ is used in this study.

We used items from the Humor Styles Questionnaire (HSQ: Martin et al., 2003) to assess this construct. The HSQ comprises items to which participants respond using a nine-point agreement scale (1=extremely disagree; 9=extremely agree). We selected items that represent aggressive and affiliative humor. People were randomized to one of three comedy sketches (about 2 minutes each). The index sketch by Stand Up for Mental Health founder, David Granirer, has more than 12,000 YouTube views. In this disclosed comic sketch, Granirer shares his ironic perceptions about the “threats” of a mental health center opening up in the community. The sketch begins with Granirer disclosing his mental illness and experiences with symptoms and medication. This condition is contrasted to a non-disclosed comic condition where Mr. Granirer does the same material except all reference to his mental illness is omitted. The control condition consisted of a two-minute sketch where a gender, race, and age-matched comic discussed his lack of understanding women's fashion. There was no mention of issues related to mental health in the control condition.

Participants completed the AQ prior to and immediately after the comic sketch. They also completed five statements about the comic on a 9-point agreement scale (9=extremely agree). Statements assessed participant perceptions about the comic being: humorous, aggressive, with/without a mental illness, “like me,” and sincere. Research participants also completed demographic questions.

Statistical Plan

First, we examine frequencies and means (standard deviations) of demographics, humor styles, and perceptions of comics across the three research conditions. We then examine change in AQ scores from pre to post assessment to test the hypothesis about disclosed and non-disclosed comedy sketch effects on stigma change. Before examining moderating effects of humor styles and perceptions of comics, we provide Pearson Product moment correlations to describe their association with AQ scores. We then tested moderator effects as planned multiple regression analyses representing the relationship between post-test AQ scores and intervention-humor style (or perceptions of comic) interactions after partialling out baseline AQ scores. Intervention represents two planned contrasts of conditions: disclosure versus no disclosure and control (1,0,0) or no disclosure versus disclosure and control (0,1,0). Hence, Table 3 represents a series of equations of the form

$$\begin{array}{c}\hfill {\mathrm{AQ}}_{\mathrm{post}}={\mathrm{AQ}}_{\mathrm{pre}}+{\mathrm{contrast}}_{(1,0,0)}\mathrm{X}\mathrm{humor}\_{\mathrm{style}}_{\mathrm{n}}+{\mathrm{contrast}}_{(0,1,0)}\mathrm{X}{\mathrm{humorstyle}}_{\mathrm{n}}\text{or}\hfill \\ \hfill {\mathrm{AQ}}_{\mathrm{post}}={\mathrm{AQ}}_{\mathrm{pre}}+{\mathrm{contrast}}_{(1,0,0)}\mathrm{X}\mathrm{comic}\_{\mathrm{perception}}_{\mathrm{n}}+{\mathrm{contrast}}_{(0,1,0)}\mathrm{X}\mathrm{comic}\_{\mathrm{perception}}_{\mathrm{n}}.\hfill \end{array}$$

Results

Examination of AQ distributions revealed seven participants showing significant floor effects; i.e., endorsing almost no stigmatizing attitudes. Hence, these outliers were removed from additional analyses reducing N fairly evenly across conditions to 342. The total sample was 70.8% male and ethnically diverse. They were 65.8% Caucasian, 20.8% African American, 9.5% Asian, and 2.9% Native American. Participants reported themselves to be 45.3% Hispanic or Latino. It was a fairly educated group with 48.5% reporting at least some college, 32.7% having earned a bachelor's degree, and 3.2% a graduate degree. Almost all participants (95.9%) reported currently working full time. Table 1 summarizes demographics across conditions where no significant differences were found. The Table also summarizes humor styles and perceptions of the comic across conditions, again with no significant differences emerging. Finally, Table 1 lists pre and post AQ scores by condition with nonsignificant main effects and interaction. These last analyses failed to show a humor effect on stigma change.

Table 1.

Differences in demographics, humor style, comic ratings, and stigma measures by condition. Frequencies (%s) are reported for categorical variables.

Variable	Sketch with disclosed comic n=117		Sketch with non-disclosed comic n=108		Control sketch n=117		Group differences
	M	SD	M	SD	M	SD
DEMOGRAPHICS
Gender female %	35.9%		28.7%		23.1%		χ2(2)=4.67, n.s.
Age	33.2 6.8		32.5 5.4		32.9 7.4		F(2,339)=.34, n.s.
Race
Caucasian	69.2%		59.3%		68.4%		χ2(8)=8.73, n.s.
African	17.1%		25.0%		20.5%
American	1.7%		5.6%		2.6%
Asian	6.8%		8.3%		6.8%
Native American Pacific Islander	5.1%		1.9%		1.7%
Hispanic Ethnicity	47.0%		44.4%		44.4%		χ2(2)=.20, n.s.
Education
Some HS	0.0%		0.9%		0.9%		χ2(12)=10.2, n.s.
HS diploma	1.7%		2.8%		0.0%
AA/AS	6.0%		13.9%		13.7%
Some college	54.7%		43.5%		47.0%
BA/BS	32.5%		33.3%		32.5%
Grad degree	2.6%		3.7%		3.4%
Other	2.6%		1.9%		2.6%
Employment							χ2(2)=.22, n.s.
Full time	95.7%		95.4%		96.6%
Part time	4.3%		4.6%		3.4%
HUMOR STYLE
Affiliative	1.99	1.14	2.07	1.05	2.09	1.03	F(2,339)=.26, n.s.
Aggressive	2.66	2.06	2.72	2.13	2.78	2.13	F(2,339)=.10, n.s.
PERCEPTIONS OF COMIC
Humorous	2.97	2.28	3.28	2.25	3.18	2.28	F(2,339)=.56, n.s.
Aggressive	2.82	2.08	2.79	2.02	2.79	1.97	F(2,339)=.01, n.s.
Has mental illness	2.85	2.10	2.90	2.10	2.77	1.95	F(2,339)=.11, n.s.
Like me	2.50	1.97	3.06	2.33	2.87	2.03	F(2,339)=2.11, n.s.
Sincere	2.46	1.67	2.85	1.95	2.71	1.93	F(2,339)=1.29, n.s.
ATTRIBUTION QUESTIONNAIRE
Pre	55.3	8.1	53.0	9.6	54.6	9.1	Condition: F(2,339)=1.35, n.s. Trial: F(1,339)=0.93, n.s.
Post	54.2	11.2	52.9	10.4	54.7	10.0	Interaction:F(2,339)=0.43, n.s.

Table 2 summarizes relationships among AQ pre and post-test scores, humor styles, and perceptions of the comic. Both pre and post AQ scores were negatively related to affiliative humor. The more participants enjoy humor, the less likely they were to endorse stigmatizing attitudes. Effect sizes for these relationships were greater than 12.5%. Although affiliative and aggressive humor were significantly inversely associated (r=-0.22, p<.001), aggressive humor was not associated with pre or post AQ scores. Table 2 also summarizes the relationship between pre and post-test AQ scores and perceptions of the comic. All of these relationships were significant; participants viewing comics as having a mental illness and relatively more humorous, aggressive, “like me,” and sincere showed diminished stigma on the AQ. Effect sizes for these relationships range from 4.8 to 22.1%. Finally, an affiliative humor style was found to be significantly associated with all perceptions of the comic. An aggressive humor style was significantly related to only viewing the comic as humorous or aggressive.

Table 2.

Pearson product moment correlations among AQ pre test scores, humor styles, and perceptions of the comics.

	HUMOR STYLES		PERCEPTIONS OF THE COMIC
	Affiliative	Aggressive	Humorous	Aggressive	Has MI	Like me	Sincere
AQ pre	−.36***	.04	−.31***	−.41***	−.47***	−.29***	−.45***
AQ post	−.41***	.03	−.32***	−.42***	−.45***	−.22***	−.44***
Affiliative		−.22***	.15**	30***	.20***	.19***	.30***
Aggressive			−.28***	−.11*	−.09	−.08	−.06
Humorous				.23***	.38***	.20***	.31***
Aggressive					.58***	.30***	.47***
Has MI						.34***	.47***
Like me							.42***

Note.

^*p<.05

^**p<.01

^***p<.001

Even though we found no significant group by trial interactions suggesting no stigma change due to comic condition, we examined interaction effects between individual conditions and measures of humor styles or perceptions of the comic to independently test these hypotheses. Table 3 provides the results of multiple regression analyses. The first two rows examined the interaction of humor style with contrasted comic sketch conditions. Results for an affiliative humor style showed significant betas for both interactions. People assigned to the disclosed comic and the non-disclosed comic condition showed significant reductions in post-test stigma on the AQ as they endorsed greater affiliative humor. Significant effects for the interactions of aggressive humor and anti-stigma interventions were not found with either condition.

Table 3.

Results of multiple regression analyses examining relationship between post-test scores, pre-test scores, and interactions (planned contrasts X comic rating or humor style). All values are betas (standardized coefficients).

		----------------------------------------------INTERACTIONS----------------------------------------------
EQ	AQ-pre	humor style or comic perception X	PLANNED CONTRAST		R2
			disclosed VERSUS non-disclosed,control (1,0,0)	Non-disclosed VERSUS disclosed,control (0,1,0)
HUMOR STYLES
1	.74***	affiliative	−.13**	−.09*	.571
2	.74***	aggressive	−.02	−.01	.557
PERCEPTIONS OF COMIC
3	.74***	humorous	−.09*	−.01	.563
4	.74***	aggressive	.08*	.04	.563
5	.74***	has MI	−.09*	−.02	.564
6	.74***	like me	−.08*	−.02	.562
7	.74***	sincere	−.09*	−.08*	.565

Note.

^*p<.05

^**p<.01

^***p<.001

The bottom half of Table 3 summarizes interactions between anti-stigma comic sketch conditions and perceptions of the comic. Perceptions of the comic when the comic self-disclosed, were associated with greater reductions in stigma from pre to post. This occurred when the comic was viewed as humorous, having a mental illness, “like me,” and sincere. The self-disclosed comic who was viewed as less aggressive also showed significantly better change in stigma on the AQ.

Discussion

Humor is used by some advocates with lived experience as a strategy to diminish public stigma towards mental illness. In this study, we attempted to examine the effects of humor on stigma change when the comic had or had not publicly disclosed his mental health status. Results failed to show a significant change in stigma due to either of the two comic sketch conditions. Given the relatively short time between pre and post test, an AQ memory effect or social desirability effect may have undermined significant change. Alternatively, comic effects may only occur when examined by moderators such as humor style and perceptions of the comic. In fact, significant stigma change was found when examining the interaction of comic conditions with humor style and perceptions of the comic. Stigma as measured on the AQ was shown to significantly decrease for the disclosed comic condition among participants who endorsed an affiliative humor style. People who seem to enjoy humor and making people laugh were shown to have significantly larger changes in the disclosed comic condition compared to the non-disclosed and control condition. An affiliative style was also found to interact with the non-disclosed comic condition which suggests comedy about mental health in general might reduce stigma in a person who enjoys laughing. An aggressive humor style was not found to interact with either comic condition, a notable finding that suggests offense at a comic's statement about mental illness does not seem to impede attitude change.

The sketch's effect on attitude change also significantly interacted with perceptions of the comic. These effects were only found for the comic who self-identified with a mental illness. Participants who viewed the disclosed comic as more humorous yielded greater reductions in stigmatizing attitudes. A specific hypothesis of the study was that the comic viewed as more authentic would produce greater stigma reduction effects. Results showed the disclosed comic who was viewed as having a mental illness, “like me,” and sincere yielded better stigma change. Although an aggressive humor style was not found to interact with comic condition, perceiving the comic as aggressive did interact with condition. Namely, perceiving the disclosed comic as aggressive or offensive led to less change in stigma.

These findings suggest researching comedy about mental health person-level variables needs to be considered in crafting these approaches. We began with humor styles here as natural moderators of the way comedy routines are experienced. There are additional humor styles in this area that might be relevant including self-defeating (e.g., “I get carried away in putting myself down.”) and self-enhancing (e.g., “My humor keeps me from getting overly depressed about things.”) humor (Martin et al., 2003). These might be especially relevant as the person weighs his or her own personal experiences with mental illness when viewing the comedy sketch. In this light, one's familiarity with mental illness is likely to mediate the comedy routine. One study showed familiar experiences with mental illness (e.g., having a family member with mental illness) mediated reductions in stigma that occurred after an education about stigma (i.e., contrasting myths and facts of mental illness) program (Holmes et al., 1999). Interestingly, those with greater familiarity showed significantly less improvement. The study also examined moderating effects of mental health literacy in the education program and found those who were more knowledgeable about mental illness were also less likely to decrease stigma after the education program. Both of these findings might reflect a floor effect; those personally familiar with and knowledgeable about mental illness are already likely at the floor of stigma measures with little room to improve. Regardless, future studies need to include measures of person-specific variables in their design.

Limitations of the study need to be considered when interpreting findings. Failure to support the main hypotheses per se is the most significant. We were unable to directly determine whether comedy and comic disclosure impact stigma. In part, this might have occurred due to selection of the study measure. The 9-item version of the AQ is much shorter than the original 27-item, which might have yielded greater power. Despite this, showing significant interaction effects definitely suggest a positive role for comedy in stigma change. Findings were also limited to decreasing stigma rather than increasing its obverse, affirming attitudes (Corrigan et al., 2013). Future research should examine the effects of comedy on positive attitudes about mental health, including recovery, self-determination, and personal empowerment. An additional limitation is that only pre and post-tests were examined in the study. We were unable to determine whether a two-minute comedy sketch yielded any changes that maintained over time. Future studies should include follow-up assessment.

This research may have implications for future humor programs attempting to change stigma. Perceptions about the comic were especially important moderators with sincere and authentic comics seeming to have greater impact. These findings support assertions of Stand Up for Mental Health that authentic comics talking about their experiences with mental health can be a powerful vehicle for stigma change. What other variables might add to this poignancy? Future research needs to examine the relative impact of videotaped versus in vivo sketches. Findings from the meta-analysis on stigma change programs found face-to-face contact leads to the greatest change (Corrigan et al., 2012). Hence, in vivo comedy sketches are expected to yield better stigma reduction. Still, videotaped and online contact improved stigma significantly, albeit less than face-to-face contact. Videotaped comedy has the potential of penetrating a much bigger audience than in vivo approaches. Hence, future research needs to consider the breadth of impact of different kinds of comedy routines.