Journal of Homosexuality

ISSN: 0091-8369 (Print) 1540-3602 (Online) Journal homepage: https://www.tandfonline.com/loi/wjhm20

Beyond the Superordinate Categories of “Gay

Men” and “Lesbian Women”: Identification of Gay
and Lesbian Subgroups

Jessica M. McCutcheon & Melanie A. Morrison

To cite this article: Jessica M. McCutcheon & Melanie A. Morrison (2019): Beyond the
Superordinate Categories of “Gay Men” and “Lesbian Women”: Identification of Gay and Lesbian
Subgroups, Journal of Homosexuality, DOI: 10.1080/00918369.2019.1627129

To link to this article: https://doi.org/10.1080/00918369.2019.1627129

Published online: 25 Jun 2019.

Submit your article to this journal

Article views: 12

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=wjhm20
JOURNAL OF HOMOSEXUALITY
https://doi.org/10.1080/00918369.2019.1627129

Beyond the Superordinate Categories of “Gay Men” and

“Lesbian Women”: Identification of Gay and Lesbian
Subgroups
Jessica M. McCutcheon, PhD and Melanie A. Morrison, PhD
Department of Psychology, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

ABSTRACT KEYWORDS
Nation-wide opinion polls and social scientific studies indicate Gay; lesbian; attitudes;
that evaluations of gay men and lesbian women have become subgroups; stereotypes
increasingly favourable. These positive trends do not explain
the widespread discrimination experiences being reported. To
assist researchers in investigating attitudes towards gay and
lesbian persons, the current research examines whether there
are multiple “types” that are identifiable and salient. Two
Canadian studies (Ns = 67 and 206) were conducted to estab-
lish the presence of gay and lesbian subgroups. Using sub-
groups generated by Study 1 participants, community and
student sub-samples selected those they perceive to exist.
Results indicated that, for gay men, the subgroups Drag
Queen and Flamboyant emerged, as did Butch for lesbian
women. Amongst students, Closeted and Feminine also
emerged for gay men, as well as Feminist and Tomboy for
lesbian women. These findings have implications for contem-
porary research on gay- and lesbian-related attitudes and the
methodology used to assess them.

Introduction
Discrimination against gay and lesbian persons is reported frequently within
health care settings, family relationships (Jewell, McCutcheon, Harriman, &
Morrison, 2011), and governmental service-providing sectors (Mattocks
et al., 2015). Critically, instances of victimization and abuse are prevalent
not only within adult sexual minority populations but are evidenced when
sampling sexual minority youth as well. In surveys of both Canadian
(Morrison, Jewell, McCutcheon, & Cochrane, 2014; Saewyc et al., 2007;
Taylor & Peter, 2011) and American (Kosciw, Greytak, Bartkiewicz,
Boesen, & Palmer, 2012) high school students, over half of the sexual
orientation minority students surveyed reported hearing the homonegative
comments “that’s so gay,” “faggot,” and “dyke,” and being verbally harassed
by other means. Approximately 10–20% of the students also reported being
targets of physical and sexual violence (e.g., evidenced by reports of gay and

CONTACT Jessica M. McCutcheon [email protected] Department of Psychology, University of

Saskatchewan, 9 Campus Drive, Arts Building, Saskatoon, Saskatchewan S7N 5A5, Canada.
© 2019 Taylor & Francis Group, LLC
2 J. M. MCCUTCHEON AND M. A. MORRISON

lesbian students being physically abused, threatened with a weapon, and

sexually assaulted). Swim, Pearson, and Johnston (2007) emphasize that the
discrimination experienced by sexual minority children often occurs regu-
larly. Of the 69 sexual minority youth who were asked to keep a diary of their
discrimination experiences during a one-week period, results indicated that,
on average, lesbian and gay students experienced two homonegative inci-
dents in the form of homonegative verbal comments (e.g., hearing anti-gay
jokes) and homonegative behaviours (e.g., exclusion, rude gestures, and/or
fear of being outed).
Survey data illustrate a decidedly more “liberalized” attitudinal trend
toward gay men and lesbian women than research focusing on their dis-
crimination experiences. Data from the General Social Survey (GSS) in
Canada revealed that, from 1988–2006, opposition to same-sex marriage
decreased by 19 percentage points (Baunach, 2012). American polling data
(i.e., Gallup 1977–2011; and the 2000 National Election Study) showed
decreases of a similar magnitude in the following domains: support for
adoption by same-sex couples, policies tackling job discrimination against
sexual minorities, and service by sexual minority persons in the military
(Hicks & Lee, 2006).
The trend toward greater “liberalization” has been echoed in myriad
academic studies, particularly amongst those using valid indicators (i.e.,
psychometrically sound scales) of global acceptance of sexual minority per-
sons. For instance, Altemeyer (2001) assessed Canadian university students’
attitudes toward sexual minority individuals in 1984 and then repeated the
assessment in 1998. As noted, scores on the Attitudes toward Homosexuals
Scale (ATHS; Altemeyer, 1988) hovered around the ATHS midpoint in 1984;
yet, decreased substantially fourteen years later. Recent studies (e.g., Hirai,
Winkel, & Popan, 2014; Stotzer, 2009) that have employed the Attitudes
toward Lesbians and Gay Men Scale (ATLG; Herek, 1988), a gold standard
measure in the field of attitudinal homonegativity, have shown vast improve-
ment from the degree of prejudice documented 20 years earlier. Finally, even
for studies assessing contemporary forms of bias toward gay and lesbian
individuals using “modern” measures of homonegativity (where attitudes are
not a function of moralistic or religious objections to homosexuality) evi-
dence similar decrements. Since Morrison and Morrison (2003) first pub-
lished their seminal research on the development and validation of the
Modern Homonegativity Scale (MHS), wherein Canadian students’ scores
on the MHS were situated above the midpoint signifying their endorsement
of a “subtler” form of homonegativity, studies published almost a decade
later (e.g., McDermott & Blair, 2012; Morrison & Morrison, 2011) now
situate Canadian participants’ scores below the MHS midpoint. To further
understand the discrimination experiences of sexual minority persons during
a time of attitudinal liberalization, researchers have begun to consider what
 JOURNAL OF HOMOSEXUALITY 3

additional complexities may exist that affect attitudes towards these evalua-
tive targets.
Research on cognition (e.g., Brewer, Dull, & Lui, 1981; Eckes, 1994; Park,
Ryan, & Judd, 1992; Richards & Hewstone, 2001) indicates that “perceivers
process group-related information at multiple levels, ranging from the
broader (i.e., overall category) to more specific (i.e., subgroup) levels”
(Brambilla, Carnaghi, & Ravenna, 2011, p. 101). It is logical to intuit, there-
fore, that individuals may possess unique attitudes toward different “types” of
gay men or lesbian women, wherein some subgroups might evoke primarily
positive attitudes, and others might evoke negative ones. Importantly, if these
distinct groups become obscured when researchers use superordinate cate-
gories (e.g., the generic terms “gay men” and/or “lesbian women,” terms
found within almost every anti-gay or -lesbian measure), interpretations of
study results may be skewed at best, or flawed at worst. Indeed, superordinate
categories of sexual minorities (e.g., “effeminate” gay men or “butch” lesbian
women) may be suppressed.
On the basis of a comprehensive review of the literature, we could locate
only three studies that have examined whether subgroups of sexual mino-
rities exist, with one of the three focusing on gay men, and the other two
focusing on lesbian women. In their study on gay male subgroups, American
researchers Clausell and Fiske (2005) contend that attitudes toward sexual
minorities might be complicated by the existence of subgroups (i.e., a smaller
subset of people with similar identifiers within a larger overarching group),
which the authors first hypothesized to exist after observing gay men’s
neutral position within a map of the Stereotype Content Model (SCM).1 In
order to determine what subgroups of gay men might exist, Clausell and
Fiske (2005) conducted a preliminary study with 44 American undergraduate
students from Princeton University. Participants were instructed to identify
attributes of gay men and then to sort them into subgroups, a process that
resulted in 73 separate subgroup terms being recorded. The researchers
retained subgroups that were generated by at least 10% of the sample,
which were (in descending order): “in the closet,” “flamboyant,” “feminine,”
“crossdresser,” “gay activist,” “hyper-masculine,” “body-conscious,” “artistic,”
“leather/biker,” and “straight-acting.” To understand the perceptions asso-
ciated with each subgroup, Clausell and Fiske (2005) then surveyed an
additional 40 undergraduate students from Princeton University.
Specifically, participants were asked about the warmth and competence of
each subgroup using a measure derived from the SCM, and constructed
originally by Fiske et al. in 2002.2 Using hierarchical cluster analysis,
Clausell and Fiske found that the 10 subgroups that had been generated in
their initial study were clustered into three distinct groups. Results indicated
that, the majority of the groups fell within the high competence-low warmth
(HC-LW; i.e., “body-conscious,” “straight acting,” “in the closet,” “artistic,”
4 J. M. MCCUTCHEON AND M. A. MORRISON

“hyper-masculine,” and “activist”) and low competence-high warmth (LC-

HW; i.e., “flamboyant” and “feminine”) quadrants. In accordance with the
authors’ hypotheses, the “leather/biker” and “crossdresser” subgroups fell
within the most derogated low competence-low warmth (LC-LW) quadrant
area.
One year later, in 2006, Geiger, Harwood, and Hummert conducted
a study to identify subgroups of lesbian women. Unlike research by
Clausell and Fiske (2005) in which the SCM was used, the authors took
a cognitive perspective to theoretically frame their study and asserted that
individuals may hold multiple stereotypes about a given group (e.g., Brewer
et al., 1981; Hummert, 1990). As such, Geiger et al. (2006) expected to find
that individuals possess both positive and negative subgroups of lesbian
women. In the first stage of their two-phase study, Geiger et al. (2006)
instructed 61 students from a mid-western American university (i.e., the
University of Missouri) to generate traits that they associated with the super-
ordinate category “lesbian.” After removing derogatory terms (e.g., “whor-
ebag”) and combining synonymous descriptors (e.g., “manly” with “male-like
”), 94 distinct traits were generated. In the second phase, 63 different under-
graduate students were instructed to sort the traits generated by participants
in Phase 1 into groups that represented “types” of lesbians, and were
instructed to generate names for each lesbian subgroup based on the traits
ascribed. A hierarchical cluster analysis revealed that participants’ categorisa-
tions resulted in two higher-level clusters; namely, positive traits and negative
traits. Four subgroups of lesbian women fell within the positive cluster:
“lipstick lesbian,” “career-oriented feminist,” “soft-butch,” and “free-spirit.”
Likewise, four subgroups comprised the negative traits cluster, which were
“hypersexual,” “sexually confused,” “sexually deviant,” and “angry butch.” In
addition, Geiger et al. (2006) interpreted a strong-weak dimension in which
the subgroups associated with sexuality were perceived as “weak,” and the
“butch” and “feminist” categories were positioned closest to the “strong”
pole.
Complementing the earlier studies of 2005 and 2006 by Clausell and Fiske
and Geiger et al., respectively, Italian researchers Brambilla et al. (2011)
conducted the first study examining the presence of lesbian subgroups out-
side the American states of Missouri and New Jersey. Using the SCM as
a theoretical and methodological framework, the researchers instructed 32
Italian undergraduate students to list subgroups of lesbian women and
provide the most salient characteristics for each subgroup they listed.
Adopting a criterion that the subgroups must be mentioned by at least 15%
or 10% of the sample (a slightly more stringent percentage than that
employed by Causell and Fiske in 2005), the study resulted in the identifica-
tion of four lesbian subgroups; namely, “butch,” “feminine,” “closeted,” and
“outed.” Following the subgroup generation process, Brambilla et al. (2011)
 JOURNAL OF HOMOSEXUALITY 5

assessed 70 Italian students’ ratings of either the four established subgroups

or the superordinate category “lesbian woman” on indicators of warmth and
competence. Of the subgroups generated, “feminine” and “outed” lesbians
fell in the HC-HW quadrant, “butch” lesbians were positioned in the HC-LW
quadrant, and “closeted” lesbians were in the LC-LW area. Importantly, the
results indicated that the superordinate category of “lesbian woman” was
situated in the middle of the warmth and competence SCM space.
Evident across the only three subgroup-generation studies conducted so
far, is the tendency for participants to construct subgroups that relate to
gender conformity and gender non-conformity. For instance, of the ten
subgroups documented by Clausell and Fiske (2005), at least half (i.e.,
“hyper-masculine,” “flamboyant,” “feminine,” “crossdresser,” and “leather/
biker”) reflect perceptions around gender roles. Of the four groups that
clustered in Brambilla et al.’s study, two (i.e., “feminine” and “butch”) relate
to gender role characteristics, as do three (i.e., “lipstick lesbian,” “soft-butch,”
and “angry butch”) of the eight highlighted by Geiger et al. (2006). As
hypothesized by Clausell and Fiske (2005), participants were expected to
rely on gender roles when stereotyping sexual minority persons such that
subgroups that replicated feminine gender roles (i.e., “feminine,” “flamboy-
ant,” and “crossdresser”) were to be rated similarly to the ratings observed for
the general social category “women.” This hypothesis was supported insofar
as both the subgroups perceived as “feminine” and “women” were found in
the LC-HW quadrant. The subgroups that were consistent with masculine
gender roles (i.e., “hyper-masculine,” “straight-acting,” “leather/biker,” and
“in-the-closet”) were expected to fall within the HC-LW space in a pattern
similar to that found for the general social category “men.” The authors’
hypotheses were largely supported on this front as well. Other research (e.g.,
Blashill & Powlishta, 2009; Glick & Fiske, 1996) further provides support for
the relationship between gender role beliefs and sexual prejudice, as well as
the centrality of heterosexuality on gender stereotyping. To conclude,
although the relationships between gender roles and the stereotyping of gay
men and lesbian women cannot account for all of the subgroups that have
been generated, it appears that gender roles are a factor in, and potentially
a driving force behind, heterosexual persons’ evaluations of and attitudes
toward subgroups of gay men and lesbian women. Consequently, a key
objective of the present study is determining the salience of perceived gender
role characteristics for participants and any linkages with the gay and lesbian
subgroups generated.

Present study and incremental advances

The current study will address the following four exploratory research ques-
tions: (1) What contemporary subgroups of gay men and lesbian women are
6 J. M. MCCUTCHEON AND M. A. MORRISON

identified by Canadian-born heterosexual individuals; (2) Will photographs

lead to the identification of additional subgroups over and above what can be
generated without prompts; (3) Are the identified subgroups related to
gender role characteristics; (4) Does level of modern homonegativity affect
the awareness of subgroups that are perceived to exist within a Canadian
context? A cognitive perspective, as used by Geiger et al. (2006), was adopted
for the theoretical framework guiding the present study. This framework
contends that individuals may hold multiple stereotypes about a given group
(Brewer et al., 1981; Hummert, 1990). Different stereotypes may be activated
depending on the characteristics of the situation or target. Based on this
framework, we expect that both positive and negative subgroups would be
generated for the overarching categories of gay men and lesbian women.
However, in the present study, similar to Geiger et al. (2006), the valence (i.e.,
positive or negative) of the descriptions provided for the generated sub-
groups were not collected, and, therefore, drawing conclusions related to
what attitudes and stereotypes are associated with the subgroups are not
possible.
The present study will supplement the small body of research on sub-
groups of gay men and lesbian women. To date, each of the three extant
studies in the area (i.e., Brambilla et al., 2011; Clausell & Fiske, 2005; Geiger
et al., 2006) identified subgroups for one overarching group only; for
instance, Clausell and Fiske (2005) examined subgroups of gay men, and
Geiger et al. (2006) and Brambilla et al. (2011) focused on subgroups of
lesbian women. In the present study, we identify the subgroups generated
empirically for both gay men and lesbian women across two studies, and
include this design feature in order to allow for their initial documentation,
as well as comparison and contrast. Further, since all three of the studies
conducted to date have examined the subgroups of gay men or lesbian
women from the perspective of university students, researchers’ ability to
generalize beyond a student perspective is questionable. In the present study,
we extend our investigation to include non-university participants (for Study
2); thus, documenting for the first time, the subgroups generated by those
outside a post-secondary milieu.3
The current study also will be the first to use a Canadian sample to identify
subgroups of gay men and lesbian women. Canadian laws (Rayside, 2008) and
public opinion surveys (Andersen & Fetner, 2008) indicate that Canadians have
comparatively more liberal attitudes toward sexual minorities than a country
such as the United States. Indeed, Adam (1995) outlines five factors that
differentiate Canadians from Americans in terms of their attitudes toward
sexual minority persons. These are: 1) fewer Canadians identify as evangelical
Protestants; 2) more Canadians are unionized; 3) Canada is multicultural
(rather than a “melting pot”); 4) Canada has a stronger history of social
democracy; and 5) Canada has a weaker militaristic tradition. Adam (1995)
 JOURNAL OF HOMOSEXUALITY 7

suggests that these factors are associated with improved attitudes toward gay
men and lesbian women and have likely influenced governmental and legal
decisions in regards to sexual minority persons. Given these cultural differ-
ences, it is possible that a set of subgroups for gay men and lesbian women that
is distinct from the subgroups found earlier by Clausell and Fiske (gay men;
2005) and Geiger et al. (lesbian women; 2006) will be observed.
In terms of intra-study differences, Study 1 entails participants generating all
possible subgroups of gay men or lesbian women they could think of, and then
providing all possible traits they believed strongly characterized each of the
subgroups generated. Importantly, after initially generating all possible sub-
groups, participants were then shown photographs of men or women to elicit
any additional subgroups that they may have been unable to think of at the time
without a prompt. Based on the results of Study 1, a master list of subgroups
was created for use in Study 2. The current study therefore uses a two-step
approach, which offers the most rigorous methodological approach to generat-
ing subgroups of gay men and lesbian women to date because it reduces the
reliance on memory retrieval. For instance, the subgroups generated in other
studies (i.e., Brambilla et al., 2011; Clausell & Fiske, 2005; Geiger et al., 2006)
were contingent on participants’ memory retrieval abilities at a particular
moment within the course of a study. It is possible that participants may be
unable to think of subgroups with which they are familiar “on the spot” and
consequently provide suboptimal responses. To counteract retrieval biases, we
utilize photographs in Study 1 in order to produce the most exhaustive pool of
possible subgroups for use in Study 2. According to Deocampo and Hudson
(2010), using photographs to prompt responses is a valid method for increasing
participant recall. It should also be mentioned that it is most often strangers
who display aggression toward sexual minority persons (Mason, 1993). Since
strangers have only a person’s appearance or limited behavioural cues to rely on
when classifying a person as a sexual minority or member of a particular sexual
minority subgroup, incorporating a photograph of a presumed sexual minority
person was considered essential in order to provide participants with similar
circumstances for subgroup classification.

Study 1
The purpose of Study 1 was to create a master list of potential subgroups of
gay men and lesbian women that can assist in the identification of subgroups
that are salient to heterosexual Canadians.

Participants
In total, 67 undergraduate students were recruited from the University
of Saskatchewan psychology participant pool and received bonus course
8 J. M. MCCUTCHEON AND M. A. MORRISON

credit for their participation. Participants were 47 (70.1%) women and

20 (29.9%) men. They ranged in age from 18 to 42 (M = 21.04;
SD = 4.53) and the majority self-identified as Caucasian (n = 50;
74.6%) and heterosexual (n = 59; 88.1%). Participants were randomly
assigned to complete the study about either gay men (n = 33) or lesbian
women (n = 34).

Procedure and materials

This study was conducted on the University of Saskatchewan campus.
Students signed up to participate via the psychology participant pool website
and received bonus course credit. On their arrival to the researcher’s labora-
tory, participants were given an informed consent sheet to read and sign. It
was explained to students that their participation was voluntary and that the
purpose of the study was to establish perceived subgroups of gay men and
lesbian women. In Phase I, participants were presented with paper-and-
pencil Subgroup Generation Forms, which instructed participants to “list
any and all subgroups of gay men that you think exist in society.” It was
noted that “these subgroups do not need to reflect [their] own opinion, but
rather should reflect societal stereotypes or associations with which [they] are
familiar.” They were also asked to list five adjectives they felt were most
associated with each of the subgroups generated. Participants were asked to
generate subgroups that they think exist from a society’s perspective to
reduce socially desirable responding (Fiske, Cuddy, Glick, & Xu, 2002).
This tactic is also used in other subgroup generation studies (Clausell &
Fiske, 2005; Geiger et al., 2006). Participants filled out one Subgroup
Generation Form per subgroup that they listed. Participants were randomly
assigned using a random number generator to complete the study for either
gay men or lesbian women.
After completing the Subgroup Generation Forms, participants moved to
a computer terminal for Phase II. At the computer terminal, participants
were presented with 50 photographs of men (if they generated subgroups of
gay men) or 50 photographs of women (if they generated subgroups of
lesbian women) and were asked to categorise them on the basis of the
subgroups they had generated by typing the subgroup name into an open-
ended text box under the photographs. Participants were provided with their
Subgroup Generation Forms to refer back to when completing the compu-
terised task. The researchers selected the 50 photographs of men and 50
photographs of women using Google Image Search.4 It was believed that the
presentation of photographs might trigger participants to think of additional
subgroups that they were unable to without assistance. Participants were
given the opportunity to propose additional subgroups in Phase II if the
photographs prompted them to think of others that were not generated in
 JOURNAL OF HOMOSEXUALITY 9

Phase I. If additional subgroups were proposed, participants completed

another Subgroup Generation Form for each of the new subgroups.
The presentation of visual stimuli would be similar to participants encoun-
tering a stranger they believe to be a sexual minority person. In these
situations, they would classify the stranger using visual appearance cues.
Past research (i.e., Brambilla et al., 2011; Clausell & Fiske, 2005; Geiger
et al., 2006) on subgroups of gay men and lesbian women were used to
guide the selection of photographs. That is, the subgroup labels (e.g., “flam-
boyant” and “butch”) were entered into the search engine, as well as the
general search terms “gay,” “lesbian,” and “homosexual,” and different com-
binations of the subgroup labels and general terms. Colour photographs
showing men and women from the area above the waist, groin, or knee
and facing forward were selected.5 Photographs depicting celebrities, fashion
models, or those containing multiple people or people who were in a state of
undress were disqualified. The photographs were cropped so that only the
area above the waist, groin, or knee was visible, and the backgrounds were
edited to be uniformly white. Some search terms (e.g., “leather/biker”)
generated many qualifying photographs and, in those cases, the photographs
that were considered to best fit the criteria (i.e., facing forward, depicting
only the waist, groin of knee-up area) and that required the least additional
editing (i.e., cropping, changing the background colour to white) were
selected.
As the final task, participants responded to four demographic questions
(i.e., age, gender, ethnicity, and sexual orientation) on the computer
terminal. Participants were then debriefed by being provided information
about the study’s purpose (i.e., establishing possible subgroups of gay men
and lesbian women). The study took approximately 30 minutes to
complete.

Analysis
To ascertain the subgroups of gay men and lesbian women, two coders
independently classified the generated subgroups into categories. The coders
were provided with the following instructions to guide their coding:

● Step 1: Group identically named subgroups together (e.g., butch, butch)

—while ensuring that associated adjectives do not imply distinct sub-
groups (e.g., Queen: elegant, royal, graceful, reserved, poised vs. Queen:
sassy, flamboyant, effeminate, eccentric, loud).
● Step 2: Group synonymously named subgroups together (e.g., manly,
masculine)—while ensuring that the associated adjectives do not imply
distinct categories. Select the subgroup name, of the synonymous terms,
that best exemplifies the category.
10 J. M. MCCUTCHEON AND M. A. MORRISON

● Step 3: Examine associated adjectives for overlap between subgroups

with different names (e.g., one subgroup might be called “Feminine,”
while another group might be called “Lipstick” but may have “Feminine”
as an associated adjective)—determine if these subgroups should be
classified together and choose the subgroup name that best exemplifies
the category.
● Step 4: Examine remaining subgroups for conceptual overlap. Use your
discretion to determine if subgroups should be categorised together.
Again, select the subgroup name that best exemplifies the composite
category.
● Step 5: Ensure that subgroups that have the same participant number are
not classified in the same category.
● Note: every subgroup should be classified into a category (either with
other subgroups or by itself)—all categories should be given a name.

These steps were completed to create a manageable list of conceptually

distinct subgroups for use in the subsequent study. The approach used in the
current study was similar to that used by Geiger et al. (2006) in their
identification of traits describing lesbian subgroups. Following the indepen-
dent coding, the coders compared their results. Any differences in opinion
between the two coders were resolved through discussion. The inter-rater
reliability for identically-labelled subgroups was 47.3%. In most cases, the
subgroups were categorised similarly but were given different names (e.g.,
drag queen vs. crossdresser, hippie vs. free-spirited). When these similarly
labeled names were considered as a match between coders, the inter-rater
reliability increased to 82.1%. To resolve these conflicts, the subgroups were
labelled with the word used most frequently by participants.

Results and discussion

Participants who completed the study in relation to subgroups of gay men
generated between 2 and 10 subgroups each. A total of 184 subgroups were
generated across all participants. The mean number of subgroups generated
was 5.58 (SD = 2.25), with a mean of 3.88 (SD = 1.83) subgroups generated
prior to the presentation of photographs, and a mean of 1.70 (SD = 1.38)
subgroups generated afterward. The coding process revealed that 51 distinct
subgroups were generated. The top five generated subgroups included: flam-
boyant (n = 30; 90.9%), drag queen (n = 15; 45.5%), masculine (n = 13;
39.4%), feminine (n = 11; 33.3%), and normal (n = 7; 21.2%).
Participants who completed the lesbian version of the study generated
between 2 and 9 subgroups, with 182 subgroups being generated in total. On
average, participants generated 5.35 (SD = 1.94) subgroups, with a mean of
3.71 (SD = 1.61) subgroups being generated before the presentation of
 JOURNAL OF HOMOSEXUALITY 11

photographs, and a mean of 1.65 (SD = 1.41) subgroups being generated after
the photographs were presented. The independent coders determined that
participants generated 38 distinct subgroups of lesbian women. The top
generated lesbian subgroups included: butch (n = 26; 76.5%), feminine (n
= 14; 42.4%), free-spirited (n = 14; 42.4%), tomboy (n = 11; 32.4%), feminist
(n = 8; 23.5%), and promiscuous (n = 8; 23.5%).
The results of Study 1 support the notion that individuals perceive sub-
groups of gay men and lesbian women beyond these overarching categories.
On average, participants perceived approximately five to six subgroups for
each sexual minority group. The inclusion of photographs, a tactic that had
not been employed in existing sexual minority subgroup studies, to trigger
the generation of other subgroups was helpful, with an average of 1.65 and
1.70 additional lesbian and gay subgroups, respectively, being cited after the
presentation of photographs.

Study 2
The purpose of Study 2 was to identify subgroups of gay men and lesbian
women that are salient to heterosexual Canadian undergraduate students and
members of the general public.

Participants
Data from two groups of participants were collected and analyzed sepa-
rately. Firstly, 106 students (80 women, 26 men) were recruited through
the introductory psychology participant pool and a general university-wide
student participant pool. They ranged in age from 17 to 50, with a mean
age of 21.79 (SD = 6.13). The majority self-identified as White (81.1%, n
= 86), with 4.7% (n = 5) identifying as Aboriginal and East Asian, 3.8% (n
= 4) as South Asian, 2.8% (n = 3) as mixed origin and .9% (n = 1) as Black
and Southeast Asian.
A general Canadian population sample also was recruited to verify the
generalizability of the subgroups identified by the student sample. In total,
100 heterosexual, Canadian adults (52 women, 48 men) were recruited from
the sample vendor, Ekos6. The general population sample ranged in age from
24 to 79, with a mean age of 31.64 (SD = 15.04). Again, the majority self-
identified as White (87.0%, n = 87), with 5.0% (n = 5) identifying as
Aboriginal, 3.0% (n = 3) as East Asian and mixed origin, and 1.0% (n = 1)
as Black and other. Half of the participants responded from Ontario (50.0%;
n = 50), 16.0% (n = 16) from British Columbia, 10.0% (n = 10) from Alberta,
8.0% (n = 8) from Saskatchewan, 4.0% (n = 4) from Manitoba, 3.0% (n = 3)
each from New Brunswick and Nova Scotia, and 2.0% (n = 2) from Quebec
and Newfoundland.
12 J. M. MCCUTCHEON AND M. A. MORRISON

Participants were randomly assigned to complete the study about gay men
(n = 55 students; n = 49 general population) or lesbian women (n = 51
students; n = 51 general population). Participants recruited through the
psychology participant pool were awarded course credit; those recruited
through the general university pool received five dollars; and Ekos panel
members voluntarily participate in research with no remuneration. Only
participants identifying as heterosexual were included in the analysis. While
individuals of any sexual orientation could participate in Study 1, the goal of
Study 2 was to establish subgroups that were salient for heterosexual indivi-
duals; thus, it was the perspective of heterosexually-identifying persons that
were documented and analyzed. It also should be mentioned that only
Canadian-born7 participants were included in the analysis as it was critical
that the subgroups be salient in a Canadian context.

Materials
Master list of subgroups
To create a large and comprehensive list of possible subgroups, the master
list combined the subgroups generated in Study 1 as well as subgroups
generated by the first author and five experts in the field of gender and
sexuality who were invited to contribute. Similarly to the instructions from
Study 1, these individuals were directed to generate any and all subgroups of
gay men and lesbian women that are perceived by members of society to
exist. They were informed that the subgroups do not need to reflect their
own opinions, but rather should reflect subgroups with which they are
familiar. For the gay men master list, 51 subgroups were taken from Study
1 and 46 additional subgroups generated by the first author and the content
experts, to yield a total of 97 subgroups. For the lesbian master list, 31
additional subgroups were generated to add to the 38 subgroups from
Study 1, for a total of 69 subgroups. The subgroups were listed alphabetically.

Homonegativity
The Modern Homonegativity Scale (MHS; Morrison & Morrison, 2003) was
used to assess participants’ homonegativity. The MHS is a 12-item scale that
measures modern or covert negative attitudes toward gay men and lesbian
women (e.g., “Many lesbians [gay men] use their sexual orientation so that
they can obtain special privileges”). Two parallel versions exist, with one
measuring attitudes toward gay men (MHS-G) and the other measuring
attitudes toward lesbian women (MHS-L). Participants completed the version
that corresponded to their assigned condition (i.e., if they were selecting
subgroups of gay men they completed the MHS-G, and if they were selecting
subgroups of lesbian women they completed the MHS-L). A 5-point Likert
scale was used (1 = strongly disagree; 5 = strongly agree) and total scores
 JOURNAL OF HOMOSEXUALITY 13

could range from 12 to 60. Higher scores represent greater endorsement of

modern homonegative attitudes. The MHS has been identified as
a psychometrically sound measure (Morrison, Kenny, & Harrington, 2005;
Morrison & Morrison, 2003; Morrison, Morrison, & Franklin, 2009). In the
present study, Cronbach’s alpha for the MHS-L was .85 (CI = .78-.91) and .95
(CI = .92-.97) for the student and general population sample, respectively.
Cronbach’s alpha was .83 (CI = .76-.89) and .94 (CI = .91-.96) for the MHS-G
for the student and general population sample, respectively.

Procedure
Data collection with students was conducted in-person, while the general
population sample completed the study entirely online. In-person partici-
pants were directed to a computer station to complete the study. All parti-
cipants were presented with an informed consent sheet, which explained
their rights as participants and indicated that the purpose of the study was
to understand perceived subgroups of gay men and lesbian women. After
providing their consent, participants were presented with the gay men or
lesbian women version of the master list. Participants completed only one
version to ensure that the selection of subgroups of one sexual orientation
group would not influence the other. Participants were instructed to check
off any subgroup with which they believed represented subgroups perceived
within society. These instructions were provided to reduce socially desirable
responding (i.e., participants not selecting certain subgroups because they do
not want to appear to personally endorse them). The intention was to
identify subgroups that are perceived to exist within their social world.
After selecting the subgroups, they were then prompted to provide
a definition of the subgroups that they selected. This component was incor-
porated into the methodology as a check to ensure that people truly under-
stood what the subgroup was that they were selecting. Finally, participants
responded to the MHS-G or MHS-L (corresponding to the condition they
were in for the master list) and six demographic questions (i.e., age, gender,
ethnicity, sexual orientation, country of birth, native tongue).

Results
In the gay men condition, student participants selected between 3 and 68
subgroups, with a mean number of 14.11 (SD = 12.74) subgroups selected.
The general population participants selected between 1 and 31 subgroups
(M = 7.67; SD = 7.66) for the gay men condition. In the lesbian condition, the
student sample selected between 2 and 32 subgroups, with a mean selection
of 10.59 (SD = 7.85) subgroups, while the general population sample chose
between 1 and 24 subgroups (M = 5.88; SD = 4.45). T-tests revealed that the
student sample selected significantly more gay, t(102) = 3.08, p = .003,
14 J. M. MCCUTCHEON AND M. A. MORRISON

d = 0.612, and lesbian subgroups, t(80) = 3.78, p < .001, d = 0.748, than the
general population sample.
Participants’ definitions were reviewed for accuracy by three independent
coders using a list of definitions created prior to their independent reviews.
The coders discussed and agreed upon the definitions for the subgroups
based on participants responses in Study 1 and through consultation with
sexual minority community members and content experts. Upon an exam-
ination of participants’ definitions, the coders determined that in the gay
men condition, Club Kid and Partier should be collapsed into one category,
as well as Drama Queen, Queen, and Flamboyant; Feminine and Femme; and
Masculine and Macho. Participants’ definitions for these categories were
nearly identical, while still being perceived as correctly defining the con-
structs. In the lesbian women condition, the coders recommended that
Activist and Feminist be merged into a single category; as well as Alpha
and Dominant; the three categories of Bulldyke, Dyke and Butch; and finally,
Feminine, Femme, and Lipstick. These categories were merged for all subse-
quent analyses and the label that was selected most commonly by partici-
pants was used. The coders categorized the definitions as either correct or
incorrect. An interrater reliability analysis using the Fleiss’ Kappa statistic
(i.e., used to assess inter-rater reliability between three or more raters; Fleiss,
Nee, & Landis, 1979) was performed. The analysis revealed a Fleiss’ Kappa of
.61 for the coding of the gay men subgroup definitions and .67 for the coding
of the lesbian women subgroups definitions provided by the student sample,
and .74 and .70 for the general population sample, respectively. These scores
denote substantial agreement among the three coders (Landis & Koch, 1977).
The review revealed that approximately 78.1% of definitions in the gay
condition were correct, and 78.9% in the lesbian condition were correct for
the student sample. In the general population sample, 69.1% of definitions in
the gay condition and 68.2% in the lesbian condition were deemed accurate,
respectively. When considering the number of correct definitions by partici-
pant, the student sample achieved an average of 78.7% definitions correct for
the gay condition and 78.6% correct for the lesbian condition. The general
population group achieved 76.2% and 74.7% correct, respectively.
Independent t-test analyses revealed that these scores did not significantly
differ from one another for either the gay condition, t(78) = 0.531, p = .597,
or the lesbian condition, t(89) = 0.957, p = .341.
A criterion of at least 50% was chosen a priori as the threshold by which
the sample was required to select and accurately define a subgroup in order
for it to be considered salient. Brambilla et al. (2011) and Fiske et al. (2002)
retained subgroups generated by at least 15% of the sample; however, this
study uses a recognition approach (i.e., selecting subgroups from a pre-
existing list), as opposed to subgroup generation (i.e., generating the sub-
groups spontaneously on one’s own). As such, the more stringent 50% cut-
 JOURNAL OF HOMOSEXUALITY 15

off used in the present study was deemed appropriate because it would
ensure that the subgroups are salient to many people. Among the student
sample, for the gay men group, four subgroups exceeded the 50% criterion;
specifically, Flamboyant (identified as a subgroup and accurately defined by
65.5% [n = 36] of participants), Drag Queen (56.4%; n = 31), Feminine
(56.4%; n = 31), and Closeted (50.9%; n = 28). Among the general population
sample, Drag Queen (55.1%; n = 27) and Flamboyant (51.0%; n = 25)
surpassed the 50% cut-off. While not attaining the 50% criterion, Feminine
(28.6%; n = 14) and Closeted (24.5%; n = 12) were the next highest recog-
nized and accurately defined subgroups among the general population sam-
ple, along with the Bear subgroup (24.5%; n = 12).
For the lesbian women group, three subgroups exceeded the criterion
among the student sample; specifically, Feminist (72.5%; n = 37), Butch
(68.6%; n = 35), and Tomboy (56.9%; n = 29). Among the general population
sample, only Butch (52.9%; n = 27) qualified as a salient subgroup; the next
highest rated subgroups from the general population, however, mapped onto
the student responses of Feminist (31.4%; n = 16) and Tomboy (27.5%; n
= 14). The Closeted subgroup (25.5%; n = 13) also was rated highly among
the general population sample as compared to other lesbian women sub-
groups. Aside from the subgroups that were identified as meeting the criter-
ion (i.e., 50% or more), all other selected subgroups fell below 33%
endorsement, suggesting that the final subgroups were particularly salient
for respondents. See Tables 1 and 2 for a list of the number of respondents
who selected and correctly defined each of the subgroups.
Student participants in the gay male subgroup condition had a mean MHS-
G score of 24.3 (SD = 7.09) and a range of 13 to 43, while the student
participants in the lesbian women subgroup condition had a mean MHS-L
score of 26.4 (SD = 7.52) and a range of 13 to 44. Among the general
population sample, participants’ scores on the MHS-G ranged from 13 to 60
and had a mean of 31.9 (SD = 12.80). On the MHS-L, scores ranged from 12 to
59 and had a mean of 31.7 (SD = 12.71). To determine if scores on the MHS
differed by sample or sexual orientation condition, a 2 (sample: student or
general population) x 2 (orientation: gay or lesbian) ANOVA was conducted.
Prior to conducting the analysis, normality was assessed using the
Kolmogorov-Smirnov and Shaprio-Wilk tests. These tests were found to be
significant (p < .001). Upon visual inspection, it was evident that the data
were positively skewed. Additionally, to assess the assumption of homoge-
neity of variance, Levene’s test was examined and found to be significant (p
< .001). To address these violations, the data were subjected to an inverse
transformation. Following the transformation, Kolmogorov-Smirnov and
Shaprio-Wilk tests were still significant (p < .001); however, Levene’s test
was no longer significant (p = .393). Given that it has been argued that
ANOVA is robust to violations of normality when the sample size is large
 16

Table 1. Frequency of recognized gay men subgroups that were accurately defined after selection.
Student Sample (N = 55) General Population Sample (N = 49)
Frequency of subgroup Frequency of subgroup
being correctly defined being correctly defined
after selection after selection
Subgroups n (%) Subgroups n (%)
Flamboyant 36 (65%) Drag Queen 27 (55%)
Drag Queen, Feminine 31 (56%) Flamboyant 25 (51%)
Closeted 28 (51%) Feminine 14 (29%)
Proud 18 (33%) Closeted, Bear 12 (24%)
Bisexual, Different 17 (31%) Bisexual 8 (16%)
Homo 14 (25%) Straight acting, Twink 7 (14%)
Activist, Artistic 13 (24%) Homo 6 (12%)
Funny 12 (22%) Bottom, Club kid, Gym rat, Leatherman, Masculine 5 (10%)
Actor, Bitchy 11 (20%) Activist, Actor, Artistic, Cruiser, Different, Theatric 4 (8%)
Attractive, Confident, Masculine 10 (18%) Amateur porn producers/performers, Dominant, Friend, Normal, Outcast, 3 (6%)
J. M. MCCUTCHEON AND M. A. MORRISON

Power bottom, Promiscuous, Proud, Rice queen, Stylish, Top

Bottom, Club kid, Outcast 9 (16%) Arts student, Bitchy, Chubby chaser, Daddy, Damaged, Diseased, Ethnic 2 (4%)
fetishizer, Gaymer, Neat freak, Pedophile
Arts student, Bear, Friend, Preppy, Straight 8 (15%) Aboriginal, Attractive, Black, Bug chaser, Cub, Domestic, Excludes ethnic 1 (2%)
acting, Top minorities, Funny, Gift giver, Homophobic, Jailbait, Nerdy, Otter, Preppy,
Professional, Smoker, Submissive, Typical, White
Happy, Queen, Theatric 7 (13%)
Gym rat, Hipster, Jock, Latino, Leatherman, 6 (11%)
Promiscuous, Twink, Weak, White
Daddy, Helpful, Neat freak, Nerdy, Normal, 5 (9%)
Professional, Submissive
Amateur porn producers/performers, American, 4 (7%)
Asian, Black, Dominant, Teenage
Body modifier, Cub, Damaged, Father, 3 (5%)
Homewrecker, Pedophile, Polygamist, Power
bottom, Typical
Aboriginal, Bro, Chubby chaser, Cruiser, 3 (4%)
Domestic, Gaymer, Goth, Homophobic,
Jailbail, Masochist, Otter, Thug, Twunk
Table 2. Frequency of recognized lesbian subgroups that were accurately defined after selection.
Student Sample (N = 51) General Population Sample (N = 51)
Frequency of subgroup being Frequency of subgroup
correctly defined after being correctly defined after
selection selection
Subgroups n (%) Subgroups n (%)
Feminist 37 (73%) Butch 27 (53%)
Butch 35 (71%) Feminist 16 (31%)
Tomboy 29 (57%) Tomboy 14 (27%)
Queer 16 (31%) Closeted 13 (25%)
Androgynous 14 (27%) Open 10 (20%)
Free spirited, Transgender 13 (25%) Free spirited 9 (18%)
Alpha 12 (24%) Dyke 8 (16%)
Alternative, Closeted, Hipster 11 (22%) Androgynous 7 (14%)
Arts student, Feminine 10 (20%) Feminine, Queer 6 (12%)
Drag king, Drama queen, Flamboyant, Lezboi, Punk 9 (18%) Baby dyke, Straight acting 5 (10%)
Athlete, Biker 8 (16%) Athlete, Aboriginal 4 (8%)
Dominant, Goth 7 (14%) Alpha, Alternative, Arab/Middle Eastern, Asian, 3 (6%)
Black, Drag king, Flamboyant, Professional, Pushy,
Uncertain
Ambiguous, Body modifier, Open, Uncertain 6 (12%) Attention seeker, Dominant, Hidden, Lezboi, 2 (4%)
Mother, South Asian, White
Attention seeker, Musician 5 (10%) Biker, Body image, Latino, Model, Punk, Submissive, 1 (2%)
Transgender
Intellectual, Promiscuous, Straight acting, White 4 (8%)
Asian, Baby dyke, Black, Confidant, Mother, Pushy, Smoker, 3 (6%)
Troublemaker
Caregiver, Reserved, South Asian, Substance user 2 (4%)
Aboriginal, Arab/Middle Eastern, Body image, Bottom, 1 (2%)
JOURNAL OF HOMOSEXUALITY

Chapstick, Fashion, Hidden, Jailbait, Masochist, Prisoner,

Submissive, Top
17
18 J. M. MCCUTCHEON AND M. A. MORRISON

(Maxwell & Delaney, 2004), the analysis was conducted using the trans-
formed data despite significant tests of normality. The analysis revealed
a significant main effect for Sample, F(1, 202) = 8.14, p = .005, η2 = .04,
with the general population respondents (M = 31.80, SD = 12.69) having
significantly higher MHS scores than the student sample (M = 25.28,
SD = 7.34).
Finally, seven t-tests were conducted for the student sample and three for
the general population sample to determine if the selection of subgroups was
contingent on level of modern homonegativity. Yes/no scores (0 = no;
1 = yes) as to whether the term was considered a subgroup was the inde-
pendent variable and scores on the MHS was the dependent variable in the
analyses. The results were non-significant (p > .05) for all analyses, suggesting
that participants recognised the existence of these highly selected subgroups
regardless of their level of homonegativity.

Discussion
The present study offers evidence for the existence of perceived subgroups
within the overarching social categories of “gay men” and “lesbian women.”
Based on the subgroup generation task in Study 1, in which participants
generated an average of 3.88 and 3.71 subgroups without any prompting, and
an additional 1.70 and 1.65 subgroups after photographs of men or women
were presented, it can be inferred that Canadian heterosexual students
acknowledge the perceived existence of gay and lesbian subgroups within
society. Using the subgroups created in Study 1, Study 2 participants, which
included an undergraduate student sample and a national general population
sample identified those that they believe exist at the societal level. The use of
both a “spontaneous generation” task and a “recognition” task to identify
salient subgroups was an important departure from the existing subgroup
generation studies (i.e., Brambilla et al., 2011; Clausell & Fiske, 2005; Geiger
et al., 2006). These studies required participants to generate their own
subgroups using recall. Cognitive theorists (e.g., Craik & McDowd, 1987;
Yonelinas & Jacoby, 1994) suggest that generation tasks are more difficult
than recognition tasks because the former requires the initial recall of
information. With the recognition task, participants were able to select
from a list of numerous potential subgroups. As such, the subgroups that
emerged as most salient were not limited by participants’ recall ability.
Another departure from past studies was the selection of a more stringent
cut-off criterion for the identification of “salient” subgroups. This cut-off
criterion refers to the percentage of participants needed to select the sub-
group for it to qualify as a viable one. Clausell and Fiske (2005) used a 10%
criterion and Brambilla et al. (2011) employed a 15% cut-off. A much larger
criterion of 50% was employed in the present study to ensure that the
 JOURNAL OF HOMOSEXUALITY 19

subgroups would be salient enough to individuals to be generalizable beyond

the present study. The sizeable gap (upwards of 18 percentage points)
between the subgroups exceeding the 50% cut-off and those that fell below
supports the selection criterion that was chosen and reinforces the likelihood
that the selected subgroups are salient to Canadians.
For the gay men condition, the student sample generated four subgroups
and the non-student sample generated two. In descending order, the sub-
groups exceeding the 50% criterion were: Flamboyant, Feminine, Drag Queen,
and Closeted for the students, and Drag Queen and Flamboyant for the non-
students. In the lesbian women condition, student participants generated three
subgroups that met the criterion and the non-student sample generated one. In
descending order, the subgroups for the student sample were Feminist, Butch,
and Tomboy. Only Butch was selected by the non-student sample. All four of
the gay men subgroups were among the 10 that emerged in Clausell and
Fiske’s (2005) study. When comparing the emergent lesbian subgroups to
those reported by Geiger et al. (2006), they were captured if you consider
Tomboy to be similar to their “Soft Butch” category and Butch as equivalent to
their “Angry Butch” subgroup. Had a lower criterion been used in the present
study, all the subgroups generated from previous research would have
emerged, with the exception of the sexuality-focused subgroups (i.e., hyper-
sexual, sexually confused, and sexually deviant), which appeared not to be
salient subgroups of lesbian women for either sample.
It is worth noting that had a criterion lower than 50% been selected for the
general population sample, the subgroups would have mapped identically on
to those generated by the student sample. That is, the top four gay men
subgroups and the top three lesbian subgroups were identical for the two
samples. This provides strong support for the saliency of these subgroups. It
is possible that a student sample may have a more varied understanding of
sexual minorities and that explains why additional subgroups emerged as
salient among this sample. Researchers (e.g., Herek & Capitanio, 1996; Herek
& Glunt, 1993) have found that younger and more educated individuals are
more likely to report greater contact with sexual minorities. As such, they
may have been exposed to a greater diversity of gay and lesbian persons.
Future research, however, is needed to determine why the subgroups that
were identified as perceived subgroups emerged over others. In particular,
a qualitative inquiry is needed that explores how people perceive subgroups
of sexual minorities, how these perceptions are developed, and what judg-
ments are used to categorise sexual minorities.
There are a few limitations related to the subgroups that emerged as salient
among participants that are worth noting. For instance, it should be acknowl-
edged that the master list was presented to participants alphabetically. As
such, it is possible that alphabetically earlier subgroups may have been
selected more often, while alphabetically later subgroups may have been
20 J. M. MCCUTCHEON AND M. A. MORRISON

selected less often and been affected by fatigue due to the high number of
subgroups. Additionally, participants may have perceived some of the sub-
groups on the master list similarly (e.g., Top and Dominant). If some sub-
groups were perceived to cluster together and participants opted to select
only one from the cluster, a lower endorsement level for some subgroups
may have resulted. Researchers should consider replicating Study 2 by ran-
domizing the master list and by either including a follow-up question related
to potential subgroup clusters or including explicit instructions to partici-
pants to select all subgroups they perceive to exist even if they seem redun-
dant. Lastly, all subgroup generation studies are limited by the historical,
geographical, and cultural context in which the studies are conducted. For
this study in particular, given that the samples were predominantly White
and the photographs used to increase the comprehensiveness of the subgroup
master list were White, it is possible that the subgroups that emerged
represent a racially homogenous subsection of gay and lesbian culture.
There are a number of inferences that can be made as a result of the
subgroups that emerged. Perceived gender role characteristics appear to be
an important determinant of gay and lesbian subgroups. Three of the gay
subgroups (i.e., Drag Queen, Feminine, and Flamboyant) and two of the
lesbian subgroups (i.e., Butch and Tomboy) relate directly to violations of
socially assigned gender roles. That is, drag queens, feminine, and flamboyant
gay men are ascribed feminine characteristics; butch and tomboy lesbian
women are considered to have masculine traits. It is only the Feminist lesbian
woman subgroup and the Closeted gay man subgroup that were not ascribed
gender role traits or appearance indicators. Arguably though, the behavioral
attributes (e.g., strong, independent, self-sufficient, aggressive) that were used
to describe Feminist could be categorized as masculine traits (Bem, 1974).
Also, Clausell and Fiske (2005) noted that their “in the closet” subgroup was
rated similarly to the other masculine/agentic subgroups and, like the general
category of “men,” fell within the HC-LW quadrant of the SCM.
Interestingly, a subgroup that would be considered “feminine” did not
emerge among lesbian women using the recognition approach in Study 2.
Few researchers have addressed the invisibility of feminine lesbian women.
An exception is Walker (1993), who provides an in-depth commentary on
the invisibility of “femme” lesbian women. She notes that butch lesbian
women are often represented as “the authentic lesbian” (p. 881), while
femme lesbian women are able to “pass” as heterosexual and are, therefore,
perceived as not experiencing the same extent of marginalisation and even
not truly desiring other women. Ciasullo (2001) approaches the topic of (in)
visibility from a different perspective by arguing that it is actually butch
lesbian women who are invisible, particularly within mainstream media, but
that feminine lesbian women are represented as not true lesbians. Ciasullo
notes that “the mainstream lesbian body is … . made into an object of desire
 JOURNAL OF HOMOSEXUALITY 21

for straight audiences through her heterosexualization, a process achieved by

representing the lesbian as embodying a hegemonic femininity” (p. 578). She
also argues that lesbian women are rarely portrayed in sexual acts with
women and as such are also “de-homosexualized.” While adopting
a different approach, Ciasullo’s (2001) and Walker’s (1993) perspectives are
complementary and support the failure of a feminine subgroup to emerge in
the present study. It is possible that heterosexual individuals, while inundated
with depictions of feminine lesbian women in mainstream media, may
dismiss feminine lesbian women as “real” lesbians or as a group that would
not be populous enough to warrant a subgroup designation. Moreover, the
omission of an explicit Masculine gay man subgroup also supports the
tendency for individuals to stereotype gay men as feminine and lesbian
women as masculine (Blashill & Powlishta, 2009).
Of the subgroups that did emerge as salient for the participants, none were
related to their scores on the Modern Homonegativity Scale (Morrison &
Morrison, 2003). Regardless of participants’ own negativity (or positivity)
towards gay men and lesbian women, they recognize the same subgroups as
being perceived to exist by society. This finding may suggest that certain
stereotypes about gay men and lesbian women are pervasive and widely
known despite participants not personally endorsing them or believing they
exist. Future research should determine if there are other factors that may
relate to whether individuals believe that the identified subgroups are per-
ceived to exist by society. Given that no subgroup was unanimously selected,
it is possible that there are variables that may be associated with participants’
recognition of certain gay and lesbian subgroups. The identification of these
factors could help in developing interventions to dispel stereotypes and
misconceptions about sexual minority persons.
As another important step in understanding the subgroups that emerged
as salient in the present study, researchers should assess the content and
valence of the attitudes that people have towards the subgroups. While the
current study used rigorous methods to identify subgroups, it did not assess
how they are perceived. Research from Clausell and Fiske (2005), Brambilla
et al. (2011), and Geiger et al. (2006) suggest that all of the subgroups that
emerged in the present study could be ranked along a hierarchy from positive
to negative. For instance, in Clausell and Fiske’s study of gay men, “in the
closet” gay men were rated most positively, while “crossdressers” were rated
most negatively. In Geiger at al.’s study of lesbian women subgroups, “career
oriented feminists” were rated positively, while “angry butch” was rated
negatively. Based on the evidence from these preliminary subgroup studies,
it would be expected that the subgroups that are perceived as “gender role
violators” would be relegated to the lowest tier of the subgroup hierarchy.
Important, too, is for researchers to evaluate how the subgroups are rated
in comparison to the overarching social categories of “gay man” and “lesbian
22 J. M. MCCUTCHEON AND M. A. MORRISON

woman,” to determine their degree of difference or alignment. Given that

these are the categories that are primarily assessed in research examining
attitudes towards sexual minority persons, it would be valuable to determine
if some of the subgroups map onto the attitudes toward the overarching
category. This would assist researchers to determine if individuals are more
likely to think of one subgroup over another when asked to evaluate the
superordinate group. An existing framework, such as the SCM (Fiske et al.,
2002), could be used to guide such an investigation, both to identify the
relative attitudes toward each group and to assess if such a theoretical and
methodological approach would be valid for the generated subgroups. Past
research using the SCM (i.e., Brambilla et al., 2011; Clausell & Fiske, 2005)
suggests that perceptions of gay and lesbian subgroups result in their posi-
tioning across the SCM quadrants; however, the framework’s validity asses-
sing attitudes towards gay and lesbian subgroups within a Canadian context
and with a general population sample has yet to be evaluated. Alternative
theories and frameworks could potentially spawn from an increased under-
standing of attitudes towards the subgroups of gay men and lesbian women,
particularly when considering the different results that emerged across the
two samples (i.e., students vs. general population).
As can be seen from conducting the same study with two separate popula-
tions, an identical number of subgroups are not necessarily going to emerge
across difference samples. Although the seven subgroups that emerged for
the student sample were all recognized by many of the participants in the
general population sample, only three of the seven reached the 50% thresh-
old. Moving forward, researchers should examine variables such as familiar-
ity or contact with sexual minorities, education levels, geographic area, and
age to determine if these differences may result in the recognition of addi-
tional gay and lesbian subgroups. For instance, urban centers such as
Toronto or Vancouver, where there is greater social diversity, may potentially
yield additional subgroups from a student sample than the present study,
which was conducted in a medium-sized urban centre. Despite these poten-
tial limitations in generalizability, the advance of the sexual minority sub-
group generation research into a non-student sample is an important step in
understanding how gay men and lesbian women are perceived. The present
study is indeed the first to identify subgroups of gay men and lesbian women
from a national sample. The findings from the two samples largely supported
one another as well as many of the subgroups in the extant literature.

Conclusion
As the first Canadian study to identify subgroups of gay men and lesbian
women that are salient among both a student sample and a general popula-
tion sample, the present study provides a good launching point to begin
 JOURNAL OF HOMOSEXUALITY 23

Canadian research that goes beyond the superordinate categories of gay men
and lesbian women. These overarching categories are possibly masking
a different reality about the attitudes that exist toward sexual minority
persons. If attitudes differ based on the perceived subgroup that a sexual
minority person is classified into then it is important that these attitudes are
being accurately assessed, with the ultimate goal to develop interventions that
could reduce prejudice and discrimination toward gay men and lesbian
women. Moving forward, researchers should consider using the subgroups
that have been identified in the present study to assess attitudes toward
sexual minority persons and should conduct research that would lead to
a deeper understanding as to how these subgroups are developed and the role
they play in the prejudice and discrimination that is directed at gay men and
lesbian women.

Notes
1. The SCM is a model that enables one to position social groups on a map according to
the dimensions of perceived “warmth” and “competence.” In 2002, Fiske et al. found
that gay men fell in the centre of the warmth X competence space, which signified
neutral perceptions toward gay men relative to other social groups. The position of gay
men seemed at odds with the well-documented stereotyping, prejudice, and discrimi-
nation reported by this social group.
2. To date, there have been no rigorous examinations of the psychometric properties of
the measure associated with the SCM that assesses a particular social group’s or
subgroup’s warmth and competence.
3. Henrich, Heine, and Norenzayan (2010) pointed out the gross reliance on Western,
Educated, Industrialized, Rich, and Democratic (WEIRD) samples in psychological
research. Through a comparative review of studies on key psychological and behavioral
variables (e.g., reasoning styles, fairness, categorization and inferential induction) they
determined that WEIRD samples are not particularly representative of the overall
human population. The researchers note that American undergraduate students, in
particular, constitute the bulk of the WEIRD samples employed in research. It is
anticipated that our general population sample will be demographically more variable
than an undergraduate student population, and therefore, more representative of the
wider population.
4. The results of Google Image Searches vary by several factors, which effect the repro-
ducibility of photograph selection in the present study. McEvoy (2015) has noted seven
factors that can affect search results that will be generated by Google Image Search.
They include: 1) device (e.g., desktop, tablet) used when searching, 2) personal search
history, 3) being logged in to a Google account, 4) geographic location, 5) browser
type, 6) Google-generated ads on the page, 7) the type of search (i.e., minor keyword
can drastically alter results). Undoubtedly, image results will also change over time.
5. Only photographs that depicted men and women who appeared to be young adults or
middle-aged were selected, as well as only photographs of Caucasian individuals.
Variations in these demographic features could potentially produce attitudinal effects
independent of, or in interaction with, sexual orientation (Woody, 2013); therefore,
efforts were made to keep age and race similar across photographs.
24 J. M. MCCUTCHEON AND M. A. MORRISON

6. Ekos is a Canadian research vendor who provides access to samples of online survey
respondents. Ekos recruits participants to their panel through random digit dialing of
Canadian telephone numbers. Participants are informed of available surveys through
an email issued by Ekos and they are not remunerated for their participation.
7. Being Canadian-born was an inclusion criterion for Study 2 as many of the interna-
tional students in the authors’ university participant pool are first-year students who
have lived in Canada for less than 6 months. This was not considered enough time to
be adequately exposed to culturally-specific gay and lesbian subgroups.

Disclosure statement
No potential conflict of interest was reported by the authors.

References
Adam, B. (1995). The rise of the gay and lesbian movement (Rev ed.). New York, NY: Twayne
Publishers.
Altemeyer, B. (1988). Enemies of Freedom. San Francisco, CA: Jossey Bass.
Altemeyer, B. (2001). Changes in attitudes toward homosexuals. Journal of Homosexuality,
42, 63–75.
Andersen, R., & Fetner, T. (2008). Cohort differences in tolerance of homosexuality:
Attitudinal changes in Canada and the United States, 1981–2000. Public Opinion
Quarterly, 72, 311–330. doi:10.1093/poq/nfn017
Baunach, D. M. (2012). Changing same-sex marriage attitudes in America from 1988 through
2010. Public Opinion Quarterly, 76, 364–378. doi:10.1093/poq/nfs022
Bem, S. L. (1974). The measurement of psychological androgyny. Journal of Consulting and
Clinical Psychology, 42, 155–162.
Blashill, A. J., & Powlishta, K. K. (2009). Gay stereotypes: The use of sexual orientation as
a cue for gender-related attributes. Sex Roles, 61, 783–793. doi:10.1007/s11199-009-9684-7
Brambilla, M., Carnaghi, A., & Ravenna, M. (2011). Status and cooperation shape lesbian
stereotypes: Testing predictions from the stereotype content model. Social Psychology, 42,
101–110. doi:10.1027/1864-9335/a000054
Brewer, M. B., Dull, V., & Lui, L. (1981). Perceptions of the elderly: Stereotypes as prototypes.
Journal of Personality and Social Psychology, 41, 656–670. doi:10.1037/0022-3514.41.4.656
Ciasullo, A. M. (2001). Making her (in)visible: Cultural representations of lesbianism and the
lesbian body in the 1990s. Feminist Studies, 27, 577–608. doi:10.2307/3178806
Clausell, E., & Fiske, S. T. (2005). When do subgroup parts add up to the stereotypic whole?
Mixed stereotype content for gay male subgroups explains overall ratings. Social Cognition,
23, 161–181. doi:10.1521/soco.23.2.161.65626
Craik, F. I., & McDowd, J. M. (1987). Age differences in recall and recognition. Journal of
Experimental Psychology: Learning, Memory, and Cognition, 13, 474–479.
Deocampo, J. A., & Hudson, J. (2010). Reinstatement of 2-year-olds’ event memory using
photographs. Memory, 11, 13–25. doi:10.1080/741938177
Eckes, T. (1994). Exploration in gender cognition: Content and structure of female and male
subtypes. Social Cognition, 12, 37–60. doi:10.1521/soco.1994.12.1.37
Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype
content: Competence and warmth respectively follow from perceived status and
competition. Journal of Personality and Social Psychology, 82, 878–902.
 JOURNAL OF HOMOSEXUALITY 25

Fleiss, J. L., Nee, J. C. M., & Landis, J. R. (1979). Large sample variance of kappa in the
case of different sets of raters. Psychological Bulletin, 86, 974–977. doi:10.1037/0033-
2909.86.5.974
Geiger, W., Harwood, J., & Hummert, M. L. (2006). College students’ multiple stereotypes of
lesbians: A cognitive perspective. Journal of Homosexuality, 51, 165–182. doi:10.1300/
J082v51n03_08
Glick, P., & Fiske, S. T. (1996). The ambivalent sexism inventory: Differentiating hostile and
benevolent sexism. Journal of Personality and Social Psychology, 70, 491–512. doi:10.1037/
0022-3514.70.3.491
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world?
Behavioral and Brain Sciences, 33, 61–235. doi:10.1017/S0140525X0999152X
Herek, G. M. (1988). Heterosexual’s attitudes toward lesbians and gay men: Correlates and
gender differences. The Journal of Sex Research, 25, 451–477. doi:10.1080/
00224498809551476
Herek, G. M., & Capitanio, J. P. (1996). “Some of my best friends”: Intergroup contact,
concealable stigma, and heterosexuals’ attitudes toward gay men and lesbians. Personality
and Social Psychology Bulletin, 22, 412–424. doi:10.1177/0146167296224007
Herek, G. M., & Glunt, E. K. (1993). Interpersonal contact and heterosexuals’ attitudes
towards gay men: Results from a national survey. Journal of Sex Research, 30, 239–244.
doi:10.1080/00224499309551707
Hicks, G. R., & Lee, -T.-T. (2006). Public attitudes toward gays and lesbians: Trends and
predictors. Journal of Homosexuality, 51, 57–77. doi:10.1300/J082v51n02_04
Hirai, M., Winkel, M. H., & Popan, J. R. (2014). The role of machismo in prejudice toward
lesbians and gay men: Personality traits as moderators. Personality and Individual
Differences, 70, 105–110. doi:10.1016/j.paid.2014.06.028
Hummert, M. L. (1990). Multiple stereotypes of elderly and young adults: A comparison of
structure and evaluation. Psychology and Aging, 5, 182–193.
Jewell, L. M., McCutcheon, J., Harriman, R. L., & Morrison, M. A. (2011). It’s like a bunch of
mosquitoes coming at you … ”: Exploring the ubiquitous nature of subtle discrimination
and its implications for the everyday experiences of LGB individuals. In T. G. Morrison,
M. A. Morrison, M. A. Carrigan, & D. T. McDermott (Eds.), Sexual orientation in the new
millennium. New York, NY: Nova Science Publishers.
Kosciw, J. G., Greytak, E. A., Bartkiewicz, M. J., Boesen, M. J., & Palmer, N. A. (2012). The
2011 national school climate survey: The experiences of lesbian, gay, bisexual, and transgen-
der youth in our nation’s schools. New York, NY: Gay, Lesbian & Straight Education
Network.
Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical
data. Biometrics, 33, 159–174. doi:10.2307/2529310
Mason, G. (1993). Violence against lesbians and gay men. Violence Prevention Today Series
No. 2. Canberra: Australian Institute of Criminology.
Mattocks, K. M., Sullivan, C., Bertand, C., Kinney, R. L., Sherman, M. D., & Gustason, C.
(2015). Perceived stigma, discrimination, and disclosure of sexual orientation among
a sample of lesbian veterans receiving care in the department of Veterans Affairs. LGBT
Health, 2, 147–153. doi:10.1089/lgbt.2014.0131
Maxwell, S. E., & Delaney, H. D. (2004). Designing experiments and analyzing data: A model
comparison perspective (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates.
McDermott, D. T., & Blair, K. L. (2012). ‘What’s it like on your side of the pond?’: A cross-
cultural comparison of modern and old-fashioned homonegativity between North
American and European samples. Psychology & Sexuality, 3, 277–296. doi:10.1080/
19419899.2012.700032
26 J. M. MCCUTCHEON AND M. A. MORRISON

McEvoy, M. (2015). 7 reasons Google search results vary dramatically. Retrieved from https://
www.webpresencesolutions.net/7-reasons-google-search-results-vary-dramatically/
Morrison, M. A., Jewell, L. M., McCutcheon, J. M., & Cochrane, D. B. (2014). In the face of
anti-LGBQ behaviour: Saskatchewan high-school students’ perceptions of school climate
and consequential impact. Canadian Journal of Education/Revue Canadienne De
L’éducation, 37, 1–29.
Morrison, M. A., & Morrison, T. G. (2003). Development and validation of a scale measuring
modern prejudice toward gay men and lesbian women. Journal of Homosexuality, 43,
15–37. doi:10.1300/J082v43n02_02
Morrison, M. A., & Morrison, T. G. (2011). Sexual orientation bias toward gay men and
lesbian women: Modern homonegative attitudes and their association with discriminatory
behavioral intentions. Journal of Applied Social Psychology, 41, 2573–2599. doi:10.1111/
jasp.2011.41.issue-11
Morrison, M. A., Morrison, T. G., & Franklin, R. (2009). Modern and old-fashioned homo-
negativity among samples of Canadian and American university students. Journal of Cross-
Cultural Psychology, 40, 523–542. doi:10.1177/0022022109335053
Morrison, T. G., Kenny, P., & Harrington, A. (2005). Modern prejudice toward gay men and
lesbian women: Assessing the viability of a measure of modern homonegative attitudes
within an Irish context. Genetic, Social, and General Psychology Monographs, 131, 219–250.
doi:10.3200/MONO.131.3.219-250
Park, B., Ryan, C. S., & Judd, C. M. (1992). Role of meaningful subgroups in explaining
differences in perceived variability for in-groups and out-groups. Journal of Personality and
Social Psychology, 63, 553–567.
Rayside, D. (2008). Queer inclusions, continental divisions: Public recognition of sexual diver-
sity in Canada and the United States. Toronto: University of Toronto Press.
Richards, Z., & Hewstone, M. (2001). Subtyping and subgrouping: Processes for the preven-
tion and promotion of stereotype change. Personality and Social Psychology Review, 5,
52–73. doi:10.1207/S15327957PSPR0501_4
Saewyc, E., Poon, C., Wang, N., Homma, Y., Smith, A., & Liebel, A. (2007). Not yet equal: The
health of lesbian, gay, & bisexual youth in BC. Vancouver, BC: McCreary Centre Society.
Stotzer, R. L. (2009). Straight allies: Supportive attitudes toward lesbians, gay men, and
bisexuals in a college sample. Sex Roles, 60, 67–80. doi:10.1007/s11199-008-9508-1
Swim, J. K., Pearson, N. B., & Johnston, K. E. (2007). Daily encounters with heterosexism:
A week in the life of lesbian, gay, and bisexual individuals. Journal of Homosexuality, 53,
31–48. doi:10.1080/00918360802101179
Taylor, C., & Peter, T. (2011). Left behind: Sexual and gender minority students in Canadian
high schools in the new millennium. In T. G. Morrison, M. A. Morrison, M. A. Carrigan, &
D. T. McDermott (Eds.), Sexual orientation in the new millennium. New York, NY: Nova
Science Publishers.
Walker, L. M. (1993). How to recognize a lesbian: The cultural politics of looking like what
you are. Signs: Journal of Women in Culture and Society, 18, 866–891. doi:10.1086/494846
Woody, I. (2013). Aging out: A qualitative exploration of ageism and heterosexism among
aging African American lesbians and gay men. Journal of Homosexuality, 61, 145–165.
doi:10.1080/00918369.2013.835603
Yonelinas, A. P., & Jacoby, L. L. (1994). Dissociations of processes in recognition memory:
Effects of interference and of response speed. Canadian Journal of Experimental
Psychology, 48, 516–535.