Humor is a fundamental human activity enjoyed daily by individuals of all ages and in all cultures (Martin & Ford, Reference Martin and Ford2018). The ability to appreciate and create humor is universal, relying on mechanisms such as surprise and incongruity during playful social interactions, which elicit humor and laughter across the globe (Apte, Reference Apte1985; Gervais & Wilson, Reference Gervais and Wilson2005). Humor serves as a unique communication tool, breaking interpersonal barriers, easing tensions, and connecting people in various social situations.
Researchers agree that humor has an evolutionary basis, although its adaptive function remains debated (Greengross & Kozbelt, Reference Greengross, Kozbelt, Chłopicki, Ford and Kuipers2024). Smiling and laughter are not unique to humans; they are observed in many other primates, rats, and possibly even dolphins (Gamble, Reference Gamble2001; Gervais & Wilson, Reference Gervais and Wilson2005; Maglieri et al., Reference Maglieri, Vantaggio, Pilenga, Böye, Lemasson, Favaro and Palagi2024; Palagi et al., Reference Palagi, Caruana and de Waal2022; Panksepp & Burgdorf, Reference Panksepp and Burgdorf2003; Preuschoft & Van-Hooff, Reference Preuschoft, Van-Hooff, Segerstrale and Molnar1997). Smiles and laughter develop spontaneously early in life, uninfluenced by culture, and are characterized by stereotypical facial expression and vocalization that are universally recognizable (Bergen, Reference Bergen and Ruch1998; Ekman, Reference Ekman1993; Provine, Reference Provine2000). In other species, smiling and laughter occur in playful or friendly teasing contexts, resembling humans’ joking relationships, and are considered antecedent to humor in humans (Eckert et al., Reference Eckert, Winkler and Cartmill2020). Although humans do not always smile or laugh in reaction to humor, these behaviors are strongly linked to humorous interactions.
The ability to produce high quality humor is considered one of the most highly valued human attributes and lies at the heart of many evolutionary explanations for humor. High humor production ability (HPA) — the capacity to craft witty remarks, generate funny ideas, and elicit laughter — is socially attractive and is highly beneficial in various social interactions. Its potential evolutionary benefits include navigating and resolving conflicts in high-tension situations, facilitating social bonding, promoting cooperation, signaling in-group affiliation, attracting mates, and sustaining long-term romantic relationships and friendships (Bryant et al., Reference Bryant, Fessler, Fusaroli, Clint, Aarøe, Apicella, Petersen, Bickham, Bolyanatz and Chavez2016; Chafe, Reference Chafe1987; Gervais & Wilson, Reference Gervais and Wilson2005; Greengross, Reference Greengross2014, Reference Greengross2021; Greengross & Kozbelt, Reference Greengross, Kozbelt, Chłopicki, Ford and Kuipers2024; Greengross & Miller, Reference Greengross and Miller2011; Kaufman et al., Reference Kaufman, Kozbelt, Bromley, Miller, Geher and Miller2008). Possessing a great sense of humor in each of these scenarios provides adaptive advantage, either directly or indirectly, by contributing to survival and reproductive success. Most people recognize the importance of humor and its associated social and mating benefits, which may explain why so many believe they have an above-average sense of humor (Chapman & Gadfield, Reference Chapman and Gadfield1976; Fine, Reference Fine1975; Omwake, Reference Omwake1937).
Despite humor’s importance, relatively little is known about the sources of individual differences in humor ability or what contributes to developing a great sense of humor. The current study aims to address this gap and evaluate, for the first time, the relative contributions of genetic and environmental factors to HPA. Previous research has examined hereditary and environmental influences on other aspects of humor, such as humor appreciation and humor styles. These studies generally find a genetic component in humor, with low additive genetic contributions to humor appreciation (0.15−0.28), and moderate heritability for aggressive and affiliative humor styles — humor used to enhance relationships with others (0.30−0.50; Baughman et al., Reference Baughman, Giammarco, Veselka, Schermer, Martin, Lynskey and Vernon2012; Vernon, Martin, Schermer, Cherkas, et al., Reference Vernon, Martin, Schermer, Cherkas and Spector2008; Vernon, Martin, Schermer, & Mackie, Reference Vernon, Martin, Schermer and Mackie2008; Weber et al., Reference Weber, Ruch, Riemann, Spinath and Angleitner2014; Wilson et al., Reference Wilson, Rust and Kasriel1977).
However, HPA is a distinct cognitive ability that is largely independent of other humor-related traits (Feingold & Mazzella, Reference Feingold and Mazzella1993; Heintz, Reference Heintz2023; Kohler & Ruch, Reference Kohler and Ruch1996; Moran et al., Reference Moran, Rain, Page-Gould and Mar2014; Ruch & Heintz, Reference Ruch, Heintz, Luria, Baer and Kaufman2019; Silvia et al., Reference Silvia, Cotter and Christensen2024). A great sense of humor is not strongly correlated with how much a person laughs, their enjoyment of comedy movies and shows, or their general cheerfulness. Large individual differences exist in HPA, ranging from stand-up comedians’ exceptional humor creativity to individuals who rarely make others laugh or even attempt say anything funny. The question remains: how much of the variation in HPA can be differentially attributed to genetic factors and environmental influences?
Studying the heritability of HPA can also add to the discussions of the heritability of cognitive abilities more broadly. There is ongoing debate about the sources of cognitive abilities, their formation, and the degree to which biological and environmental factors play a role in their development (Plomin et al., Reference Plomin, DeFries, Knopik and Neiderhiser2013). Many studies suggest that substantial individual variability in cognitive abilities can be attributed to genetic differences, with heritability estimates for intelligence and creativity ranging from approximately 0.40 to 0.80 (Haier et al., Reference Haier, Colom and Hunt2023; Haworth et al., Reference Haworth, Wright, Luciano, Martin, de Geus, van Beijsterveldt, Bartels, Posthuma, Boomsma and Davis2010; Piffer & Hur, Reference Piffer and Hur2014; Polderman et al., Reference Polderman, Benyamin, De Leeuw, Sullivan, Van Bochoven, Visscher and Posthuma2015; Turkheimer et al., Reference Turkheimer, Haley, Waldron, d’Onofrio and Gottesman2003). Given HPA’s moderate correlation with intelligence and creativity (e.g., 0.22 with fluid intelligence, 0.49 with crystallized intelligence, and a stronger positive association with verbal tasks than with abstract reasoning tests), HPA is expected to exhibit moderate heritability as well (Christensen et al., Reference Christensen, Silvia, Nusbaum and Beaty2018; Greengross & Miller, Reference Greengross and Miller2011; Kellner & Benedek, Reference Kellner and Benedek2017).
Additionally, this research aims to explore an area yet to be studied, namely the genetic and environmental contributions to the relationship between humor ability and health. Humor is widely believed to positively impact physical and mental health (e.g., Cousins et al., Reference Cousins, Monat and Lazarus1991), though findings have been mixed (Greengross & Martin, Reference Greengross and Martin2018; Martin, Reference Martin2001; Martin, Reference Martin and Raskin2008). Identifying a genetic basis for the connection between humor ability and health could illuminate humor’s evolutionary role and clarify biological and environmental influences on cognition and health. Understanding this relationship could also advance knowledge of cognitive decline, as humor ability changes with aging (Greengross, Reference Greengross2013).
To achieve these objectives, the study utilizes the twin study paradigm, the most common method for estimating genetic and environmental contributions (Segal, Reference Segal2000, Reference Segal2017). Based on previous research on the heritability of intelligence and creativity, we hypothesize that individual differences in HPA will be influenced by additive genetic factors and nonshared environmental factors, with a smaller contribution from the shared environment (Harris, Reference Harris2009; Plomin, Reference Plomin2011; Plomin & Daniels, Reference Plomin and Daniels1987; Turkheimer & Waldron, Reference Turkheimer and Waldron2000). Furthermore, we hypothesize that the correlations between humor ability and both health and intelligence will have a moderate genetic basis.
Materials and Methods
Sample
Participants were recruited from the Twins UK registry, the most clinically detailed adult twins’ registry in the UK. In the summer of 2022, active twins in the registry with email addresses were approached and offered the opportunity to participate in the study. Twins and their co-twins were instructed to complete the study separately and not discuss it until both had finished. Of the 17,100 individual twins registered, 3332 took part in the study (about 19%). Only monozygotic (MZ) and dizygotic (DZ) same-sex pairs in which both twins completed all measures were included in the study. After removing mixed-sex twins (n = 19), individuals of unknown zygosity (n = 130), cases where at least one twin did not complete all measures (n = 14), and cases with an anomaly recorded (n = 4), the final sample included 1363 twins (89.4% women), age range 21−89 (mean = 61.5, SD = 14.6; median = 66). In total, 196 complete DZ pairs and 448 complete MZ pairs were included in the final analyses. Data from singletons (incomplete pairs) were included for 8 DZ individuals and 18 MZ individuals, as they contributed to estimation of the mean, but not to the variance decomposition.
Measures
Health
Twins were asked to assess their general physical health on the day of the study using a 6-point scale: Excellent, Very Good, Good, Fair, Poor, and Don’t know/Prefer not to answer. Responses were converted to a Likert-type scale from 1−5, excluding nonanswers. The question was designed to provide a subjective estimate of health.
Cognitive ability
The Cognitive Reflection Test (CRT) is a three-question measure often used to estimate cognitive ability (Frederick, Reference Frederick2005). Each question presents an intuitive, but incorrect answer, requiring effortful thought to identify the correct response. Participants received 1 point for each correct answer, resulting in a total of 0−3. CRT scores are strongly correlated with intelligence, particularly with verbal and numerical abilities, working memory, mechanical-spatial ability, and numeracy (Otero et al., Reference Otero, Salgado and Moscoso2022).
Humor production ability: Objective method
While no universally agreed method exists for measuring humor creativity, researchers frequently assess it using varied procedures (Greengross et al., Reference Greengross, Silvia and Nusbaum2020; Ruch & Heintz, Reference Ruch, Heintz, Luria, Baer and Kaufman2019; Silvia, Rodriguez et al., Reference Silvia, Cotter and Christensen2024). One robust and reliable approach for measuring spontaneous humor involves asking participants to create humorous captions for cartoons that have no captions. This method, based on The New Yorker cartoon caption contest, has been widely employed in many studies (Greengross & Miller, Reference Greengross and Miller2011; Kellner & Benedek, Reference Kellner and Benedek2017; Mickes et al., Reference Mickes, Walker, Parris, Mankoff and Christenfeld2012).
For this study, we used two cartoons from previous cartoon contests. In the first, a restaurant customer places an order before a wet bear acting as a waiter. In the second, a real estate agent shows a couple a house located in space, with planets visible through the window. Twins were asked to write one funny caption for each cartoon. There was no time limit, but twins were encouraged to spend some time considering their responses.
Because of the large number of responses, a planned missing rating design was used (Forthmann et al., Reference Forthmann, Goecke and Beaty2023; Primi et al., Reference Primi, Silvia, Jauk and Benedek2019). After data collection, captions were grouped into five blocks, each containing approximately 500 participants and their captions. Twin pairs were always placed in the same block. Forty judges (27 women, 13 men) from Aberystwyth University and the surrounding area rated the captions. Each block was rated by 5−10 judges. Judges rated captions on a scale from 1 (not funny) to 5 (very funny) and were instructed to maintain consistency across ratings (i.e., rate similar responses at similar funniness levels). Each judge assessed both captions for a given participant, but only reviewed a subset of the total dataset to minimize effects of fatigue, rater drift, and reduced sensitivity to humor. To avoid biases, judges were blind to any characteristic of the twins (e.g., sex, age, ethnicity), and did not know that participants were twins. In the next step, three additional ‘anchor judges’ assessed a sample of 20 participants from each block, rating both captions. Captions in the sample were chosen to represent the full range of funniness ratings, as established by the original 40 judges. These anchor ratings ensured that the ratings were psychometrically connected across blocks (see Eckes, Reference Eckes2023; Wind & Stager, Reference Wind and Stager2019).
We applied the Many-Facet Rasch Model (Eckes, Reference Eckes2023; Engelhard Jr, & Wind, Reference Engelhard and Wind2017) to derive an IRT-based humor ability score for each twin. This model-based approach recognizes the faceted design (items by raters) and estimates a humor ability score for each person that is appropriately adjusted for the facets: possible differences in the ‘difficulty’ of the items they completed (e.g., it might be easier to generate funny responses for some items than for others), and, crucially for rater-mediated assessment, the ‘severity’ of the judges who rated their responses (e.g., some raters might be more lenient, tending to give higher humor scores, and others might be more severe, tending to give lower scores). The analyses were conducted in R using TAM (Robitzsch et al., Reference Robitzsch, Kiefer and Wu2024). This model yields an overall latent humor ability score for each respondent based on the judges’ ratings for both items. The reliability for this global ability score, expressed as expected a posteriori (EAP) reliability of the latent variable, estimated by the IRT model, was .760, which as expected was higher than the reliability for the scores for each individual cartoon (Cartoon 1: .697, Cartoon 2: .682). We used the overall latent humor ability score based on the faceted design for all subsequent analyses.
Subjective method
Twins were also asked to self-rate how funny they believe they are on a scale from 1 (not funny at all) to 10 (very funny) and how often they make others laugh using the same scale. The two ratings were then averaged to create a measure of self-rated humor.
Co-twin perception
Twins were asked to rate their co-twin’s humor using the same two questions: how funny they believe their co-twin is (on a scale from 1 to 10) and how often their co-twin makes others laugh (on the same scale). The two items were then averaged to create a measure of co-twin perceived humor. This variable represents how twin B perceives and rates twin A’s humor, which was independently measured through self-rated humor.
Statistical Analyses
To describe the data, we examined cross-trait individual level correlations to evaluate the degree to which measures of health, humor, and cognitive ability were related.
To quantify the sources and magnitude of variance in study outcomes, we utilized within-trait twin correlations stratified by zygosity and twin variance decomposition models. These models partition trait variance and covariance into additive genetic (A), shared environmental (C), genetic dominance (D), and/or unique environmental (E) sources. The choice of components (ACE, ADE, AE, and/or E only) was determined by evaluation of the twin correlations and model comparisons via the likelihood ratio test. We utilized the direct symmetric parameterization (Verhulst et al., Reference Verhulst, Prom-Wormley, Keller, Medland and Neale2019) to avoid upwards bias in parameter estimates. In the twin models, CRT was treated as an ordinal variable and self-rated health, self-rated humor, and latent humor ability scores were all treated as continuous variables.
Results
Table 1 presents the descriptive statistics for all outcome variables, the sex differences among them, and the effect sizes. Men scored significantly higher than women on self-rated humor, co-twin rated humor, objective humor ability, and CRT. There was no significant sex difference in self-rated health.
Descriptive statistics and comparisons and Cohen’s effect sizes between male and female twins on outcome variables
Note: Full sample: N = 1363; Males: n = 145, females: n =1218. Positive effect sizes mean that men scored higher than women.
Table 2 presents individual level correlations between outcomes. Notably, twins’ self-rated humor is moderately positively correlated (.48) with how their co-twin views their sense of humor.
Individual level correlations between outcomes
Note: Bolded values indicate significant difference from zero at p < .05. All correlations are Pearson correlations, CRT scores treated as continuous for descriptive purposes. Co-twin perception refers to how twin B perceives and rates twin A’s humor, which is independently measured through self-rated humor.
Twin Models
Within-trait twin intraclass correlations and results of univariate twin models are presented in Table 3. In general, MZ twin correlations were larger than DZ twin correlations, though the latent humor ability score is a clear exception. Univariate model (Table 3) results indicate that all outcomes are moderately heritable, except latent humor ability scores, for which the additive genetic effects were nonsignificant. See Table S1 on the Cambridge Core website for a full description of full and reduced model fit indices. Other than Latent Humor Ability, we did not observe significant shared environmental effects for any outcome. To further understand the covariation between latent humor ability and CRT scores observed at the individual level (see Table 1), we evaluated a bivariate twin model. The bivariate twin model indicated that the relationship between latent humor ability and CRT is via the unique environment, considering that latent humor ability is best represented by a CE model and CRT as AE.
Twin correlations and univariate model-fitting results for outcome variable phenotypes
Note: Bolded values indicate significant difference from zero at p < .05. All correlations are intraclass correlations with the exception of CRT, which is a polychoric correlation.
Table 4 presents the individual correlations between self-ratings of humor and ratings of the co-twin across all twins. Tables 5 and 6 display these correlations separately for MZ and DZ twins. Overall, the strongest correlation is between how twins rate themselves and how they rate their co-twin. In general, MZ correlations tend to be higher than DZ correlations.
Individual correlations of twins and co-twins humor ratings
Note: Bolded values indicate significant difference from zero at p < .05. All correlations are Pearson correlations.
MZ individual correlations of twins and co-twins humor ratings
Note: Bolded values indicate significant difference from zero at p < .05. All correlations are Pearson correlations.
DZ Individual correlations of twins and co-twins humor ratings
Note: Bolded values indicate significant difference from zero at p < .05. All correlations are Pearson correlations.
Discussion
Our study is the first to evaluate the relative contributions of genetic and environmental factors to humor ability. The results indicate that self-rated humor has a substantial heritable component, along with significant non-shared environmental influence. In contrast, the objective measure of HPA) showed, at best, a small heritability component, with most individual differences attributed to shared and nonshared environmental factors. This finding does not support our hypothesis that HPA has a considerable genetic basis and contradicts previous studies that found moderate to high additive genetic component for other cognitive abilities such as intelligence and creativity (Haier et al., Reference Haier, Colom and Hunt2023; Haworth et al., Reference Haworth, Wright, Luciano, Martin, de Geus, van Beijsterveldt, Bartels, Posthuma, Boomsma and Davis2010; Piffer & Hur, Reference Piffer and Hur2014; Polderman et al., Reference Polderman, Benyamin, De Leeuw, Sullivan, Van Bochoven, Visscher and Posthuma2015; Turkheimer et al., Reference Turkheimer, Haley, Waldron, d’Onofrio and Gottesman2003), as well as to other humor-related attributes (Baughman et al., Reference Baughman, Giammarco, Veselka, Schermer, Martin, Lynskey and Vernon2012; Vernon, Martin, Schermer, Cherkas, et al., Reference Vernon, Martin, Schermer, Cherkas and Spector2008; Vernon, Martin, Schermer, & Mackie, Reference Vernon, Martin, Schermer and Mackie2008).
We conducted post-hoc power analyses according to the process outlined by Verhulst (Reference Verhulst2017). Power to detect a significant variance component in a continuous twin model depends on the overall sample size, ratio of MZ to DZ twin pairs, and magnitude of both A and C effects. With our sample size and variance components, we had a power of 0.27 to detect a significant A component for HPA. We would require a sample size approximately 6 times larger than the present study to have power of 0.80 to detect a significant A component, given the same underlying variance decomposition and ratio of MZ to DZ twins. Future work in a larger sample size and more even MZ:DZ ratio is necessary to understand the variation underlying HPA.
What else might explain the lack of evidence for heritability in HPA, besides statistical power? One possibility relates to the task used to measure HPA. There is no consensus on the best method for assessing humor ability, nor is there a standardized test for evaluating it (Ruch & Heintz, Reference Ruch, Heintz, Luria, Baer and Kaufman2019; Silvia et al., Reference Silvia, Cotter and Christensen2024). Most researchers agree that the cartoon captioning task is a useful tool for distinguishing individuals with high humor ability from those with lower humor ability, and it is the most widely used measure of humor creativity (Greengross et al., Reference Greengross, Silvia and Nusbaum2020). However, the task may not accurately reflect an individual’s true humor prowess, as it is somewhat artificial. Most humor arises spontaneously in natural conversations, whereas the task requires individuals to be funny on demand. Additionally, the task is cognitively demanding — creating a funny caption on the spot requires concerted effort and can be stressful, especially for older individuals, such as those in our sample (median age: 66). Moreover, rating captions may introduce bias, as the judges were younger than the twin participants. Despite multiple ratings per caption, the overall assessments may not fully capture the twins’ true humor ability. Nonetheless, the task has been successfully used in many studies and is highly correlated with other measures of HPA (Christensen et al., Reference Christensen, Silvia, Nusbaum and Beaty2018; Greengross & Miller, Reference Greengross and Miller2011; Kozbelt & Nishioka, Reference Kozbelt and Nishioka2010; Nusbaum et al., Reference Nusbaum, Silvia and Beaty2017). Additionally, our measurement of HPA and CRT both are impacted by range restriction via ceiling or floor effects, resulting in low variability. This range restriction results in lower reliability and increased measurement error, which is captured by the E component in the twin model. A higher E proportion necessarily means a lower A and/or C proportion, which feeds back into the statistical power analyses for the univariate twin models.
Assuming the validity of the task, how can a possible lack of heritability in HPA be interpreted? This result contradicts studies showing a significant genetic basis for other humor attributes (Baughman et al., Reference Baughman, Giammarco, Veselka, Schermer, Martin, Lynskey and Vernon2012; Vernon, Martin, Schermer, Cherkas, et al., Reference Vernon, Martin, Schermer, Cherkas and Spector2008; Vernon, Martin, Schermer, & Mackie, Reference Vernon, Martin, Schermer and Mackie2008). However, producing high quality humor is more complex than other humor-related traits and requires creativity, intelligence, and playfulness. Other factors, such as personality, mood, and cognitive flexibility also play a role. HPA may be comparable to traits composed of multiple distinct capacities. For instance, romantic love styles and materialism arise from various personality traits and attitudes and, despite their intuitive everyday meaning, are difficult to define and measure. Similar to HPA, these traits exhibit zero to low heritability, with nearly all individual differences attributed to shared and nonshared environmental factors (Giddens et al., Reference Giddens, Schermer and Vernon2009; Waller & Shaver, Reference Waller and Shaver1994).
More importantly, a lack of heritability in HPA challenges the widely accepted evolutionary basis of humor (Greengross & Kozbelt, Reference Greengross, Kozbelt, Chłopicki, Ford and Kuipers2024). A great sense of humor can help ease tension in dangerous situations, foster cooperation, break down interpersonal barriers, and attract mates — all of which enhance survival and reproduction. It is possible that our humor ability task, while valid, does not capture the type of humor used in conversation and social interactions, which may be the primary driver of evolutionary change. Another possibility is HPA is not under selection in the current environment — though this seems unlikely given the strong evidence supporting the importance of humor in various social settings, particularly for choosing a mate (Greengross & Kozbelt, Reference Greengross, Kozbelt, Chłopicki, Ford and Kuipers2024).
Assortative mating, in which individuals select mates with similar traits, might also contribute to the observed results. Nonrandom mating could obscure genetic effects (Gonggrijp et al., Reference Gonggrijp, Silventoinen, Dolan, Boomsma, Kaprio and Willemsen2023). However, this explanation is unlikely because men and women value humor differently in mate selection. Research shows that women prioritize high HPA in a partner more than do men, whereas men value a women’s ability to appreciate their humor (Bressler & Balshine, Reference Bressler and Balshine2006; Bressler et al., Reference Bressler, Martin and Balshine2006; Greengross & Kozbelt, Reference Greengross, Kozbelt, Chłopicki, Ford and Kuipers2024). Women’s high HPA does not necessarily make them more attractive to men. This suggests that men experience stronger selection pressure to be funny to impress women, leading to men having slightly higher humor ability, on average — a finding supported by our study (see Greengross et al., Reference Greengross, Silvia and Nusbaum2020, for a review). Furthermore, men rated themselves as funnier than women, likely reflecting an awareness of humor’s importance in female mate choice. Indeed, studies show that men emphasize their humor ability on dating sites, whereas women’s humor ability does not enhance their attractiveness (Wilbur & Campbell, Reference Wilbur and Campbell2011). Our sample was highly imbalanced, with only about 10% men. A more balanced sex ratio might have yielded different results with greater variability in HPA.
Age may also play a role. Our sample consisted of relatively older individuals, limiting our ability to examine age effects on heritability. Studies on cognitive abilities such as IQ typically include younger participants and find that additive genetic factors increase with age, peaking in adulthood (Bergen et al., Reference Bergen, Gardner and Kendler2007; Bouchard Jr, Reference Bouchard2013; Haworth et al., Reference Haworth, Wright, Luciano, Martin, de Geus, van Beijsterveldt, Bartels, Posthuma, Boomsma and Davis2010; Plomin et al., Reference Plomin, DeFries, Knopik and Neiderhiser2016; Tucker-Drob et al., Reference Tucker-Drob, Briley and Harden2013). However, research on elderly populations suggests that the heritability of cognitive abilities decreases after age 65 (Lee et al., Reference Lee, Henry, Trollor and Sachdev2010). A review of 27 studies on the heritability of verbal ability — a trait strongly correlated with HPA (see: Greengross & Miller, Reference Greengross and Miller2011) — found that verbal ability declines after age 60, while nonadditive genetic factors increase (Reynolds & Finkel, Reference Reynolds and Finkel2015). Our study design did not allow us to capture nonadditive genetic influences, but future studies with extended family designs might detect such effects. Additionally, a younger twin cohort might reveal a different genetic architecture, which includes additive genetic influences.
One might wonder whether the talents of Jerry Seinfeld, George Carlin, Joan Rivers, and other great comedians, can be attributed solely to environmental factors. Do comedians known for their improvisational skills and spontaneous witty repartee — such as Robin Williams, Jonathan Winters, and Billy Connolly — possess no genetic basis for their talent? While this may seem intuitively unlikely, our study could not identify such a basis.
It is also possible that the ability to make others laugh in everyday social situations differs from the ability to appeal to mass audiences. Indeed, comedians who took the cartoon caption task scored several orders of magnitude higher than the general population (Greengross et al., Reference Greengross, Martin and Miller2012). Thus, the characteristics of top comedians do not necessarily reflect the variation in the humor ability among the general population. Comedians represent the extreme end of humor spectrum — one that perhaps needs to be addressed separately.
In contrast with HPA, self-rated humor exhibited a strong additive genetic component. However, self-rated humor and HPA were only weakly correlated. Self-assessment of humor is considered a poor indicator of actual humor ability and is unreliable for evaluating HPA (Feingold & Mazzella, Reference Feingold and Mazzella1993; Silvia, Greengross et al., Reference Silvia, Rodriguez and Karwowski2021). Most people believe they have above average HPA, a statistical impossibility given a normal distribution, with men in particular overestimating their humor ability (Crawford & Gressley, Reference Crawford and Gressley1991; Mickes et al., Reference Mickes, Walker, Parris, Mankoff and Christenfeld2012; Myers et al., Reference Myers, Ropog and Rodgers1997; Silvia, Greengross, et al., Reference Silvia, Rodriguez and Karwowski2021). For example, in one study, 93% of men and 87% of women believed they had either an average or above average sense of humor (Chapman & Gadfield, Reference Chapman and Gadfield1976). The high heritability of self-assessed humor may indicate a strong genetic component for psychological traits unrelated to HPA. In particular, extroverted people have higher opinion of their humor ability, while people high on conscientiousness tend to have lower confidence in their ability to make others laugh (Beins & O’Toole, Reference Beins and O’Toole2010; Silvia, Greengross et al., Reference Silvia, Rodriguez and Karwowski2021; Silvia, Rodriguez et al., Reference Silvia, Rodriguez and Karwowski2021). Some studies find a weak correlation between openness to new experiences and self-assess humor, whereas openness is much more strongly positively correlated with HPA (Greengross & Miller, Reference Greengross and Miller2011; Nusbaum, Reference Nusbaum2015; Silvia, Greengross et al., Reference Silvia, Greengross, Cotter, Christensen and Gredlein2021).
CRT also exhibited high heritability. This approximate but widely used measure of cognitive ability is usually correlated with other cognitive attributes, such as intelligence (Otero et al., Reference Otero, Salgado and Moscoso2022). However, our study revealed practically zero correlation between CRT and both self-rated humor and HPA, contrary to previous studies that did find a positive correlation between HPA and cognitive ability. HPA is strongly correlated with intelligence, and, in particular, verbal ability (Greengross & Miller, Reference Greengross and Miller2011). Unlike the weak correlation between HPA and self-assessed humor, which is predictable and highlights the difference between true humor ability and the perception of that ability, the weak relationship between HPA and CRT — two domains of cognitive ability — is surprising and may indicate inaccurate estimates of HPA, CRT or both in our study. The very small correlation between HPA and CRT did not allow us to estimate the genetic covariance of both measures.
Self-assessed health showed strong additive genetic and nonshared environmental components. Many common diseases are partially heritable, but the underlying genetic relationship between humor ability and health has not been studied. Such a link could be adaptive, demonstrating the role plays in enhancing survival. However, our study found no substantial correlations between self-assessed health and all humor measures. This lack of association should be interpreted with caution, as no objective health measures were taken.
Conclusion
Our study’s finding that HPA lacks heritability is surprising, as it contradicts most research on the heritability of cognitive abilities and is therefore interesting (Davis, Reference Davis1971). Since this is the first study to examine the heritability of HPA, these results should be interpreted with caution. Small genetic effects cannot be ruled out, and some studies on traits with significant heritability also fail to identify an additive genetic component. Even research on other humor-related attributes, such as humor styles and humor appreciation, which generally find genetic effects, occasionally fail to identify an additive genetic component for humor appreciation (Cherkas et al., Reference Cherkas, Hockberg, MacGregor, Snieder and Spector2000). Humor ability is a multifaceted phenomenon involving various cognitive processes that are difficult to assess. It is a complex and unique trait influenced by numerous psychological attributes and personality characteristics and varies across different social contexts. These factors, along with the unique characteristics of our sample, may explain the lack of heritability in HPA. On the other hand, the current results may indicate that we are missing something fundamental in our understanding of humor ability. If these results hold in further studies, we may need to rethink the role humor play in various theoretical frameworks. Future studies with different twin demographics and perhaps alternative methods to assess HPA will provide a clearer picture of how heritable HPA is, if at all.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/thg.2025.10010.
Acknowledgements
TwinsUK is funded by the Wellcome Trust, Medical Research Council, Versus Arthritis, European Union Horizon 2020, Chronic Disease Research Foundation (CDRF), Wellcome Leap Dynamic Resilience Programme (co-funded by Temasek Trust), Zoe Ltd, the National Institute for Health and Care Research (NIHR) Clinical Research Network (CRN) and Biomedical Research Centre based at Guy’s and St Thomas’ NHS Foundation Trust in partnership with King’s College London.
Financial support
The work was supported by funding from Julie Aitken Schermer, a jointly appointed Professor in the departments of Psychology and Management and Organizational Studies at The University of Western Ontario.
Competing interests
None
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. The study was carried out in TwinsUK under the TwinsUK BioBank (REC reference 19/NW/0187, IRAS ID 258513). Ethical approval was also obtained from the Department of Psychology at Aberystwyth University.
Open access
