Sex differences in humor production ability: A meta-analysis

Abstract

We offer the first systematic quantitative meta-analysis on sex differences in humor production ability. We included studies where participants created humor output that was assessed for funniness by independent raters. Our meta-analysis includes 36 effect sizes from 28 studies published between 1976 and 2018 (N = 5057, 67% women). Twenty of the 36 effect sizes, accounting for 61% of the participants, were not previously published. Results based on random-effects model revealed that men’s humor output was rated as funnier than women’s, with a combined effect size d = 0.321. Results were robust across various moderators and study characteristics, and multiple tests indicated that publication bias is unlikely. Both evolutionary and cultural explanations were considered and discussed.

Full text

1
****** Pre-proof final accepted version ******
Title: Sex differences in humor production ability: A meta-analysis
Gil Greengross (corresponding author), Paul J. Silvia , and Emily C. Nusbaum
3.
1. Department of Psychology, Penbryn 5, Penglais Campus, Aberystwyth University,
Aberystwyth, UK, SY23 3UX. Tel: 01970 622120. Email: humorology@gmail.com
2. Department of Psychology, University of North Carolina at Greensboro, 296 Eberhart
Bldg, PO Box 26170, Greensboro, NC 27402-6170.
3. Department of Psychology, University of North Carolina at Greensboro, PO Box
26170, Greensboro, NC 27402-6170.
Acknowledgments
We thank the following researchers for providing unpublished data and other information
included in our meta-analysis: Ori Amir, Mathias Benedek, Judit Boda-Ujlaky, Dov
Cohen, Glenn Geher, Varda Inglis, Olivia Jewell, Scott Barry Kaufman, Barry
Kudrowitz, Karl-Heinz Renner, Vassilis Saroglou, and Shoshana Shiloh.
Author contribution: Conception of the study – GG, PS, EN. Searching for studies, data
collection, coding: GG, PS. Data analysis and interpretation: GG, PS, EN. Drafting the
article: GG. Critical revisions and final approval of the article: GG, PS, EN.
1 2

2
Abstract
We offer the first systematic quantitative meta-analysis on sex differences in humor
production ability. We included studies where participants created humor output that was
assessed for funniness by independent raters. Our meta-analysis includes 36 effect sizes
from 28 studies published between 1976 and 2018 (N = 5057, 67% women). Twenty of
the 36 effect sizes, accounting for 61% of the participants, were not previously published.
Results based on random-effects model revealed that men’s humor output was rated as
funnier than women’s, with a combined effect size d = 0.321. Results were robust across
various moderators and study characteristics, and multiple tests indicated that publication
bias is unlikely. Both evolutionary and cultural explanations were considered and
discussed.
Keywords: Humor; humor production ability; sex differences; evolutionary psychology;
meta-analysis
Public Significance Statements
This meta-analysis suggests that, on average, men have higher humor production ability
than women, as judged by independent raters assessing the humor produced by both
sexes. This difference is small-to-moderate and may reflect divergent preferences and
experiences of humor for men and women, stemming from evolutionary and
environmental influences.

3
1. Introduction
There is an ongoing debate about whether men and women differ in their cognitive
abilities, how big the differences are, and how to explain them, if they exist (Halpern,
2011; Halpern et al., 2007; Hyde, 2005, 2014; Lindberg, Hyde, Petersen, & Linn, 2010;
Lippa, Collaer, & Peters, 2010; Spelke, 2005; Voyer, Postma, Brake, & Imperato-
McGinley, 2007; Voyer, Voyer, & Bryden, 1995; Voyer, Voyer, & Saint-Aubin, 2017;
Zell, Krizan, & Teeter, 2015). Humor production ability (HPA)—a cognitive trait defined
as the ability to produce funny remarks, create funny ideas, and make others laugh—is
one such domain (Greengross, 2014; Greengross & Miller, 2011; Hooper, Sharpe, &
Roberts, 2016; Lampert & Ervin-Tripp, 1998; Martin, 2014). Social stereotypes about sex
differences in humor—particularly the stereotype that “women are not funny”—are
culturally pervasive (Hitchens, 2007; Shlesinger, 2017). To date, no systematic review
has evaluated the evidence for whether men and women differ in their humor production
ability. After reviewing different schools of thought that seek to explain the role of
biological sex in humor production, we meta-analytically synthesize the literature that has
accumulated that can inform this question.
1.1. Considering Sex Differences in Humor Production Ability
1.1.1. Evolutionary Explanations
The universality of humor, its early developmental onset, and the fact that humans are not
the only species that smiles and laughs have led researchers to suggest that humor has an
evolutionary basis (Alexander, 1986; Chafe, 1987; Davila-Ross, Owren, & Zimmermann,
2009; Gamble, 2001; Hurley, Dennett, & Adams Jr., 2011; Miller, 2000a; Preuschoft &

4
Van-Hooff, 1997; Ramachandran, 1998; Viana, 2017; Weisfeld, 1993). The most relevant
evolutionary theory pertaining to the possibility of sex differences in HPA is the mental
fitness indicator theory, an extension of sexual selection theory (Darwin, 1871;
Greengross & Miller, 2011; Howrigan & MacDonald, 2008; Miller, 2000a, 2000b).
According to sexual selection theory, males’ and females’ distinct behaviors and
preferences are shaped due to asymmetrical parental investment (Trivers, 1972). In
sexually reproducing species, the sex that bears the higher costs of reproduction is the
choosier one, in most cases the females. Miller (2000a, 2000b) proposed that various
cognitive capacities, such as language, arts, sports, and humor, evolved through mutual
mate choice to advertise mate quality. These hard-to-fake mental fitness indicators serve
to promulgate one’s cognitive prowess, and are honest signals of intelligence that
underline an individual’s genetic quality. Humor is hypothesized to be one such fitness
indicator, and HPA is positively correlated with various intelligence measures, most
strongly with verbal aptitude (Christensen, Silvia, Nusbaum, & Beaty, 2018; Greengross
& Miller, 2011; Howrigan & MacDonald, 2008; Kellner & Benedek, 2017). The mental
fitness indicator theory proposed by Miller (2000a, 2000b) predicts that since women are
choosier than men, men will be more motivated to advertise HPA in an effort to attract
women, while women will put more effort in selecting mates based on men’s ability to
produce and showcase high levels of HPA. Why should this lead to higher levels of HPA
among men? The answer lies in selection pressures involving a stronger male-male
competition (intra-sexual selection), which drives men to ever improve their HPA in an
effort to be funnier than their rivals in an attempt to attract women. Thus, we expect men

5
to have a better HPA than women, as there is weaker evolutionary pressure for women to
use humor to attract mates (Miller, 2000b).
Several studies support the various predictions stemming from the evolutionary
explanation for viewing HPA as a sexually selected trait. Across cultures, sense of humor
is found to be a more desirable trait in a mate for women choosing a mate, than for men
(Buss, 1988; Feingold, 1992; Goodwin, 1990; Lippa, 2007; Sprecher & Regan, 2002;
Todosijević, Ljubinković, & Arančić, 2003; Toro-Morn & Sprecher, 2003). However,
one study found no difference (McGee & Shevlin, 2009), and another found the opposite
trend (i.e., that men view women as more attractive and more suitable as mates, when the
women portrayed a great sense of humor, but not the opposite) (Antonovici & Turliuc,
2017). The apparent contradictory results could be explained by the lack of clarity in the
meaning of the term ‘sense of humor’. Saying that someone has a great sense of humor
could mean that the person exhibits a high level of HPA, or that he or she enjoys humor
or laughs often and easily. The lack of distinction leaves participants to have their own
interpretation of the term and to mixed results. To resolve the issue, Bressler et al. (2006)
specifically tested whether women prefer men who display high HPA, and men prefer
women who appreciate their humor. The study found that although both men and women
valued a good sense of humor in their respective partners, women showed a preference
for a man with great HPA over a man that appreciated their humor production, while men
preferred a woman that would appreciate their humor over a woman that would make
them laugh. These results support the notion that when men and women talk about

6
wanting a partner with a great sense of humor, they mean vastly different things. Men
want a humor appreciator, while women want a humor producer.
Other studies from more ecologically valid situations, such as personal ads in newspapers
and online dating sites, where people have low incentive to lie about their true
preferences, show that women seek a mate who portrays humor ability twice as much as
men do, and that men are more likely to declare how funny they are, or attempt humor,
compared to women (Smith, Waldorf, & Trembath, 1990; Wilbur & Campbell, 2011).
Women, on the other hand, express a desire for mates that will make them laugh, much
more than men do, corroborating the prediction that men try to advertise their humor
ability, and women are the appreciators of humor. Lastly, research shows a direct link
between HPA and mating success. Adding humor to personal ads made men more
attractive to women, but had little effect on women’s attractiveness (Wilbur & Campbell,
2011). In another study, individuals who were rated high for HPA reported higher mating
success as measured by number of sexual partners, age of first intercourse, and more
sexual encounters, compared to individuals with low HPA (results were true for both
sexes) (Greengross & Miller, 2011). However, women who have humorous partners did
report having more and stronger vaginal orgasms, compared to women who have less
funny partners, while men’s sexual satisfaction was not related to women’s HPA (Gallup,
Ampel, Wedberg, & Pogosjan, 2014). These results highlight the significance of partner’s
high HPA on women, something that may contribute to higher reproductive success.
1.1.2. Social Factors and Humor Production Ability

7
Many social and cultural factors influence the way men and women create humor and
how it is perceived. The notion that men are funnier than women is a widely-held
stereotype and a cultural trope (e.g., Hitchens, 2007; Shlesinger, 2017). For example,
when asked to describe an individual with a great sense of humor, or to name which sex
is funnier, both men and women are much more likely to describe or choose a man
(Crawford & Gressley, 1991; Nevo, Nevo, & Yin, 2001). In one study, 94% of the men
and 89% of the women agreed to the stereotype that men are funnier than women
(Mickes, Walker, Parris, Mankoff, & Christenfeld, 2012). In another study with both
sexes surveyed, 62% of the participants believed that men have a greater HPA than
women, 34% thought that men and women are equally funny, and only 4% viewed
women as the funnier sex (Hooper et al., 2016). In addition, both men and women are
more likely to attribute funny captions to male writers, and non-funny cartoons to
women, even where the identity of the humor producer is concealed (Mickes et al., 2012).
Such a stereotype may suppress women’s willingness, and hinder their ability to create
humor, ultimately putting them at a disadvantage compared to men.
Social role theory may be the best cultural framework in which observed sex differences
are understood, and offers an alternative to the evolutionary explanation (Archer, 1996).
According to the theory, sex differences emerge as part of the historical demarcation of
men’s and women’s roles within a society, which place them in unequal positions. In
many societies, men have higher status and hold more power than women, while
controlling the majority of the resources (Eagly & Wood, 1999). This power asymmetry
leads the more powerful men to more masculine and dominant behaviors, while less

8
powerful women exhibit behaviors that are subordinate and passive. Societal expectations
to fit into sex-specific roles put pressure on both men and women to acquire the skills and
adopt behaviors that will conform to their role requirements. If the stereotype that men
have higher HPA than women is pervasive, cultural practices will work to sustain such a
notion, and both men and women might try to fit into that expectation. Such practices
were common throughout history and may still persist today. For example, for many
years women tended to be the objects of jokes, often disparaging and sexist in nature, but
rarely the subject producing the humor (Kotthoff, 2006). Specifically, women were
prevented from using humor in the public sphere, not allowed to tell jokes and perform
comedy routines, and confined to tell jokes only in private, while men were free to
exhibit their humor in any form and platform they wished. These expectations, especially
if indoctrinated from early ages, may contribute to observed sex differences in HPA.
The mechanisms in which such expectations transfer into behavior are often referred to as
cultural scripts (Goddard & Wierzbicka, 2004). Cultural scripts include norms, values
and practices that serve as collective guides to people on how to behave. Cultural scripts
are highly influenced by gender stereotypes and could apply to various uses of humor. A
specific script, the traditional courtship script, may come into play when using humor to
attract mates. In the traditional courtship script, men are expected to be more active and
take the initiative, while women are assumed to have a more passive role, being the
recipients of men’s romantic invitations (Eaton & Rose, 2011). In relation to humor,
research suggests that men use humor to attract women, while women serve as
appreciators by evaluating men’s humor, thus fulfilling traditional gender role

9
expectations (Wilbur & Campbell, 2011). How men and women view HPA in the
traditional courtship script may be connected to traditional views of masculinity and
femininity, with HPA being a masculine trait and humor appreciation viewed as
feminine. Ross and Hall (in press) found that for both sexes, producing high quality
humor in a courtship setting was associated with trait masculinity, with high masculine
participants reporting using humor more to attract mates, compared to less masculine
people of the same sex. In addition, these men also believed their HPA was better than
same sex peers. Moreover, women adhering to traditional courtship behaviors, such as
not making the first move when initiating a relationship, were less likely to use humor to
attract men. Thus, the use of humor in courtship conforms to traditional gender roles of
men as the pursuer and women the appreciator, with HPA to attract mates perceived as a
masculine trait. The study illustrates the possible influences of sexual courtship scripts on
how HPA is perceived, regardless of its real quality, and more generally, how cultural
norms may affect how both men and women use humor.
1.2. Assessing Humor Production Ability
Understanding the empirical literature on sex or gender, and humor production requires
understanding how HPA is commonly assessed. As illustrated in Figure 1, the
measurement of humor creation ability typically consists of four steps, two for the
creation of the humor, and two for the evaluation of the humor. First, a participant
attempts to produce a humorous response to a non-funny stimulus or prompt provided by
the researchers. Second, judges rate the participants’ responses, and their ratings for each
participant are summarized to create the individual HPA score.

10
Humor Creation Humor Evaluation
Figure 1. Measuring humor production ability. Participants are asked to generate humor
in response to a non-funny stimulus. The responses are later evaluated by independent
judges, and an overall score of humor production ability is calculated.
Creating humor which will be later judged for funniness involves two steps: first,
introducing a stimuli, and second, creating a funny response relevant to the stimuli
presented. In the most common variation, researchers present participants with a picture
or a cartoon with no caption and ask them to write a funny caption (Babad, 1974;
Brodzinsky & Rubien, 1976; Eysenck, 1943; Feingold & Mazzella, 1991, 1993; Kohler
& Ruch, 1996; Koppel & Sechrest, 1970; Treadwell, 1970; Turner, 1980; Ziller,
Behringer, & Goodchilds, 1962). Other tasks seek a verbal response from a verbal
stimulus. For example, McGhee (1974) presented children with absurd riddles such as
“Why did the old man and his wife drive to the North Pole?” and asked them to provide
funny answers that were later judged for funniness. Similarly, Shiloh (1982) asked
participants to answer unusual questions in a funny way, such as “What would have
happened if the oceans were full of orange juice?”. Another example is to ask participants
to provide a funny definition to a nonexistent term, such as Yoga Bank and Fruit Jar, or
to write a funny ending to a social scenario that sets up funny responses (Christensen et
al., 2018; Nusbaum, Silvia, & Beaty, 2017). Only a few studies have sought visual
output. For example, Howrigan and MacDonald (2008) asked participants to draw a
Non-funny stimuli Humor response Ratings of stimuli Humor score

11
picture of four animals (e.g., a giraffe), and four people with a specific profession (e.g., a
professor).
Regardless of the type of stimuli and response employed, task administration methods are
wide ranging. First, there is no standard time frame for how long each task requires to
complete, ranging from as little as 30 seconds to write a funny caption for a captionless
cartoon (e.g., Moran, Rain, Page-Gould, & Mar, 2014) to unlimited time (e.g., Feingold
& Mazzella, 1993; Ziv, 1981b, 1983), and anything in between. Second, the number of
stimuli introduced to participants, and the number of responses allowed for each stimulus,
vary substantially. For example, some researchers may ask for only one response per
stimuli (e.g., Babad, 1974; Saroglou, 2002; Saroglou & Jaspard, 2001) while others will
not restrict the number of captions the participant can produce (e.g., Greengross & Miller,
2011; Kohler & Ruch, 1996).
Evaluating individual differences in funniness also requires two steps. The first step is to
ask judges to assess the level of funniness for each of the responses created by the
participant. There is no consensus on how to evaluate and score individual humor
production. With a few exceptions (e.g., Freiheit, Overholser, & Lehnert, 1998; Kozbelt
& Nishioka, 2010), most judges of the humor stimuli are college students (both
undergraduate and graduate) and professors (e.g., Nusbaum et al., 2017). The number of
judges also differ, from 2 to as many as 81 (Amir & Biederman, 2016; Mickes et al.,
2012). Most often judges are asked to rate the stimuli for “funniness”, but in some case
judges are tasked with rating the “humorousness” (e.g., Saroglou, 2002; Saroglou &

12
Jaspard, 2001,) or “wittiness” (e.g., Kohler & Ruch, 1996) of the caption or joke. How
much each judge rates varies as well. In some cases, every judge evaluates the humor for
all humor outputs, but sometimes, judges assess only a portion of the overall humor
produced, mostly because there are too many stimuli for one person to reasonably handle.
The rating scales themselves also vary, from merely a dichotomous distinction of whether
the stimulus was funny or not (Saroglou, 2002; Saroglou & Jaspard, 2001; Ziv, 1983), to
ratings ranging from 1-10 (Ruch, Beermann, & Proyer, 2009), and anything in between.
In sum, there is much variability in the identity of judges, how many stimuli they rate
from the overall sample, and what they are asked to do, when evaluating humor
creativity.
The second part in evaluating humor is to create an overall score for HPA. As with other
aspects of measuring humor ability, researchers use many methods to get an overall
rating of funniness. For starters, some researchers use the raw scores of judges’ ratings as
the basis for their analyses (e.g., Moran et al., 2014), and others standardize the scores
(e.g., Greengross & Miller, 2011). Furthermore, when more than one humor output is
measured, and there are multiple judges rating the humor, there is a need to summarize all
the scores to one statistic. Some researchers average all the responses across raters (e.g.,
Brodzinsky & Rubien, 1976), others take the highest score from each judge and average
them (e.g., Greengross & Miller, 2011), and still others use many-facet Rasch models to
distill the faceted design to a single score (e.g., Nusbaum et al., 2017).

13
In summary, there is no standardized procedure for assessing HPA. Nonetheless, while
there is much variation in the tasks, the premises and procedures tend to follow the same
mechanism. Overall, different measures of humor following the four-step process
described in Figure 1 are strongly correlated with each other (Christensen et al., 2018;
Howrigan & MacDonald, 2008; Nusbaum et al., 2017).
1.3. The Current Research
It is clear that there is a pervasive stereotype that men have higher humor production
abilities than women. Regardless of the reasons for such a belief, or why men and women
might differ in their HPA, the veracity of sex differences in humor ability has not been
systematically evaluated to date. Our goal is to create the first meta-analytic review on
this topic. While a few qualitative reviews on sex differences in HPA exist, their
conclusions were inconsistent with each other, and the scope of the reviews is limited
(Greengross, 2014; Kotthoff, 2006; Martin, 2014). Thus, we had three main goals for our
meta-analysis. First, we aimed to gather all available data on sex differences in HPA, and
create the largest database on the topic to date. Second, we planned to estimate
quantitatively the magnitudes of sex differences in HPA based on weighed effect sizes,
from all available data. Third, we tested the possible influence of various moderators on
sex differences in HPA (see below).
2. Method
2.1. Literature Search

14
A key feature of the literature on humor ability is that relatively few studies have looked
at and reported information about sex differences in HPA. Many publications that
evaluated humor ability directly do not consider sex as an important factor of interest, and
often do not report sex-specific data on HPA beyond the number of men and women
participants. Other times, analyses of sex or gender differences are included, but the
information is too limited to extract an effect size (e.g., Martin & Lefcourt, 1983). In
either case, data on men’s and women’s humor ability might still exist and could be
recovered, as sex is a common variable recorded in most studies. Thus, to avoid missing
potentially relevant data, we searched for any study that had a measure of HPA,
regardless of whether the information on sex differences was provided in the publication
or not. If the researchers did not include data on men’s and women’s humor ability, or
there was no information or analyses that could be reverse engineered to extract an effect
size, we contacted the authors to attempt to retrieve the relevant data.
We employed multiple strategies to locate all relevant studies and data. First, we used a
backward search (Card, 2015, p. 49-50) and examined the reference lists of key review
publications that covered either sex or gender differences in humor or humor production
in general. These reviews included Greengross (2014); Kaufman et al. (2008); Lampert
and Ervin-Tripp (1998); Martin (2014); Nusbaum (2015); O'Quin and Derks (1997); and
Ruch (2008). Second, we used forward search (Card, 2015, p. 50-51) to look for studies
that cited key articles on humor production, or ones that introduced new measures of
HPA. These articles are: Babad (1974); Brodzinsky and Rubien (1976); Feingold and
Mazzella (1991); Kohler and Ruch (1996); Koppel and Sechrest (1970); Masten (1986);

15
Treadwell (1970); and Turner (1980). Third, we emailed the listserv of the International
Society for Humor Research (ISHS) asking for any unpublished data. Fourth, we
searched the conference proceeding of the American Psychological Association,
Association for Psychological Science, and ISHS, for any humor paper or poster
presented since 2000, where data were available. Fifth, we searched the following
databases: PsycINFO, PsycARTICLES, Web of Science, PubMed, ProQuest Dissertation
& Theses Global, National Library of Israel, and Google Scholar. For searches in the
databases we included the following phrases and their combinations: humor ability,
humor creat*, humor product*, and cartoon caption*. All of the above searches were
also performed using the non-American spelling, humour. Sixth, we contacted prominent
researchers in the field of humor studies that come from non-English speaking countries,
and asked if they knew about any publications in their field. In total, our literature
searches included the following languages: Chinese, English, French, German, Hebrew,
Hungarian, and Japanese. Publications in Hebrew were translated by the first author, and
an author of a Hungarian paper translated all the relevant data we requested. Seventh, all
authors of the current paper had unpublished data that was relevant to the meta-analysis,
and knew about other researchers with additional unpublished data that we were able to
obtain. We read the abstract, the full text, or both to determine if they contained relevant
data for our meta-analysis. We concluded our search in August 2018.
2.2. Inclusion and Exclusion Criteria
Our approach was to include studies of humor production tasks that comprise of the two
elements that we believe are the most essential for assessing true individual differences in

16
humor ability with minimum biases: one, the creation of a new humor output (i.e., not a
completion of a known joke or a joke recall), and two, the evaluation of the humor
produced by independent judges who do not know the identity or any characteristics of
the humor producer. Thus, to be included in the meta-analysis, studies had to include the
following criteria:
1. The study must have included a sample of men and women.
2. The humor production task must have included creation of verbal humor.
3. Participants must have generated spontaneous new or innovative humor as part of
the humor production task. Studies that were based on self-reports, or scales such
as the Multidimensional Sense of Humor Scale (MSHS) (Thorson & Powell,
1993), or studies where participants had to complete a joke from multiple possible
responses, given the setup, were excluded.
4. Judges had to be blind to any characteristic of the humor producers. This excludes
studies in which judges observed participants in the lab or on video (e.g., Inglis,
Zach, & Kaniel, 2014).
5. Judges rated the humor for funniness. Studies in which there were no judges, and
the ratings of HPA were solely based on counting the number of responses or
humor attempts, rather than actual ratings of the humor produced, were excluded
(e.g., Hall, 2015). One study where judges were instructed to rate responses based
on consensual sense of humor, i.e., if the answers could be seen as funny in
principle by most people, regardless of the judges’ own appraisal (essentially

17
evaluating humor attempts rather than actual humor), was excluded (Hull, Tosun,
& Vaid, 2017).
A couple of exclusion criteria were also applied. Studies on children (preadolescence)
were excluded (e.g., Masten, 1986). In addition, studies that were conducted on
participants with brain damage were not considered, though no such study met all the
other inclusion criteria.
In cases where a study met all the inclusion criteria, but lacked sufficient information for
a calculation of an effect size, we contacted the authors to obtain the information. A total
of 22 authors were contacted, and eight of them provided data that could be included in
our meta-analysis. In a few cases, the paper was too old and the data were lost, and one
author refused to share the data. Additionally, the authors of the present paper had three
unpublished datasets and knew of one more unpublished dataset that met our criteria. All
these data were procured and included in the meta-analysis (see Figure 2 for the PRISMA
schematic screening of the studies) (Moher, Liberati, Tetzlaff, & Altman, 2009).
Any disagreements about what studies should be included were discussed and resolved
among the authors. Altogether, 28 studies met our inclusion criteria, with 36 independent
samples, and a total sample of 5057 participants (1677 men, 3380 women, 67% women).

18
Figure 2. PRISMA flow chart of screening process of studies selected for the meta-
analysis.
Records identified through
database searching
(n =
6
,
774
)
Screening
Included
Eligibility
Identification
Additional records identified
through other sources
(n =
27
)
Records after duplicates removed
(n = 4,237)
Records screened
(n = 4,237)
Records excluded
(n = 4,188)
Full-text articles assessed
for eligibility
(n = 49)
Full text articles (n =
21)
excluded due
to:
 Inadequate humor creation
measure (n = 8)
 Data insufficient or unavailable
even after contacting authors (n
= 11)
 Data on children (n = 2)
Studies included
in meta
-
analysis
(n = 28)

19
2.3. Coding Procedure
All studies were coded by two independent coders, one of whom was the first author. All
disagreements were discussed and resolved between the coders. For each of our samples,
we estimated the effect sizes based on standardized mean differences (Cohen’s d)
(Cohen, 1988). Most calculations were based on means, SD’s and samples sizes, but
when they were not available, we employed inferential statistics to calculate the effect
size (see Card, 2015). These procedures included the use of F-statistics, t-statistics, and p-
values. We elected to use Cohen’s d as our estimate for the standardized mean difference
over Hedges' g as most samples were at least modest in size, and results based on Hedges'
g were nearly identical to those based on Cohen’s d (Card, 2015, p. 90-91). A positive
sign of Cohen’s d denotes a higher humor ability for men.
Several studies included more than one measure of HPA per sample. In such cases, we
averaged the effect sizes to produce one d effect size for each sample. In total, we
calculated average effect size for 12 out of the 36 samples with multiple outcomes. One
exception was Mollica (1983), where we chose one effect size from the two reported. We
believe this is justifiable as the effect size was based on the averages across all captions
produced, and the one excluded was computed based on the average of the first caption
only. Both measures are highly correlated with each other, and the overall average score
embodies within it the average for the first caption.
Note that the procedures used to produce the overall HPA scores for each study vary, as
there is no consensus on the best way to calculate a humor score. As discussed earlier,

20
researchers employ various protocols to calculate a humor score, which are partially
dependent on the tasks employed, and whether one or more stimuli are used to generate
humorous responses. For example, when multiple stimuli are used, researchers may
average all responses across raters (Brodzinsky & Rubien, 1976), or only average the best
outputs across raters (Greengross & Miller, 2011). For single tasks, there is no need to
average responses, but they might be combined with other humor production tasks to
create one overall score (e.g., Nusbaum et al., 2017). Still, despite the variability in the
scoring of HPA, evidence suggests that different methods and scoring procedures yield
similar results (Christensen et al., 2018; Mollica, 1983; Nusbaum et al., 2017).
2.4. Coding of Potential Moderators
We identified several variables that could potentially moderate sex differences in humor
ability. The aim of the moderators was twofold. First, we wanted to test possible biases in
the studies or the data collected. Some of these potential biases are associated with
research on sex differences, such as the sex of the researcher, while other possible biases
may relate to study characteristics, such as the type and nature of the samples or the
country in which data were collected. Publication bias was also a concern, and we
included several variables that specifically address this possibility (see below). Second,
we included moderators that could illuminate or provide new insights into the source of
sex differences in HPA, if it exists. These moderators pertain mostly to the measures,
procedures, and evaluations associated with the humor ability task. It is possible that
some moderators may serve both functions and both help detect biases and also elucidate
the nature of sex differences in HPA. For example, publication year can both illustrate

21
how sex differences in HPA has varied over the years, but also indicate changes in
measurements of HPA across time.
We coded for 26 variables that were intended to be used as moderators, but not all were
included in the moderator analyses, either due to little variability in the outcome, or
because very few studies reported the relevant information. The list of the six excluded
moderators and the justification for their exclusion appear in Table 1. In total, 20
variables served as moderators.
Publication year.
Publication year could help identify possible changes and trends of
sex differences in HPA over time.
Affiliation of first author.
The country in which the first author’s affiliation appears
on the manuscript, or in the case of unpublished data, where the researcher resides. This
categorical variable includes the following countries/regions: North America (US &
Canada), Europe (Austria, Belgium, Germany, Hungary, United Kingdom), and Israel.
We grouped together the European countries as there were too few studies for each
country to be included as a separate category. This variable may roughly represent the
country in which the data were collected (i.e., where participants came from), but not in
all cases (see the sample country variable). This moderator allows us to test whether the
results are consistent cross-culturally.
Sex of first author.
We dichotomously coded whether the first author was a man or a
woman. This moderator aims to test for possible sex bias by the researcher conducting the
study.

22
Single-sex team.
This is a complementary, dichotomous variable to the sex of first
author with yes/no coding, also intended to gauge a possible bias in publications with
authors consisting of only one sex.
No. of authors.
A larger number of authors may be less susceptible to biases.
Publication status.
This is a dichotomous variable with peer-reviewed/not peer-
reviewed coding. Note that for the peer-reviewed publications, the actual data on sex
difference in HPA may not have been reported (e.g., was not paramount to the study), but
there was still sufficient information to conclude that such data exist, and could be
obtained. This moderator is one of the most common methods to assess publication bias.
Data availability in the peer-reviewed publication.
This moderator aims to examine
possible biases within peer-reviewed publications using yes/no coding that denotes
whether sufficient data to calculate an effect size was included in a peer-reviewed
publication (when applicable). In several cases, the peer-reviewed study did not contain
the relevant data, though such data existed. The relevant information was obtained by
contacting the authors directly.
Overall data publication status.
This dichotomous moderator codes whether sufficient
data on sex differences was included in a peer-reviewed publication, compared to all
other manuscripts. Coding of ‘yes’ marks that data were published in a peer-reviewed
paper and available for analysis. Coding of ‘no’ means that the data were either not
included in a peer-reviewed publication but could still be accessed, or that it was not
peer-reviewed (thesis, dissertation or unpublished data). This moderator could be viewed
as a more precise estimate of publication bias, as it distinguishes between data that were
fully reported in a peer-reviewed publication and all other data.

23
Sample group.
This is a dichotomous variable with college students/non-college
categories. We grouped the non-student samples under one category as they included
many different groups (high school students, professional and amateur stand-up
comedians and comedy writers, candidates for a tour guiding course abroad, online
participants, and adolescent inpatients). Two studies that included a mix of students with
other participants were excluded from the sample group moderator analysis.
Sample country.
Similar to the affiliation of first author moderator, this categorical
variable included the following countries/regions: North America (US & Canada),
Europe (Austria, Belgium, Germany, Hungary, United Kingdom), Israel, and Worldwide.
These comparisons will allow to test if there are country specific effects.
Mean sample age.
This quantitative variable aimed at testing whether effects are
consistent across different age means. For studies reporting a range of the participants’
ages, the middle point was used to denote their mean age.
Sample size.
The total number of participants is used to estimate publication bias.
When publication bias is present, smaller sample sizes are associated with larger effect
sizes.
Humor creation task.
Most humor creation tasks employ the cartoon/picture caption
paradigm, with no other task being common enough to comprise a category of its own.
Therefore, ‘new caption’ denotes studies where participants were introduced with a
captionless cartoon or picture, and were instructed to produce funny captions, and ‘other’
indicates studies with all other humor creativity tasks. The ‘other’ category included tasks
based on verbal stimuli, in which participants were asked to complete a joke or a sentence
in a funny way, write a funny story, write a funny resume or profile, write a funny

24
definition for an absurd term, narrate a film in a funny way, or a composite score based
on various non-caption stimuli. In a few cases, participants were asked to produce new
captions in addition to other tasks and thus were coded as ‘both’. This moderator can
shed light on whether the type of the production humor task has any effect on sex
differences in HPA.
No. of humor creation tasks.
This is a quantitative variable denoting the number of
humor creation items that served as a stimuli. More items may allow for more flexibility
in the responses, and to express higher quality of humor.
No. of responses per task.
This dichotomous variable coded ‘one’ vs. ‘multiple’
compares humor creation tasks that require the participants to produce only one funny
outcome per each stimuli vs. studies which ask to produce multiple outcomes (usually
unlimited number). The moderator allows us to compare humor tasks that limit
participants to their best possible funny output, to those which give participants the
freedom to try various attempts at being funny.
No. of levels in funniness scale.
Most researchers use a numeric scale (e.g., 1-7), but
for the studies using a dichotomous variable (e.g. funny vs. non-funny) the variable was
coded as two. This moderator could illuminate whether effect sizes are more or less
pronounced depending on the number of levels in the scale.
Time limited.
This dichotomous variable with yes/no categories indicates whether
participants had a limited time to produce the humorous output, or were given unlimited
time to do so. Time restriction may reduce the quality of the humor, as there is less time
to think, and more pressure to produce. On the other hand, giving participants unlimited

25
time to write funny responses may induce fatigue or boredom. These time effects may
differ by sex, so we will use it as a moderator.
Average task time.
This moderator complements the time limited variable and is more
accurate in evaluating the effect of time on participant’s performance by measuring in
minutes the average time participants were allocated to complete the humor production
task (only for time-limited tasks). The moderator was calculated by dividing the total task
time by the number of tasks participants had to complete. Sex differences in HPA may be
more or less pronounced as a function of the amount of time allocated for each task.
No. of judges.
A higher number of judges are likely to create a more precise
evaluation of the humor produced.
Male to female judge ratio.
This is a quantitative variable representing the ratio of
male to female judges. Values above 1 mean that there were more male raters than
females. This moderator can help explain if sex differences in HPA are due to
disproportional number of male or female judges.
We used Comprehensive Meta-Analysis software to analyze the data (Borenstein,
Hedges, J., & Rothstein, 2005), and our figures were produced with the JASP software
(JASP Team, 2018). We first describe our overall analytic approach, overview the studies
and their effect sizes, and compute the combined effect. We then address the possibility
of publication bias, and report results for all moderators’ analyses.
We analyzed the data based on both fixed- and random-effects models and the results
were comparable to each other, with similar conclusions. Thus, we only report the

26
random-effects results, which allow a more generalized inference about the mean of a
distribution of effect sizes, and not just a single effect size (Card, 2015, p. 230-256).
Random-effects models also require fewer assumptions about the statistical model and
yield more conservative estimates, and thus are usually preferred over fixed-models
(Borenstein, Hedges, Higgins, & Rothstein, 2009, p. 61-86).
3. Results
Tables 2 and 3 present an overview of the studies included in the meta-analysis, along
with the estimated effect sizes and moderators’ codings. We estimated the overall effect
size of sex difference in HPA across 36 independent samples (N = 5057). Results from
the random-effects model analysis show a mean estimated d = 0.321, 95% CI [.237,
.405], Z = 7.46, p < .0001, see Figure 3. Put in a different way, using the standard normal
cumulative distribution function, the combined effect size of d = 0.321 indicates that 63%
of men score above the mean HPA of women (Card, 2015, p. 124). Homogeneity test was
significant, Q(34) = 56.14, p < .02, with I
2
= 37.66. This result indicates that effect sizes
have a small to moderate amount of heterogeneity, i.e., not estimates of a single
population value, thus adding further justification for the use of random-effects models.
Figure 3 presents a forest plot of all independent effect sizes and the overall combined
effect size, along with their respective 95% confidence intervals (CI), and each sample’s
relative weights. Effect size ranged from -.13 to 1.13, with four negative effect sizes and
32 positive ones.

27
Figure 3. Forest plot displaying effect sizes with 95% confidence intervals and relative
weights, as well as the combined effect. All estimates are based on a random-effect
model.

28
3.1. Assessing Publication Bias
We took several steps to reduce the risk of publication bias and to test for its existence.
We first addressed the issue in the initial stages of the meta-analysis, during the literature
search, in an attempt to minimize publication bias. Additionally, we report analyses that
tested whether such bias exists after the data were collected.
3.1.1. Publication status
In an attempt to reach as many studies as possible and minimize publication bias, we did
not explicitly search for sex or gender differences, and did not include the words ‘sex’ or
‘gender’ in our searches. It is possible that studies focusing on sex differences (and
hence, include ‘sex’ or ‘gender’ as key words), are more likely to report that such
differences exist. By using broader search terms, we were able to minimize this bias and
still access many studies that contained relevant data on sex differences in HPA. These
data were either ancillary to the main findings or were not reported at all. In total, 12 of
the 29 peer-reviewed papers used in the meta-analysis did not include sufficient data on
sex differences, but the data were later obtained from the authors. Overall, 20 of the 36
effect sizes in the current meta-analysis (61% of the participants) were not previously
published, either coming from unpublished manuscripts (e.g. theses), or retrieved from
authors of published papers that did not report the data. Additionally, from all 16 effect
sizes published in peer-review publications, seven showed significant sex differences
results and nine did not. Thus, whether sex differences in HPA existed did not seem to be
a major factor determining the publication of the paper, minimizing the file drawer
problem (Rosenthal, 1979).

29
3.1.2. Funnel plot
One of the most common ways to evaluate publication bias is by graphically displaying
effect sizes against their standard errors through a funnel plot (Card, 2015, p. 263-266).
Visual inspection of the funnel plot in Figure 4 reveals a roughly symmetric distribution,
with no obvious outliers. The figure shows lower variability in effect sizes for larger
samples, as one would expect if no publication bias exists. In addition to visually
inspecting the funnel plot, we conducted two tests for the asymmetry of the funnel plot.
The Begg and Mazumdar rank correlation test for funnel plot asymmetry tests the
correlation between effect sizes and standard errors using Kendall’s rank correlation
coefficient (Begg & Mazumdar, 1994). Results show no evidence for asymmetry (Tau =
.117, Z = 1.01, p = .31). The second test for the funnel plot asymmetry was Egger’s
regression test (Egger, Smith, Schneider, & Minder, 1997). As with the previous test,
there was no indication for publication bias (b intercept = -.028, 95% CI [-1.187, 1.129]),
t(34) = .050, p = .96. Overall, these tests and the graphs reveal no evidence of publication
bias due to small sample sizes having large effects.

30
Figure 4.
Funnel plot of effect sizes against their standard errors. The solid line indicates
the combined effect size. White area represents 95% of CI
.
3.1.3. Trim and fill analysis
The trim and fill method is a procedure to correct the estimate of the combined effect size
by imputing potentially missing effect sizes to make the funnel plot symmetric (Duval &
Tweedie, 2000). A new combined estimate is produced by adding the effect sizes that
would fall left of the mean. This method identifies how many studies are missing and
what their effect sizes are. However, our analysis revealed that no missing studies were
identified and needed to be added, resulting in the same adjusted point estimate and

31
confidence interval as the main results (d = 0.321, 95% CI [.237, .405]). Thus, this
analysis showed no evidence of publication bias.
3.1.4. Normal Q-Q plot
Another graphic examination for publication bias is the normal quantile plot. In this plot,
observed quantiles are plotted against the normal distribution quantiles. If the observed
data are normally distributed, we would expect the points on the plot to fall roughly on a
straight line (Wang & Bushman, 1998). Figure 5 displays the Normal Q-Q plot. As can
be seen from the graph, the points are roughly on a straight line, and all effects fall within
95% of the normal line. These results show no evidence of the data deviating from a
normal distribution, or of publication bias.

32
Figure 5.
Normal quantile plot. The diagonal line represents normal distribution of the
dots close to the diagonal line indicate normal distribution of effect sizes. Dashed lines
mark 95% CI.
3.2. Quantitative Moderator Analyses
Moderator analyses attempt to explain the heterogeneity of the effect sizes and identify
variables contributing to a lower or higher combined effect size (Card, 2015, p. 198-228).
To assess the impact of moderators on the combined effect size, we first ran a series of
univariate meta-regressions for each of the nine quantitative moderator variables. In these
tests, we predict the combined effect size from every moderator variable. All models
included an intercept and test whether the coefficient is equal to zero. Table 4 reports the

33
range of the results, the median, the coefficient b, 95% CI, Z scores, Q values, and p-
values. As can be seen from the table, none of the moderators was significant (all tests are
for p-value of .05). Most notably, the variable sample size is a commonly used moderator
for testing publication bias. If publication bias exists, and the combined effect size is
positive, we would expect to find a negative association between sample size and effect
size, meaning that a small sample size produces large effect sizes (Card, 2015, p. 266).
Results showed no association between sample size and effect size, thus there is no
evidence for publication bias.
Though none of the moderator analyses was significant, we further visually inspected the
distribution of every moderator against the effect sizes to see if any trends emerged.
Three moderators had unusual range. The first, number of humor creation tasks, ranged
from 1-34, with a quarter of studies administering more than 16 tasks. However, the
moderator had no discernible correlation with effect size. Second, there were four studies
with 30 or more judges, but their effect size ranged from -0.07 to 0.68, showing no
particular trend. Third, funniness scales ranged from 2-10. As with the previous two
moderators, there was no apparent association between scale’s range and effect sizes.
3.3. Categorical Moderation Analyses
Using mixed-effects categorical moderator analysis (Borenstein et al., 2009), we
calculated effect sizes and 95% CI for each level of all 11 categorical moderators. Results
are displayed in Table 5, along with all between group heterogeneity tests. Of all 11
variables tested, only one moderator, number of responses per task, was significant with

34
p = .012. As three samples included in this variable had no available data, a further
analysis directly comparing the 15 studies that used one response per task to the 18
studies that included multiple responses per task, was performed. Results were still
significant (Q = 5.213, p = .022), meaning that studies that included humor creation tasks
which required only one response per stimuli showed a larger sex difference in HPA (d =
0.425), compared to studies that allowed for multiple responses per humor creation task
(d = 0.220), with both effect sizes still significant (p < .0001). Additionally, the
moderator time limited, had a p-value of .056. Since it included one study that was both
time limited and not (with two different tasks), and six studies where data were unknown,
we decided to compare No and Yes levels directly. Analysis revealed no differences
between these two groups (Q = 3.323, p = .068).
Following Borenstein et al. (2009) we planned to perform a multiple regression analysis
only on the moderators that were found significant in the moderator analysis. However,
as only one moderator was significant (number of responses per task), there was no need
for further analysis. However, as the only significant result was found in one moderator
out of 20 moderator tests, it is possible that such a result is reached by chance because of
the multiple tests (Type I error). Thus, we need to interpret the results of this moderator
with caution.
4. Discussion
The present meta-analysis provides the first comprehensive quantitative evaluation of sex
differences in humor production ability. We sought to collect all available studies

35
assessing new humor outputs (mostly verbal) rated for funniness by independent judges
blind to any characteristic of the producer. We believe that these types of humor
production tasks and the method by which they are evaluated reflect the most objective
measure of true humor abilities. Results reveal a small to moderate effect, with men
scoring higher than women.
We took several steps to address and minimize the possibility of publication and other
potential biases. First, we used random-effects models and weighted means to estimate
the combined effect size, a more conservative approach for analyzing the data
(Borenstein et al., 2009; Card, 2015). Second, our statistical analyses and tests, together
with visual inspections of funnel and Normal Q-Q plots, showed no indication of
publication bias. Overall, more than 60% of the data (20 out of 36 samples) were
unpublished, either from unpublished datasets, theses and dissertations, or based on peer-
reviewed publications where the data did not appear in the paper. The latter consists of 12
samples (44% of the published data), and included studies that did not focus on sex
differences in HPA. Thus, it is unlikely that the presence or absence of sex differences in
HPA played any role in the decision to publish the study. In fact, the combined effect size
for peer-reviewed publications was almost identical to that of not peer-reviewed studies
(.333 and .280 respectively). Moreover, given the debate surrounding sex differences in
HPA, it is unclear what type of bias might have existed. Results suggesting that men and
women do not differ in their HPA are as important and informative as data indicating that
sex differences in HPA do exist. Indeed, roughly half of the peer-reviewed publications
revealed no significant sex differences, while the other half showed men having higher

36
HPA. Thus, it is unlikely that the reason an unpublished manuscript was not published
had anything to do with the presence or absence of sex differences in HPA. In sum, all
indications are that there is little evidence for publication bias, and our view is that such
bias is unlikely.
Third, to address the possibility of additional biases, we included in our meta-analysis a
large number of moderators that could potentially influence the results. One aim was to
test whether there is any bias
stemming from the authors’ or judges’ sex. We found no
evidence for such biases. The sex of the first author, and whether a single or mixed sex
team conducted the study, had no influence on the results. In addition, despite variation in
the male to female judge ratio across studies, the overall number of male and female
judges was nearly identical, 171 and 169 respectively (based on all studies that reported
the figure, 26 out of 36 samples). A moderator test revealed no effect for this variable on
the combined effect; thus, it seems unlikely that judges’ sex had any impact on the
results.
Only one of the variables yielded a significant effect on the estimated combined effect
size. This moderator, number of responses per task suggests that asking participants to
produce one humorous output may result in higher sex differences in HPA for men,
compared with tasks that allow for multiple humour outputs. This finding may indicate
that men might have a further advantage when asked to produce their best humor once,
while women may be funnier when given the chance to create multiple responses. Still, it

37
is important to remember that in either scenario, men still scored significantly higher than
women.
Though our moderator analyses did not reveal other statistically significant moderators’
effects, some results may still illuminate important trends in the sources of sex
differences in HPA. In particular, an important aspect of all humor creation tasks is
whether the participants produced humor under time constraints or not. The results
showed that having unlimited time to complete the task was associated with larger sex
differences in HPA, compared to tasks restricted in time. Having unlimited time to create
humor may have reduce the stress involved in creating humor, and allow participants
more time to think about the task, something that benefited men more than women.
However, it is also important to note that for all studies where participants were limited in
the amount of time to produce humor, more time was not associated with larger sex
differences. In other words, just knowing that the time is limited seemed to have some
effect on the magnitude of sex differences.
It was also interesting to find some null moderator effects. For example, sex differences
remained similar in more than 40 years of research, though there are relatively few
studies prior to 2000 for which the data were available. Similarly, the results were
surprisingly similar across cultures and samples, with college students and non-college
students showing almost identical sex differences. Nonetheless, these results are
provisional, as researchers may find different results in samples that are more diverse in
the future.

38
4.1. Explaining Sex Differences in Higher Humor Production Ability
Sex differences in HPA found in this study may reflect evolved sex differences in mating
preferences and strategies that were shaped by sexual selection theory (Darwin, 1871).
Women’s higher parental investment and larger reproduction costs than men make them
choosier when selecting a mate, and more attentive to traits that could result in higher
fitness (Buss, 2016; Trivers, 1972). HPA could be one such trait, as people vary in their
ability to produce humor, and it is a reliable, hard to fake signal of intelligence, a highly
desirable trait that increases fitness and serves as a mental fitness indicator to attract
mates (Christensen et al., 2018; Greengross & Miller, 2011; Howrigan & MacDonald,
2008; Kaufman, DeYoung, Reis, & Gray, 2011; Miller, 2000a, 2000c). As women are
choosier than men, we would expect women to be more sensitive and attuned to men’s
display of high HPA. As a consequence, a stronger intra-sexual competition among men
ensues, resulting in an overall higher average HPA for men (Bressler & Balshine, 2006;
Bressler et al., 2006; Miller, 2000a, 2000b). Based on this logic, when selecting a mate,
men should use humor more often and more creatively to attract women and signal their
mate value, while women should be more sensitive to men producing high quality humor.
Various research supports this theory, and the view that HPA is valued differently and
divulges disparate information for men and women. Compared to men, choosier women
value humor as a more important trait when selecting a mate, while men make more
effort to impress women and advertise their humor ability, including in real ecological
settings, such as dyadic conversations and on dating sites (Lippa, 2007; Provine, 1993;
Sprecher & Regan, 2002; Todosijević et al., 2003; Wilbur & Campbell, 2011). Women

39
also prefer a man with higher HPA, while men are more attracted to a woman that laughs
at their humor, rather than a woman with high HPA, as smiles and laughter signal the
woman may have a romantic interest in them (Bressler & Balshine, 2006; Bressler et al.,
2006; Hone, Hurwitz, & Lieberman, 2015). Viewing HPA as a mental fitness indicator
relies on the connection between HPA and intelligence (Miller, 2000a, 2000c), and
numerous studies have shown positive correlations between the two attributes
(Christensen et al., 2018; Greengross & Miller, 2011; Howrigan & MacDonald, 2008;
Kellner & Benedek, 2017).
Nevertheless, while a universal phenomenon, humor varies across cultures and reflects
societal norms (T. Jiang, Li, & Hou, 2019). Little is known about sex differences in any
facet of humor among non-Western populations, thus, the universality of sex differences
in HPA found in our meta-analysis should be taken with caution. Most of our data come
from WEIRD (Western, Educated, Industrialized, Rich, and Democratic) countries,
which may delineate only a fraction of all human populations (Henrich, Heine, &
Norenzayan, 2010b). Often, findings that are true for WEIRD samples do not replicate in
non-WEIRD populations (Gurven, Von Rueden, Massenkoff, Kaplan, & Lero Vie, 2013;
Henrich, Heine, & Norenzayan, 2010a; Henrich et al., 2010b). It is possible that non-
WEIRD countries use and experience humor differently, which could influence the
direction and magnitude of sex differences in HPA. If we want to draw conclusions that
will be applicable to all humans, as the evolutionary explanation suggests, we need
information about more diverse populations. For example, research suggests that East
Asian people, such as from China and Taiwan, laugh less, and view themselves as less

40
funny, compared to Western cultures such the United States and Canada (F. Jiang, Yue,
& Lu, 2011). Western countries generally tend to value humor more, perceive humor as a
more socially desirable trait, and view humorous people more positively, compared to
Eastern countries (Yue, Jiang, Lu, & Hiranandani, 2016).
It is possible that the observed difference between men’s and women’s HPA is an artifact
of the fact that certain types of humor, the ones that are considered of high quality, are
more freely expressed by men than by women. For example, men may feel no restrictions
in telling sexual and aggressive jokes, while women may be more inhibited in the use of
these types of humor. However, some research suggests that while women are less likely
than men to tell jokes in general, when they do, women are just as likely as men to use
sexual and aggressive themes (Johnson, 1991). Mickes et al. (2012) compared the themes
that men and women used when producing humor using the cartoon captioning task.
Their results showed that when generating the humorous captions, men produced
significantly more sexual humor and used more profanity than women, though the overall
usage of such humor was low (4.30% for men, 1.95% for women). More importantly, the
use of sexual humor and profanity did not give men any advantage and did not contribute
to their total higher HPA ratings compared to those of women. Also, neither men nor
women judges rated these types of humor as funnier. The authors concluded that the
higher humor ability of men could not be attributed to the use of sexual humor and
profanity, but to other factors.

41
It is also important to note that despite the belief held by many that women enjoy sexual
and aggressive humor less than men, reviews of the literature show mixed support to such
claims (Lampert & Ervin-Tripp, 1998; Martin, 2014). Earlier research found women to
enjoy sexual and aggressive humor less than men (Lampert & Ervin-Tripp, 1998; Martin,
2014), however, many of the jokes and cartoons used in these studies portray women as
the target of the jokes, and the jokes tended to be sexist. When the targets of the jokes are
men, or the jokes are not sexist or have neutral themes, women and men express similar
levels of appreciation to the humor (Lampert & Ervin-Tripp, 1998). In our analysis there
is little evidence to suggest that men and women judges evaluated the humor produced by
either sex differently. Five studies in our meta-analysis tested for sex differences between
judges’ ratings, and only one found a significant difference (Mickes et al., 2012). Mickes
et al. (2012) reported that both male and female raters judged men’s HPA as higher than
women’s, but male raters gave male participants slightly higher ratings than female
raters. In contrast, four other studies that tested for sex differences in judges’ rating did
not find any significant differences (Brodzinsky & Rubien, 1976; Greengross, Jones, &
Sanoudaki, 2017; Greengross, Martin, & Miller, 2012; Greengross & Miller, 2011).
Given the small number of studies to date, the role of judges clearly deserves more
attention in future research.
There is some evidence that sexual and aggressive stimuli used to elicit the humor
production may have an effect on the overall magnitude of sex differences in HPA.
Brodzinsky and Rubien (1976) asked participants to produce spontaneous new humor in
response to captionless cartoons that contained either sexual, aggressive, or neutral

42
themes. Sex differences in HPA were found for sexual and aggressive themes, but not for
the neutral cartoons. With the exception of Brodzinsky and Rubien (1976), no other study
in our analysis reported whether the stimuli were sexual or aggressive in nature.
Another possible explanation for the observed difference lies in the nature of the task
itself. The typical HPA task requires a crisp, focused response. Some research suggests
that men are more likely to tell jokes, while women prefer telling funny stories and
anecdotes (Crawford & Gressley, 1991). Although most tasks included in the meta-
analysis do not exactly imitate a traditional canned joke structure, they are fairly
constrained and do not afford longer, narrative-oriented responses. In addition, joke-
telling is relatively a small fraction of everyday use of humor, where most humor arises
during spontaneous social interactions (Martin & Kuiper, 1999; Provine, 2000). Varying
the types of contexts of HPA tasks seems like a particularly promising direction for future
work.
4.2. Limitations and Suggestions for Future Research
As with any meta-analysis, there is always a possibility that more data exists. In fact, we
know with certainty about several studies that matched our inclusion criteria, but for
which the data could not be retrieved. Many researchers included measures of HPA in
their study but did not report the relevant information, mostly because they did not focus
on studying sex differences in humor. Most of these data come from older studies, and
the raw data are now lost. Nonetheless, as our analysis on publication bias revealed,

43
adding more data is unlikely to change the overall results. Still, we welcome any new
studies on this topic, perhaps with new methodologies for measuring HPA.
Our meta-analysis included studies that used verbal humor as their main measure of
HPA. It is possible that by focusing on non-verbal humor, the results would have been
different. Another limitation is that the measures of HPA included in our meta-analysis
are somewhat artificial, and do not represent everyday production of humor. Requesting
people to produce humor on demand is challenging, and perhaps disadvantaged women
more than men. It also ignores the social context in which most humor is produced
(Provine, 2000), context that if taken into account may benefit women more than men.
Perhaps sex differences in HPA vary depending on the environment in which it is
produced. For example, women may have equal HPA scores as men when interacting
with other women. Thus, studying various dyadic interactions of men and women in more
ecologically valid situations, such as natural conversations, is crucial for fully
understanding when and how sex differences in HPA emerge. Relatively few researchers
conduct these types of studies (Hall, 2015; Provine, 1993, 2000). Still, humor is largely a
social phenomenon and most humor is created in a social context while interacting with
other people. Studying humorous interactions in the lab (Hall, 2015), or observing them
in natural settings (Provine, 1993, 2000) should be a fruitful endeavor that requires more
of our effort.
Another limitation are the ages of participants included in our meta-analysis. The samples
contained somewhat restricted ages, ranging from 15-35, with a median age of 21.7 (see

44
Table 4.). Clearly, such samples are not representative of the whole population, but they
do represent individuals at peak reproductive age. At these ages, following sexual
maturity, people are at peak fertility. This is the period when the competition over mates
is the strongest (Buss, 2016). As a result, and due to women’s choosiness, it might
represent the time when men try to impress women with their humor the most, thus
resulting in higher HPA than women, as our meta-analysis found. Hence, results might be
different for younger or older populations, with different theoretical implications.
4.3. Conclusion
The research presented here focused on one specific aspect of humor that is largely
under-investigated in humor research, humor production ability. Despite finding men to
have higher humor creation abilities than women on verbal humor, this difference should
not be seen as representative of other types of humor, including non-verbal humor
production ability. In fact, for most aspects of humor, men and women seem to exhibit
many similarities, with relatively few differences (Martin, 2014). In regard to humor
production abilities, the topic of sex differences is often reduced to blunt assertions such
as that “Women are not funny” (e.g.,Hitchens, 2007). We hope that our meta-analysis
will help advance a more nuanced discussion on the topic based on a systematic
evaluation of the available scientific data. Examination of such data suggest that
regardless of the underlying source of variability, men exhibit higher humor ability than
women on the kinds of verbal tasks included in our sample of studies. It is important to
remember that though robust, these differences are small to medium in size, and are
based on averages. They do not reflect individual abilities, as both men and women vary

45
largely in their abilities to produce humor. We tried to illuminate possible sources for the
differences in HPA, what they might mean, theoretical implications, considerations for
future research, and limitations. Humor is an important experience for most people, one
that is largely unique to humans. We hope that our results will further foster the study of
humor, advance theories pertaining to understanding and explaining sex differences in
humor and other cognitive abilities, as well as foster research on humor ability.
Footnotes
The study was not preregistered. Data will be shared in a public repository before
publication.
The authors received no financial support for the research, authorship, and/or publication
of this article.
The authors declare no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.

46
References
References marked with an asterisk indicate studies included in the meta-analysis.
Alexander, R. D. (1986). Ostracism and indirect reciprocity: The reproductive significance of humor.
Ethology and Sociobiology, 7(3-4), 253-270. doi:http://dx.doi.org/10.1016/0162-3095(86)90052-X
*Amir, O., & Biederman, I. (2016). The Neural Correlates of Humor Creativity. Frontiers in Human
Neuroscience, 10(597). doi:https://doi.org/10.3389/fnhum.2016.00597
Antonovici, L., & Turliuc, M.-N. (2017). Humour and Mate Selection in a Romanian Sample. Journal of
Experiential Psychotherapy, 20(1), 38-45.
Archer, J. (1996). Sex differences in social behavior: Are the social role and evolutionary explanations
compatible? American psychologist, 51(9), 909-917. doi:http://dx.doi.org/10.1037/0003-
066X.51.9.909
Babad, E. Y. (1974). A multi-method approach to the assessment of humor: A critical look at the humor
tests. Journal of Personality, 42, 618-631. doi:https://doi.org/10.1111/j.1467-6494.1974.tb00697.x
Begg, C. B., & Mazumdar, M. (1994). Operating characteristics of a rank correlation test for publication
bias. Biometrics, 1088-1101. doi:https://doi.org/10.2307/2533446
Borenstein, M., Hedges, L., J., H., & Rothstein, H. (2005). Comprehensive Meta-Analysis Version 3.
Biostat, Englewood, NJ.
Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2009). Introduction to Meta-Analysis.
Chichester, UK: John Wiley & Sons.
Bressler, E., & Balshine, S. (2006). The influence of humor on desirability. Evolution and Human
Behavior, 27(1), 29-39. doi:https://doi.org/10.1016/j.evolhumbehav.2005.06.002
Bressler, E., Martin, R. A., & Balshine, S. (2006). Production and appreciation of humor as sexually
selected traits. Evolution and Human Behavior, 27(2), 121-130.
doi:http://dx.doi.org/10.1016/j.evolhumbehav.2005.09.001
*Brodzinsky, D. M., & Rubien, J. (1976). Humor production as a function of sex of subject, creativity, and
cartoon content. Journal of Consulting and Clinical Psychology, 44(4), 597-600.
doi:http://dx.doi.org/10.1037/0022-006X.44.4.597
Buss, D. M. (1988). The evolution of human intrasexual competition: Tactics of mate attraction. Journal of
personality and social psychology, 54(4), 616-628. doi:http://dx.doi.org/10.1037/0022-
3514.54.4.616
Buss, D. M. (2016). The Evolution of Desire: Strategies of Human Mating (4 edition). New York, NY:
Basic Books.
Card, N. A. (2015). Applied meta-analysis for social science research. New York, NY: Guilford
Publications.
Chafe, W. (1987). Humor as a disabling mechanism. American Behavioral Scientist, 30(1), 16-26.
doi:http://dx.doi.org/10.1177/000276487030003003
*Christensen, A. P., & Silvia, P. J. (2016). Unpublished data.
*Christensen, A. P., Silvia, P. J., Nusbaum, E. C., & Beaty, R. E. (2018). Clever People: Intelligence and
Humor Production Ability. Psychology of Aesthetics, Creativity, and the Arts, 12(2), 136-143.
doi:http://dx.doi.org/10.1037/aca0000109
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (second edition). Hillsdale, NJ:
Routledge.
Crawford, M., & Gressley, D. (1991). Creativity, Caring, and Context: Women's and Men's Accounts of
Humor Preferences and Practices. Psychology of Women Quarterly, 15, 217-231.
doi:https://doi.org/10.1111/j.1471-6402.1991.tb00793.x
Darwin, C. (1871). The Descent of Man, and Selection in Relation to Sex. London: John Murray.

47
Davila-Ross, M., Owren, M. J., & Zimmermann, E. (2009). Reconstructing the evolution of laughter in
great apes and humans. Current Biology, 19(13), 1106-1111.
doi:https://doi.org/10.1016/j.cub.2009.05.028
Duval, S., & Tweedie, R. (2000). Trim and fill: a simple funnel‐plot–based method of testing and adjusting
for publication bias in meta‐analysis. Biometrics, 56(2), 455-463.
doi:https://doi.org/10.1111/j.0006-341X.2000.00455.x
Eagly, A. H., & Wood, W. (1999). The origins of sex differences in human behavior: Evolved dispositions
versus social roles. American psychologist, 54(6), 408-423. doi:http://dx.doi.org/10.1037/0003-
066X.54.6.408
Eaton, A. A., & Rose, S. (2011). Has dating become more egalitarian? A 35 year review using Sex Roles.
Sex roles, 64(11-12), 843-862. doi:https://doi.org/10.1007/s11199-011-9957-9
*Edwards, K. R., & Martin, R. A. (2010). Humor Creation Ability and Mental Health: Are Funny People
more Psychologically Healthy? Europe’s Journal of Psychology, 6(3), 196-212.
doi:https://doi.org/10.5964/ejop.v6i3.213
Egger, M., Smith, G. D., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple,
graphical test. BMJ, 315(7109), 629-634. doi:https://doi.org/10.1136/bmj.316.7129.469
Eysenck, H. J. (1943). An experimental analysis of five tests of "appreciation of humor.". Educational and
Psychological Measurement, 3(1), 191-214. doi:https://doi.org/10.1177/001316444300300119
Feingold, A. (1992). Gender differences in mate selection preferences: A test of the parental investment
model. Psychological Bulletin, 112(1), 125-139. doi:http://dx.doi.org/10.1037/0033-
2909.112.1.125
Feingold, A., & Mazzella, R. (1991). Psychometric Intelligence and Verbal Humor Ability. Personality &
Individual Differences, 12(5), 427-435. doi:https://doi.org/10.1016/0191-8869(91)90060-O
Feingold, A., & Mazzella, R. (1993). Preliminary validation of a multidimensional model of wittiness.
Journal of Personality, 61(3), 439-456. doi:https://doi.org/10.1111/j.1467-6494.1993.tb00288.x
*Freiheit, S. R., Overholser, J. C., & Lehnert, K. L. (1998). The association between humor and depression
in adolescent psychiatric inpatients and high school students. Journal of Adolescent Research,
13(1), 32-48. doi:https://doi.org/10.1177/0743554898131003
Gallup, G. G., Ampel, B. C., Wedberg, N., & Pogosjan, A. (2014). Do Orgasms Give Women Feedback
about Mate Choice? Evolutionary Psychology, 12(5).
doi:https://doi.org/10.1177/147470491401200507
Gamble, J. (2001). Humor in Apes. HUMOR: International Journal of Humor Research, 14(2), 163-179.
doi:https://doi.org/10.1515/humr.14.2.163
*Geher, G., Betancourt, K., & Jewell, O. (2017). The Link between Emotional Intelligence and Creativity.
Imagination, Cognition and Personality, 37(1), 5-22.
doi:https://doi.org/10.1177/0276236617710029
Goddard, C., & Wierzbicka, A. (2004). Cultural scripts: What are they and what are they good for?
Intercultural pragmatics, 1(2), 153-166. doi:https://doi.org/10.1515/iprg.2004.1.2.153
Goodwin, R. (1990). Sex differences among partner preferences: Are the sexes really very similar? Sex
roles, 23(9), 501-513. doi:https://doi.org/10.1007/BF00289765
Greengross, G. (2014). Male production of humor produced by sexually selected psychological adaptations.
In V. A. Weekes-Shackelford & T. K. Shackelford (Eds.), Evolutionary perspectives on human
sexual psychology and behavior. (pp. 173-196). New York, NY: Springer.
*Greengross, G., Jones, M., & Sanoudaki, E. (2017). Enjoyment and production of humour among
bilingual speakers. Unpublished data.
*Greengross, G., Martin, R. A., & Miller, G. F. (2012). Personality traits, intelligence, humor styles, and
humor production ability of professional stand-up comedians compared to college students.
Psychology of Aesthetics, Creativity and the Arts, 6, 74-82.
doi:http://dx.doi.org/10.1037/a0025774
*Greengross, G., & Miller, G. F. (2011). Humor ability reveals intelligence, predicts mating success, and is
higher in males. Intelligence, 39, 188-192. doi:https://doi.org/10.1016/j.intell.2011.03.006
Gurven, M., Von Rueden, C., Massenkoff, M., Kaplan, H., & Lero Vie, M. (2013). How universal is the
Big Five? Testing the five-factor model of personality variation among forager–farmers in the
Bolivian Amazon. Journal of personality and social psychology, 104(2), 354-370.
doi:10.1037/a0030841

48
Hall, J. A. (2015). Sexual Selection and Humor in Courtship: A Case for Warmth and Extroversion.
Evolutionary Psychology, 13(3), 1-10. doi:https://doi.org/10.1177/1474704915598918
Halpern, D. F. (2011). Sex differences in cognitive abilities (4th ed.). New York, NY: Psychology press.
Halpern, D. F., Benbow, C. P., Geary, D. C., Gur, R. C., Hyde, J. S., & Gernsbacher, M. A. (2007). The
Science of Sex Differences in Science and Mathematics. Psychological Science in the Public
Interest, 8(1), 1-51. doi:https://doi.org/10.1111/j.1529-1006.2007.00032.x
Henrich, J., Heine, S. J., & Norenzayan, A. (2010a). Most people are not WEIRD. Nature, 466, 29.
doi:10.1038/466029a
Henrich, J., Heine, S. J., & Norenzayan, A. (2010b). The weirdest people in the world? Behavioral and
Brain Sciences, 33(2/3), 61-135. doi:https://doi.org/10.1017/S0140525X0999152X
Hitchens, C. (2007). Why women aren't funny. Vanity Fair, 557, 54-59.
Hone, L. S. E., Hurwitz, W., & Lieberman, D. (2015). Sex Differences in Preferences for Humor: A
Replication, Modification, and Extension. Evolutionary Psychology, 13(1), 167-181.
doi:https://doi.org/10.1177/147470491501300110
Hooper, J., Sharpe, D., & Roberts, S. G. B. (2016). Are Men Funnier Than Women, or do we just think
they are? Translational Issues in Psychological Science, 2(1), 54-62.
doi:http://dx.doi.org/10.1037/tps0000064
*Howrigan, D. P., & MacDonald, K. B. (2008). Humor as a Mental Fitness Indicator. Evolutionary
Psychology, 6(4), 652-666. doi:https://doi.org/10.1177/147470490800600411
Hull, R., Tosun, S., & Vaid, J. (2017). What's so funny? Modelling incongruity in humour production.
Cognition and Emotion, 31(3), 484-499. doi:https://doi.org/10.1080/02699931.2015.1129314
Hurley, M., Dennett, D., & Adams Jr., R. (2011). Inside Jokes: Using Humor to Reverse-Engineer the Mind
(1 ed.). Cambridge, MA: The MIT Press.
Hyde, J. S. (2005). The gender similarities hypothesis. American psychologist, 60(6), 581-592.
doi:http://dx.doi.org/10.1037/0003-066X.60.6.581
Hyde, J. S. (2014). Gender similarities and differences. Annual review of psychology, 65, 373-398.
doi:https://doi.org/10.1146/annurev-psych-010213-115057
Inglis, V., Zach, S., & Kaniel, S. (2014). Humor Creator and the Audience-A Multi-dimensional Model
Supported by In-vivo Methodology. American Journal of Educational Research, 2(7), 503-512.
doi:https://doi.org/10.12691/education-2-7-12
JASP Team (2018). JASP (Version 0.8.6) [Computer software].
Jiang, F., Yue, X. D., & Lu, S. (2011). Different attitudes toward humor between Chinese and American
students: evidence from the Implicit Association Test. Psychological reports, 109(1), 99-107.
doi:https://doi.org/10.2466/09.17.21.PR0.109.4.99-107
Jiang, T., Li, H., & Hou, Y. (2019). Cultural Differences in Humor Perception, Usage, and Implications.
Frontiers in Psychology, 10(123). doi:https://doi.org/10.3389/fpsyg.2019.00123
Johnson, A. M. (1991). Sex Differences in the Jokes College Students Tell. Psychological reports, 68(3),
851-854. doi:https://doi.org/10.2466/pr0.1991.68.3.851
*Kaufman, S. B. (2016). Unpublished data.
Kaufman, S. B., DeYoung, C. G., Reis, D. L., & Gray, J. R. (2011). General intelligence predicts reasoning
ability even for evolutionarily familiar content. Intelligence, 39(5), 311-322.
doi:https://doi.org/10.1016/j.intell.2011.05.002
Kaufman, S. B., Kozbelt, A., Bromley, M. L., & Miller, G. (2008). The role of creativity and humor in
human mate selection. In G. Geher & G. Miller (Eds.), Mating intelligence: Sex, relationships, and
the mind's reproductive system. (pp. 227-262). Mahwah, NJ, US: Lawrence Erlbaum Associates
Publishers.
*Kellner, R., & Benedek, M. (2017). The role of creative potential and intelligence for humor production.
Psychology of Aesthetics, Creativity, and the Arts, 11(1), 52-58.
doi:http://dx.doi.org/10.1037/aca0000065
*Kim, E., Zeppenfeld, V., & Cohen, D. (2013). Sublimation, culture, and creativity. Journal of personality
and social psychology, 105(4), 639-666. doi:http://dx.doi.org/10.1037/a0033487
Kohler, G., & Ruch, W. (1996). Sources of variance in current sense of humor inventories: How much
substance, how much method variance? HUMOR: International Journal of Humor Research, 9(3-
4), 363-397. doi:https://doi.org/10.1515/humr.1996.9.3-4.363

49
Koppel, M. A., & Sechrest, L. (1970). A multitrait-multimethod matrix analysis of sense of humor.
Educational and Psychological Measurement, 30(1), 77-85.
doi:https://doi.org/10.1177/001316447003000107
Kotthoff, H. (2006). Gender and humor: The state of the art. Journal of Pragmatics, 38(1), 4-25.
doi:https://doi.org/10.1016/j.pragma.2005.06.003
Kozbelt, A., & Nishioka, K. (2010). Humor comprehension, humor production, and insight: An exploratory
study. HUMOR: International Journal of Humor Research, 23(3), 375-401.
doi:https://doi.org/10.1515/humr.2010.017
*Kudrowitz, B. M. (2010). Haha and aha!: Creativity, idea generation, improvisational humor, and
product design. (Doctoral dissertation). Massachusetts Institute of Technology, Boston.
Lampert, M. D., & Ervin-Tripp, S. M. (1998). Exploring Paradigms: The Study of Gender and Sense of
Humor Near the End of the 20th Century. In W. Ruch (Ed.), The Sense of Humor: Explorations of
a Personality Characteristic (pp. 231-270). Berlin: Walter de Gruyter.
*Lehman, K. M., Burke, K. L., Martin, R., Sultan, J., & Czech, D. R. (2001). A reformulation of the
moderating effects of productive humor. HUMOR: International Journal of Humor Research,
14(2), 131-161. doi:https://doi.org/10.1515/humr.14.2.131
Lindberg, S. M., Hyde, J. S., Petersen, J. L., & Linn, M. C. (2010). New trends in gender and mathematics
performance: a meta-analysis. Psychological Bulletin, 136(6), 1123-1135.
doi:http://dx.doi.org/10.1037/a0021276
Lippa, R. A. (2007). The preferred traits of mates in a cross-national study of heterosexual and homosexual
men and women: An examination of biological and cultural influences. Archives of Sexual
Behavior, 36(2), 193-208. doi:http://dx.doi.org/10.1007/s10508-006-9151-2
Lippa, R. A., Collaer, M. L., & Peters, M. (2010). Sex differences in mental rotation and line angle
judgments are positively associated with gender equality and economic development across 53
nations. Archives of sexual behavior, 39(4), 990-997. doi:http://dx.doi.org/10.1007/s10508-008-
9460-8
Martin, R. A. (2014). Humor and gender: An overview of psychological research. In D. Chiaro & R.
Baccolini (Eds.), Gender and Humor: Interdisciplinary and International Perspectives. Berlin:
Mouton de Gruyter.
Martin, R. A., & Kuiper, N. (1999). Daily Occurrence of Laughter: Relationships with Age, Gender, and
Type A personality. HUMOR: International Journal of Humor Research, 12(4), 355-384.
doi:https://doi.org/10.1515/humr.1999.12.4.355
Martin, R. A., & Lefcourt, H. (1983). Sense of humor as a moderator of the relation between stressors and
moods. Journal of personality and social psychology, 45(6), 1313-1324.
doi:http://dx.doi.org/10.1037/0022-3514.45.6.1313
Masten, A. S. (1986). Humor and competence in school-aged children. Child Development, 57(2), 461-473.
doi:http://dx.doi.org/10.2307/1130601
McGee, E., & Shevlin, M. (2009). Effect of humor on interpersonal attraction and mate selection. Journal
of Psychology: Interdisciplinary and Applied, 143(1), 67-77.
doi:https://doi.org/10.3200/JRLP.143.1.67-77
McGhee, P. E. (1974). Development of Children's Ability to Create the Joking Relationship. Child
Development, 45(2), 552-556. doi:https://doi.org/10.2307/1127988
*Mickes, L., Walker, D., Parris, J., Mankoff, R., & Christenfeld, N. (2012). Who's funny: Gender
stereotypes, humor production, and memory bias. Psychonomic Bulletin & Review, 19(1), 108–
112. doi:http://dx.doi.org/10.3758/s13423-011-0161-2
Miller, G. F. (2000a). The mating mind: How sexual selection shaped the evolution of human nature. New
York, NY: Anchor books.
Miller, G. F. (2000b). Mental traits as fitness indicators: Expanding evolutionary psychology's
adaptationism. In D. Lecroy & P. Moller (Eds.), Evolutionary perspectives on human reproductive
behavior (pp. 62-74). New York, NY US: New York Academy of Sciences.
Miller, G. F. (2000c). Sexual selection for indicators of intelligence. In G. Bock, J. Goode, & K. Webb
(Eds.), The nature of intelligence (pp. 260-275): Novartis Foundation Symposium 233. John
Wiley.
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic
reviews and meta-analyses: the PRISMA statement. Annals of internal medicine, 151(4), 264-269.
doi:https://doi.org/10.1371/journal.pmed.1000097

50
*Mollica, M. A. (1983). Paradox Recognition: A Proposed Common Cognitive Process Between Creativity
And Humor (Doctoral dissertation). DePaul, Ann Arbor.
*Moran, J. M., Rain, M., Page-Gould, E., & Mar, R. A. (2014). Do I amuse you? Asymmetric predictors
for humor appreciation and humor production. Journal of Research in Personality, 49, 8-13.
doi:https://doi.org/10.1016/j.jrp.2013.12.002
Nevo, O., Nevo, B., & Yin, J. L. (2001). Singaporean Humor: a Cross-Cultural, Cross-Gender Comparison.
Journal of General Psychology, 128(2), 143-156. doi:https://doi.org/10.1080/00221300109598904
Nusbaum, E. C. (2015). A meta-analysis of individual differences in humor production and personality.
(Ph.D.). The University of North Carolina at Greensboro, Ann Arbor. ProQuest Dissertations &
Theses Global database.
*Nusbaum, E. C., Silvia, P. J., & Beaty, R. E. (2017). Ha ha? Assessing individual differences in humor
production ability. Psychology of Aesthetics, Creativity, and the Arts, 11(2), 231-241.
doi:http://dx.doi.org/10.1037/aca0000086
O'Quin, K., & Derks, P. (1997). Humor and Creativity: A Review of the Empirical Literature. In M. Runco
(Ed.), Creativity research handbook (Vol. 1, pp. 223-252). Cresskill, NJ: Hampton Press.
Preuschoft, S., & Van-Hooff, J. A. (1997). The social function of 'smile' and 'laughter': Variations across
primate species and societies. In U. Segerstrale & P. Molnar (Eds.), Nonverbal communication:
Where nature meets culture (pp. 171-190). Mahwah, NJ: Lawrence Erlbaum Associates.
Provine, R. (1993). Laughter Punctuates Speech: Linguistic, Social and Gender Contexts of Laughter.
Ethology, 95(4), 291-298. doi:https://doi.org/10.1111/j.1439-0310.1993.tb00478.x
Provine, R. (2000). Laughter: A scientific investigation. New York: Viking.
Ramachandran, V. S. (1998). The neurology and evolution of humor, laughter, and smiling: the false alarm
theory. Medical Hypotheses, 51(4), 351-354. doi:https://doi.org/10.1177/147470490600400129
*Renner, K.-H., & Manthey, L. (2018). Relations of Dispositions toward Ridicule and Histrionic Self-
Presentation with Quantitative and Qualitative Humor Creation Abilities. Frontiers in Psychology,
9(78). doi:https://doi.org/10.3389/fpsyg.2018.00078
Rosenthal, R. (1979). The file drawer problem and tolerance for null results. Psychological Bulletin, 86(3),
638-641. doi:http://dx.doi.org/10.1037/0033-2909.86.3.638
Ross, E., & Hall, J. A. (in press). The Traditional Sexual Script and Humor in Courtship. HUMOR:
International Journal of Humor Research.
Ruch, W. (2008). Psychology of humor. In V. Raskin (Ed.), The primer of humor research (Vol. 8, pp. 17-
100). Berlin, Germany: Mouton de Gruyter.
Ruch, W., Beermann, U., & Proyer, R. T. (2009). Investigating the humor of gelotophobes: Does feeling
ridiculous equal being humorless? Humor-International Journal of Humor Research, 22(1-2),
111-143. doi:https://doi.org/10.1515/HUMR.2009.006
*Saroglou, V. (2002). Religiousness, religious fundamentalism, and quest as predictors of humor creation.
International Journal for the Psychology of Religion, 12(3), 177-188.
doi:https://doi.org/10.1207/S15327582IJPR1203_04
*Saroglou, V., & Jaspard, J.-M. (2001). Does religion affect humour creation? An experimental study.
Mental Health, Religion & Culture, 4(1), 33-46.
doi:http://dx.doi.org/10.1080/13674670010016756
*Séra, L., Boda-Ujlaky, J., & Gyebnár, V. (2015). A Humorstílus és a kreativitás különbözo′ aspektusainak
összefüggései [Relationships between different aspects of humor style and creativity]. Magyar
Pszichologiai Szemle, 70(2), 295-312. doi:https://doi.org/10.1556/0016.2015.70.2.1
*Shiloh, S. (1982). Humor Creation: Its Relationship to Individual and Situational Characteristics.
(Doctoral dissertation). Tel Aviv University, Tel Aviv, Israel.
Shlesinger, I. (2017). Girl Logic: The Genius and the Absurdity. New York, NY: Hachette Books.
Smith, J. E., Waldorf, V., A., & Trembath, D. L. (1990). "Single white male looking for thin, very
attractive...,". Sex roles, 23(11/12), 675-685. doi:https://doi.org/10.1007/BF00289255
Spelke, E. S. (2005). Sex Differences in Intrinsic Aptitude for Mathematics and Science?: A Critical
Review. American psychologist, 60(9), 950-958. doi:http://dx.doi.org/10.1037/0003-
066X.60.9.950
Sprecher, S., & Regan, P. C. (2002). Liking some things (in some people) more than others: Partner
preferences in romantic relationships and friendships. Journal of Social and Personal
Relationships, 19(4), 463-481. doi:https://doi.org/10.1177/0265407502019004048

51
Thorson, J. A., & Powell, F. C. (1993). Development and validation of a multidimensional sense of humor
scale. Journal of Clinical Psychology, 49(1), 13-23. doi:https://doi.org/10.1002/1097-
4679(199301)49:1<13::AID-JCLP2270490103>3.0.CO;2-S
Todosijević, B., Ljubinković, S., & Arančić, A. (2003). Mate selection criteria: A trait desirability
assessment study of sex differences in Serbia. Evolutionary Psuchology, 1, 116-126.
doi:https://doi.org/10.1177/147470490300100108
Toro-Morn, M., & Sprecher, S. (2003). A cross-cultural comparison of mate preferences among university
students: The United States vs. the People's Republic of China (PRC). Journal of Comparative
Family Studies, 34(2), 151-170.
*Townsend, J. W. J. (1982). Relationships among humor, creative thinking abilities, race, sex, and
socioeconomic factors of advantagedness and disadvantagedness of a selected sample of high
school students. (Doctoral dissertation). University of Georgia, Ann Arbor. ProQuest Dissertations
& Theses Global database.
Treadwell, Y. (1970). Humor and creativity. Psychological reports, 26, 55-58.
doi:https://doi.org/10.2466/pr0.1970.26.1.55
Trivers, R. L. (1972). Parental investment and sexual selection. In B. Campbell (Ed.), Sexual Selection and
the Descent of Man: 1871-1971 (pp. 136-179). Chicago, IL: Aldine.
Turner, R. G. (1980). Self-monitoring and humor production. Journal of Personality, 48(2), 163-172.
doi:https://doi.org/10.1111/j.1467-6494.1980.tb00825.x
Viana, A. (2017). Humour and laughter as vestiges of evolution. The European Journal of Humour
Research, 5(1), 1-18. doi:http://dx.doi.org/10.7592/EJHR2017.5.1.viana
Voyer, D., Postma, A., Brake, B., & Imperato-McGinley, J. (2007). Gender differences in object location
memory: A meta-analysis. Psychonomic bulletin & review, 14(1), 23-38.
doi:https://doi.org/10.3758/BF03194024
Voyer, D., Voyer, S., & Bryden, M. P. (1995). Magnitude of sex differences in spatial abilities: A meta-
analysis and consideration of critical variables. Psychological Bulletin, 117(2), 250-270.
doi:http://dx.doi.org/10.1037/0033-2909.117.2.250
Voyer, D., Voyer, S. D., & Saint-Aubin, J. (2017). Sex differences in visual-spatial working memory: A
meta-analysis. Psychonomic Bulletin & Review, 24(2), 307-334.
doi:https://doi.org/10.3758/s13423-016-1085-7
Wang, M. C., & Bushman, B. J. (1998). Using the normal quantile plot to explore meta-analytic data sets.
Psychological Methods, 3(1), 46-54. doi:http://dx.doi.org/10.1037/1082-989X.3.1.46
Weisfeld, G. E. (1993). The Adaptive Value of Humor and Laughter. Ethnology and Social Biology, 14(2),
141-169. doi:https://doi.org/10.1016/0162-3095(93)90012-7
Wilbur, C. J., & Campbell, L. (2011). Humor in romantic contexts: Do men participate and women
evaluate? Personality and Social Psychology Bulletin, 37(7), 918-929.
doi:https://doi.org/10.1177/0146167211405343
Yue, X., Jiang, F., Lu, S., & Hiranandani, N. (2016). To Be or Not To Be Humorous? Cross Cultural
Perspectives on Humor. Frontiers in Psychology, 7(1495).
doi:https://doi.org/10.3389/fpsyg.2016.01495
Zell, E., Krizan, Z., & Teeter, S. R. (2015). Evaluating gender similarities and differences using
metasynthesis. American psychologist, 70(1), 10-20. doi:https://doi.org/10.1037/a0038208
Ziller, R. C., Behringer, R. D., & Goodchilds, J. D. (1962). Group creativity under conditions of success or
failure and variations in group stability. Journal of Applied Psychology, 46(1), 43-49.
doi:http://dx.doi.org/10.1037/h0045647
*Ziv, A. (1981a). The psychology of humor. Israel: Yachdav.
*Ziv, A. (1981b). The self concept of adolescent humorists. Journal of Adolescence, 4, 187-197.
doi:https://doi.org/10.1016/S0140-1971(81)80038-3
Ziv, A. (1983). The influence of humorous atmosphere on divergent thinking. Contemporary Educational
Psychology, 8(1), 68-75. doi:https://doi.org/10.1016/0361-476X(83)90035-8

52
Table 1
Excluded Moderators from the Meta-Analysis
Moderator Reason for exclusion
Department of first author Lack of variability: 78% of samples were from a psychology department
Language of publication Lack of variability: 92% of samples were in English
Judges’ identity Lack of variability: 89% of humor was rated by students and professors
Humor fluency Limited data: only 28% of studies reported this variable
Reliability of judges’ ratings
Lack of variability: 80% of the samples with Cronbach’s α reported a
reliability of .7 or higher
Sex differences in judges’ ratings Limited data: only 14% of studies reported this variable
Note. Humor fluency measures men’s and women’s number of responses for studies
where participants were allowed to produce an unlimited number of humor outputs. For
the reliability variable, various reliability measures were calculated, with 43% of the
samples reporting Cronbach’s α, and 31% of samples using other reliability measures

47
Table 2
Overview of Studies, Effect Sizes, and Moderators Included in the Meta-Analysis
Sample size
Study
Men
Women
Effect
size
(d)
SE
Rel.
weig
ht
Affiliation
of First
a
uthor
1
Sex of
first
author
Single
sex
team
No. of
authors
Publication
status
Data
availability
in the PR
publication
Overall
data
publication
s
tatus
Amir & Biederman (2016)
2
Comedians 17 3 0.68 0.64 0.42 NAM Male Yes 2 PR No No
Controls 10 7 0.05 0.49 0.69 NAM Male Yes 2 PR No No
Brodzinsky & Rubien (1976) 40 44 0.48 0.22 2.59 NAM Male No 2 PR Yes Yes
Christensen et al. (2016) 38 232 0.08 0.18 3.52 NAM Male No 4 PR No No
Christensen & Silvia (2016) 42 139 0.66 0.18 3.41 NAM Male No 3 NPR N/A No
Edwards & Martin (2010) 92 123 0.23 0.14 4.53 NAM Female No 2 PR Yes Yes
Freiheit et al. (1998)
Adolescent inpatients 23 32 1.13 0.29 1.69 NAM Female No 3 PR Yes Yes
High school students 43 42 0.19 0.22 2.66 NAM Female No 3 PR Yes Yes
Geher et al. (2017) 26 65 0.27 0.23 2.41 NAM Male No 3 PR No No
Greengross & Miller (2011) 200 200 0.37 0.10 5.84 NAM Male Yes 2 PR Yes Yes
Greengross et al. (2017) 38 79 0.43 0.20 2.99 Europe Male No 3 NPR N/A No
Greengross et al. (2012) 42 8 0.15 0.39 1.07 NAM Male Yes 3 PR No No
Howrigan & MacDonald (2008) 70 150 0.40 0.15 4.29 NAM Male Yes 2 PR Yes Yes
Kaufman (2016) 103 642 0.41 0.11 5.62 NAM Male
NPR N/A No
Kellner & Benedek (2017) 41 110 -0.09 0.18 3.33 Europe Female No 2 PR No No
Kim et al. (2013) 34 62 0.41 0.22 2.69 NAM Female No 3 PR No No
Kudrowitz (2010) 52 32 0.18 0.22 NAM Male Yes 1 NPR N/A No
Lehman et al. (2001) 21 39 0.36 0.27 1.90 NAM Female No 5 PR Yes Yes
Mickes et al. (2012) 16 16 0.24 0.35 1.23 NAM Female No 5 PR Yes Yes
Mollica (1983)
3
21 49 0.29 0.26 2.02 NAM Male Yes 1 NPR N/A No

48
Table 2 (continued)
Sample size
Study
Men
Women
Effect
size
(d)
SE
Rel.
weight
Affiliation
of First
a
uthor
Sex of
first
author
Single
sex
team
No. of
authors
Publication
status
Data
availability
in the PR
publication
Overall
data
publication
s
tatus
Moran et al. (2014) 66 93 0.29 0.16 3.85 NAM Male No 4 PR Yes Yes
Nusbaum et al. (2017)
Study 1
38
125
0.19
0.19
3.26
NAM
Female
No
3
PR
No
No
Study 2 44 116 0.33 0.18 3.44 NAM Female No 3 PR No No
Study 3 45 93 0.37 0.18 3.33 NAM Female No 3 PR No No
Renner & Man
they (2018)
83
254
0.18
0.13
4.91
Europe
Male
No
2
PR
No
No
Saroglou & Jaspard (2001)
Humorous video 12 15 0.93 0.41 0.97 Europe Male No 2 PR Yes Yes
No video/control
8
21
0.35
0.42
0.93
Europe
Male
No
2
PR
Yes
Yes
Religious video 12 17 0.54 0.38 1.08 Europe Male No 2 PR Yes Yes
Saroglou (2002) 18 54 0.89 0.28 1.80 Europe Male Yes 1 PR No No
Séra et al. (2015)
43
90
-
0.07
0.19
3.28
Europe
Male
No
3
PR
Yes
Yes
Shiloh (1982)
Study 1 36 57 -0.07 0.21 2.73 Israel Female Yes 1 NPR N/A No
Study 2
36
64
0.33
0.21
2.78
Israel
Female
Yes
1
NPR
N/A
No
Townsend (1982) 47 63 -0.13 0.19 3.12 NAM Male Yes 1 NPR N/A No
Ziv (1981b)
4
162 182 0.69 0.11 5.45 Israel Male Yes 1 PR Yes Yes
Ziv (1983)
Control
5
28 30 0.17 0.26 2.01 Israel Male Yes 1 PR Yes Yes
Experiment 30 32 0.51 0.26 2.07 Israel Male Yes 1 PR Yes Yes
Note. Positive d denotes higher humor ability for men. Rel. weight = relative weight for the random model; NAM = North America;
N/A = not applicable; PR = peer-reviewed; NPR = Non-peer-reviewed.

49
Table 3
Moderators Included in the Meta-Analysis
Study
Sample
group
Sample
country
1
Sample age
mean
Humor
creation
task
No. of
humor
creation
tasks
No. of
responses
per task
Funniness
scale
Time
limited
Average
task time
(minutes)
No. of
judges
Male to
female
judge
ratio
Amir & Biederman (2016)
Comedians NCS NAM 32.05 Caption Variable Multiple 7 Yes 0.25 81 1.45
Controls
6
NAM 24.90 Caption Variable Multiple 7 Yes 0.25 81 1.45
Brodzinsky & Rubien (1976)
CS
NAM
Caption
12
One
6
Yes
5.00
6
1.00
Christensen et al. (2016) CS NAM 19.08 Both 3 One 5 No N/A 3 2.00
Christensen & Silvia (2016) CS NAM 19.10 Other 9 3 No N/A 8 0.60
Edwards & Martin (2010)
CS
NAM
18.58
Both
5
Multiple
5
Yes
2.50
6
0.50
Freiheit et al. (1998)
Adolescent inpatients NCS NAM 15.44 Caption 8 One 5 Yes 1.88 6 1.00
High school students
NCS
NAM
15.13
Caption
8
One
5
Yes
1.88
6
1.00
Geher et al. (2017) NCS World 26.32 Caption 2 One 5 No N/A 3 0.50
Greengross & Miller (2011) CS NAM 20.60 Caption 3 Multiple 7 Yes 3.33 6 0.50
7
Greengross et al. (2017)
CS
8
Europe
21.94
Caption
3
Multiple
5
Yes
3.33
12
1.40
Greengross et al. (2012) NCS NAM 35.46 Caption 3 Multiple 7 Yes 3.33 6 0.50
Howrigan & MacDonald (2008) CS NAM 22.00 Other 11 Multiple 7 No N/A 4
9
1.00
Kaufman (2016)
CS
NAM
24.19
Caption
4
Kellner & Benedek (2017) CS Europe 23.10 Caption 6 Multiple 4 Yes 2.5 10
Kim et al. (2013) CS NAM Caption 5 Multiple Yes 1.2 6
Kudrowitz (2010)
NCS
10
NAM
28
Caption
3
Multiple
3
Yes
1.67
12
9
0.79
Lehman et al. (2001) CS NAM Other 1 5 2
Mickes et al. (2012) CS NAM Caption 20 One 6 Yes 2.25 81 0.72
Mollica (1983)
CS
NAM
Caption
18
Multiple
6
Yes
2.22
5
0.67

50
Table 3 (continued)
Study
Sample
group
Sample
country
Sample age
mean
Humor
creation
task
No. of
humor
creation
tasks
No. of
responses
per task
Funniness
scale
Time
limited
Average
task time
(minutes)
No. of
judges
Male to
female
judge
ratio
Moran
et
al.
(2014)
NCS
World
30.00
Caption
34
One
7
Yes
0.5
4
0.00
Nusbaum et al. (2017)
Study 1
CS
NAM 19.00 Both 3 One 5 Yes N/A 4 1.00
S
tudy 2
CS
NAM
19.00
Both
3
One
5
Yes
2.00
5
0.67
Study 3
CS
NAM 18.70 Both 3 One 5 No N/A 2 1.00
Renner & Manthey (2018)
CS
8
Europe 33.17 Caption 6 Multiple 10 Yes 0.42 3 0.00
Saroglou & Jaspard (2001)
Humorous video
CS
Europe Caption 24 One 2 2 1.00
No video/control
CS
Europe Caption 24 One 2 2 1.00
Religious video
CS
Europe
Caption
24
One
2
2
1.00
Saroglou (2002)
CS
Europe 23.42 Caption 24 One 2 2 1.00
Séra et al. (2015)
11
Europe 29.50 Caption 6 Multiple 9 Yes 2.5 14
Shiloh (1982)
Study 1
NCS
Israel 17.50 Other 4 Multiple 2 Yes 1.00 30
Study 2
NCS
Israel 34.53 Other 1 Multiple 2 Yes 4.00
Townsend (1982)
NCS
NAM
17.00
12
C
aption
4
Multiple
6
Yes
0.75
2
Ziv (1981b)
NCS
Israel 15.50
12
Caption 10 One 6 No N/A 3
Ziv (1983)
Control
NCS
Israel
15.50
12
Caption
20
Multiple
2
No
N/A
Experiment
CS
Israel 15.50
12
Caption 20 Multiple 2 No N/A
Note. Blank indicates no available data; NCS = non-college students; CS = college students; NAM = North America; World =
worldwide; Caption = new caption; Other = Verbal stimuli; Both = new caption and verbal stimuli; N/A = not applicable.

51
Table 4
Univariate Meta-Regressions for Quantitative Moderator Analysis
Moderator K Range Median b 95% CI Z Q(1) p
Sample size 36 17-745 94.5 0.0001 [-0.0004, 0.0006] 0.42 0.18 0.672
Publication year 36 1976-2018 2010.5 -0.0026 [-0.0086, 0.0033] -0.87 0.75 0.386
No. of authors 35 1-5 2 -0.0111 [-0.0966, 0.0743] -0.26 0.07 0.799
Sample age mean 28 15.13-35.46 21.7 -0.0098 [-0.0264, 0.0069] -1.15 1.32 0.250
No. of humor creation tasks 34 1-34 6 0.0089 [-0.0018, 0.0197] 1.63 2.64 0.104
No. of levels in funniness scale 34 2-10 5 -0.0261 [-0.0700, 0.0177] -1.17 1.36 0.243
Average Task Time 21 15 sec-5 min 2 min 0.0566 [-0.0298, 0.1429] 1.28 1.65 0.199
No. of judges 32 2-81 5.5 -0.0032 [-0.0101, 0.0037] -0.91 0.82 0.365
Male to female judge ratio 25 0-2 1 0.0293 [-0.1435, 0.2021] 0.33 0.11 0.739

52
Table 5
Mixed-Effects Categorical Moderator Analysis
Moderator/ Group k Men Women d 95% CI Q p
Main effect 36 1677 3380 0.321 [0.237, 0.405]
Affiliation of first author 0.067 0.967
North America 23 1130 2375 0.326 [0.247, 0.405]
Europe 8 255 640 0.302 [0.053, 0.551]
Israel 5 292 365 0.354 [0.049, 0.659]
Sex of first author 0.823 0.364
Female 12 469 879 0.266 [0.122, 0.410]
Male 24 1208 2501 0.348 [0.245, 0.451]
Single sex team 0.813 0.666
No 21 805 1807 0.298 [0.193, 0.404]
Yes 14 769 931 0.331 [0.175, 0.488]
Unknown 1 103 642 0.405 [0.196, 0.614]
Publication status 0.253 0.615
Not peer-reviewed 8 375 1125 0.280 [0.097, 0.463]
Peer-reviewed 28 1302 2255 0.333 [0.236, 0.430]
Availability in PR publication 2.816 0.093
No 12 436 1129 0.238 [0.114, 0.362]
Yes 16 866 1126 0.392 [0.261, 0.523]
Overall data publication status 2.201 0.138
No 20 811 2251 0.265 [0.160, 0.370]
Yes 16 866 1126 0.392 [0.261, 0.523]
Sample group 4.917 0.086
College students 21 1016 2580 0.337 [0.256, 0.418]
Non-college students 13 608 703 0.316 [0.125, 0.506]
Mixed 2 53 97 -0.059 [-0.399, 0.281]
Sample country 0.175 0.981
North America 21 1038 2217 0.329 [0.240, 0.417]
Europe 8 255 640 0.302 [0.053, 0.551]
Israel 5 292 365 0.354 [0.049, 0.659]
Worldwide 2 92 158 0.282 [0.022, 0.543]
Humor creation task 1.321 0.517
New caption 26 1215 2242 0.337 [0.225, 0.449]
Other 5 205 449 0.353 [0.125, 0.581]
Both 5 257 689 0.235 [0.087, 0.384]
No. of responses per task 8.792 0.012
One 15 591 1147 0.425 [0.277, 0.573]
Multiple 18 920 1413 0.221 [0.128, 0.314]
Unknown 3 166 820 0.460 [0.289, 0.630]

53
Table 5 (continued)
Moderator k Men Women d 95% CI Q p
Time limited? 7.542 0.056
No 8 441 923 0.422 [0.253, 0.591]
Yes 21 1024 1544 0.239 [0.137, 0.340]
Both 1 38 125 0.190 [-0.175, 0.555]
Unknown 6 174 788 0.475 [0.305, 0.646]
Note. PR = peer-reviewed; Other = Verbal stimuli; Both = new caption and verbal stimuli.

54
1
For North American countries (NAM), all are from the US except Edwards & Martin (2010).
2
Humor production was based on recall of captions that were generated under a fMRI scan.
Following the scan, participants were asked to write down the same captions they thought about while
in the machine.
3
The dissertation included two effect sizes, but we calculated the effect size based on only the
averages across all captions. The effect size based on the average of the first caption was excluded
(see also coding procedures).
4
The sd of 7.54 reported in Ziv (1981b) is probably wrong, as it is incongruent with women’s sd of
17.28. Ziv (1981a) reports a sd of 17.54, which we used here.
5
The study included non-significant results, therefore we used a conservative p-value of .51 to
calculate the effect size.
6
Sample group moderator data were excluded as it included a mix of college and graduate students, as
well as faculty members.
7
The original paper mistakenly reported that there were four men and two women judges, where in
fact there were four women and two men judges.
8
Majority of participants were college students.
9
Total number of judges was higher, but this was the number that judged each task.
10
Majority of participants were non-college students.
11
Sample group moderator data were excluded as half the participants were students and half not.
12
Middle point was used for a range of ages.