Download HomophobicNameCalling

Social Development Vol 24 No. 1 184–205 February 2015
doi: 10.1111/sode.12085
Homophobic Name-calling, Peer-groups, and
Masculinity: The Socialization of
Homophobic Behavior in Adolescents
Michelle Birkett1 and Dorothy L. Espelage2
1
Northwestern University
University of Illinois at Urbana-Champaign
2
Abstract
Social network analysis and multilevel modeling were used to examine the formation
of homophobic name-calling behavior in adolescents. Specifically, peer group contextual and socialization effects on homophobic name-calling as well as the influence of
masculinity attitudes, general bullying perpetration, and victimization were tested.
Participants included 493 fifth- through eighth-grade students from two middle
schools. Results indicated that peer groups play an important role in the formation of
homophobic name-calling. Additionally, students who were victims of homophobic
name-calling over time increased their own perpetration of homophobic name-calling.
Non-homophobic bullying was also related to homophobic name-calling, but only for
male peer groups. And finally, the role of masculinity attitudes was shown to be
complex, as peer group masculinity attitudes were significantly predictive of an individual’s homophobic perpetration; however, this effect did not remain significant over
time. Results suggest that homophobic name-calling during early adolescence is
strongly influenced by peers and rooted in gender and masculinity.
Keywords: sexuality; victimization; peers/peer relations; socialization
Introduction
Homophobic name-calling is one of the most common forms of victimization present
in schools. Studies have shown that students report hearing words like ‘faggot’, ‘dyke’,
or ‘queer’ on a nearly daily basis (Kosciw & Diaz, 2006). Recently, several suicides of
gay teenagers have highlighted the impact of victimization on gay students (McKinley,
2010). Research has also shown homophobic name-calling, as well as bullying
and victimization of lesbian, gay, bisexual, and transgender (LGBT) students, to be
This research was supported by Centers for Disease Control & Prevention (#1U01/CE001677) to
Dorothy Espelage (PI).
Correspondence should be addressed to Michelle Birkett, Department of Medical Social Sciences,
Northwestern University, 625 N. Michigan Avenue, Suite 2700, Chicago, IL 60611, USA. Email:
[email protected]
© 2014 John Wiley & Sons Ltd
Homophobic Name-calling and Peer Groups 185
correlated with increased depression, increased suicidality, feeling unsafe in school,
and continued mental health problems into adulthood (D’Augelli & Hershberger,
1993; Hershberger & D’Augelli, 1995). Understanding the formation of homophobic
behaviors in early adolescence may be a key in preventing the later victimization of
LGBT students. This longitudinal study was conducted to understand the formation of
homophobic behavior in adolescents over time. Specifically, it utilized multilevel
modeling and social network analysis to understand how peer groups influence the
development of homophobic name-calling behavior and what role masculinity attitudes play in the transmission of these behaviors.
Peer Contextual Influence
One social context that is of particular importance during adolescence is the peer group
(Brown, 2004). A peer group is typically defined by a small friendship cluster of
adolescents who spend large amounts of time together (Brown, 2004). An increasing
amount of research has shown that a wide range of behaviors and attitudes can be
influenced by adolescent peers (Brechwald & Prinstein, 2011). It has been shown that
children’s identification with their peers reaches its peak in early adolescence (Gavin
& Furman, 1989). Likewise, research has shown that between the ages of 10 and 14
years, children show deficits in their ability to resist peer influence (Steinberg &
Monahan, 2007). Both an increased identification with peers and an inability to resist
peer influence make early adolescence a time when peers are particularly influential on
children (Rubin, Bukowski, & Parker, 2006).
Homophily Hypothesis. A robust explanatory theory within the adolescent peer
research literature is the homophily hypothesis (Kandel, 1978; Lazarsfeld & Merton,
1954). Homophily is the predisposition of children to be connected to peers who are
similar across attitudinal and behavioral characteristics. Homophily involves two separate processes: selection and socialization. Selection suggests that children who are
similar to each other are more likely to form groups (Kandel, 1978). Socialization
refers to how social groups also influence children over time, with children often
picking up attitudes and behaviors of other group members (Kandel, 1978). The
homophily hypothesis has been used to examine a number of sociodemographic
characteristics (McPherson, Smith-Lovin, & Cook, 2001) including gender (Ibarra,
1992), race (Mollica, Gray, & Trevino, 2003), education and occupation (McPherson
& Smith-Lovin, 1987), and attitudes and behaviors such as achievement (Ryan, 2001),
aggression (Espelage, Green, & Wasserman, 2007; Espelage, Holt, & Henkel, 2003),
and frequency of smoking and drug use (Ennett & Bauman, 1994; Kandel, 1978). The
homophily hypothesis will be used in this study as a framework to examine peer
influence and homophobic behavior.
Peers’ Influence on Prejudice. A majority of the studies which have examined peer
groups and prejudice have focused on describing homophily within groups (see
McPherson et al., 2001 for review) with fewer studies examining peer group influence
on prejudice. Those studies that have examined peer influence on prejudice have
included largely racial attitude measures (Aboud, 1989; Aboud & Doyle, 1996). For
example, Aboud and Doyle (1996) used an experimental manipulation to show that
discussing racial issues with peers can lower prejudice. In their study, 8- to 11-year-old
children first completed a measure of racial prejudice. High prejudiced children were
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
186
Michelle Birkett and Dorothy L. Espelage
then paired with low prejudiced children, and they were instructed to discuss their
ratings on the measures. Students who originally reported high racial prejudice
reported lower levels of prejudice after this discussion. This research emphasizes the
strong influence peers are able to exert on prejudiced attitudes.
Poteat (2007) extended previous work on peer contextual effects to homophobic
attitudes and behaviors. In his study, significant similarity in homophobic attitudes
within friendship groups at the initial time point was found. Additionally, evidence of
peer group socialization was found, as adolescent attitudes resembled their peers eight
months later even after controlling for their original attitudes (Poteat, 2007). Another
study by Poteat, Espelage, and Green (2007) indicated that aggressive peer groups
influenced the use of more homophobic epithets. These studies underscored the peer
process underlying the development and proliferation of prejudiced attitudes. However,
this study was limited because it focused on a high school sample in a single
university-based school. This school consisted of a largely affluent and high-achieving
population, and thus it is not clear whether these same processes will hold for an
economically and racially diverse sample of middle school students.
Predictors of Homophobic Attitudes
Although few studies have examined the peer-level influences of sexual prejudice,
various individual-level traits have been found to be correlated with homophobic
attitudes and behavior across multiple studies. One of the strongest predictors of
homophobic attitudes has been traditional masculinity, or masculinity that values
dominance, assertiveness, and a lack of emotion (Epstein, 2001; Mandel & Shakeshaft,
2000; Phoenix, Frosh, & Pattman, 2003). In Whitley’s (2001) meta-analysis of genderrole variables and their prediction of homophobia, traditional gender-role beliefs,
modern sexism, and hyper-masculinity predicted increased homophobic attitudes. In
another study by Whitley, the strongest predictors of homophobic attitudes were being
male, endorsing traditional gender role beliefs, and endorsing sexist attitudes (Whitley,
2001). Masculinity researchers have posited that homophobia is one of the key principles of traditional masculinity (Epstein, 2001; Kimmel & Mahler, 2003). It is hypothesized that men’s fears of being seen as gay themselves often pushes their behavior and
attitudes to become hyper-masculine (Kimmel & Mahler, 2003). Despite the association of homophobic attitudes and masculinity, no research has yet examined the
influence of peer group masculinity attitudes on homophobia.
In addition to masculinity, race has also been shown to be a significant predictor of
homophobic attitudes, with African-Americans often holding more negative attitudes
than White people (Levant, Majors, & Kelley, 1998). However, this finding is likely
related to masculinity attitudes, as African-Americans, particularly African-American
males, are more likely to hold a traditional masculine ideology and more rigid gender
roles (Levant et al., 1998).
In summary, this study capitalized on Aboud (2005)’s recommendation to conduct
research that examines the development of prejudice by assessing variables from
various contextual levels. Most scholarship in the area of prejudice formation has
failed to examine the interaction among contexts adequately. This is especially evident
in research on homophobic attitudes, where most studies focus on individual-level
correlates. Therefore, this current study utilized complex statistical methods to tease
out the influences of individual and peer-level variables on homophobic name-calling
in order to elucidate a greater understanding of these complex systems.
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 187
Study Hypotheses
The current study expanded upon the existing literature by using an ecological framework to study the formation of homophobia in adolescents. Specifically, both a crosssectional contextual model and a longitudinal socialization model were built to
examine peer influence on homophobic name-calling. Additionally, the individual and
peer-group-levels of masculinity, bullying, and victimization were examined for their
possible influence on homophobic name-calling. Hierarchical linear modeling
(Raudenbush & Bryk, 2002) and social network analysis methodology were utilized to
test the study hypotheses.
Individual Predictors of Homophobic Name-calling. Based on the extant literature,
gender, racial, and grade-level differences on homophobic name-calling were
expected. It was hypothesized (1) that males would score higher than females, (2) that
African-Americans would score higher than White people, and (3) that students in
seventh and eighth grades would score higher than students in fifth and sixth grades on
homophobic behaviors, as levels of peer bullying often increase throughout middle
school (DeVoe & Kaffenberger, 2005). Additionally, as bullying has been shown to
correlate with homophobic name-calling (Poteat & Espelage, 2005), bullying perpetration was expected to predict homophobic name-calling.
Peer Contextual Model of Homophobic Name-calling. As peer groups have been
shown to influence individual behaviors and attitudes (Brown, 2004; Poteat, 2007;
Steinberg & Monahan, 2007), it was hypothesized that (4) peer groups would differ
across their levels of homophobic name-calling and that peer-level masculinity attitudes would account for the differences between peer groups, with more traditional
peer-level masculinity attitudes leading to higher prevalence rates of individual
homophobic name-calling. In addition to masculinity attitudes, individual and peer
level of general bullying perpetration and victimization were examined for their influences on homophobic name-calling. And finally, the grade, gender, and racial composition of the peer group were tested as factors accounting for peer group differences in
homophobic name-calling.
Peer Socialization of Homophobic Name-calling. Beyond contextual effects, the possible peer group socialization effect on homophobic name-calling was also tested using
a longitudinal design. It was hypothesized that (5) peer group homophobic namecalling would significantly predict an individual’s homophobic name-calling over time,
even while controlling for an individual’s previous name-calling behavior. It was
expected that members of peer groups with high homophobic name-calling behavior
would increase their own homophobic name-calling behavior.
Method
Participants and Consent Procedures
Participants included 493 fifth- (N = 51), sixth- (N = 134), seventh- (N = 132), and
eighth- (N = 176) grade students from two Midwestern middle schools. School A
contained sixth through eighth grades whereas school B contained fifth through eighth
grades. The sample was 45.2% White, 36.7% Black, and 18.1% Other. Additionally,
the sample was 50.7% female (N = 250) and 49.3% male (N = 243). The survey was
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
188
Michelle Birkett and Dorothy L. Espelage
administered in two waves, once in the spring of 2008 and once in the fall of 2008.
Parental permission forms were sent to all students registered at the middle schools
prior to data collection, and parents were asked to sign and return the consent form
only if they wished that their child would not participate in the study. Students were
also read an assent script prior to data collection and could opt out of the survey. This
study was approved by the university institutional review board and a certificate of
confidentiality was granted for the data. Participation rates for all grades across both
schools ranged from 90% to 95%, with a total participation rate of 93% at wave 1.
Retention rates from wave 1 to wave 2 ranged from 76% to 84%, with a total retention
of 79%.
Measures
Demographic Variables. Self-reports of gender, grade, and race were elicited to determine demographic characteristics.
Homophobic Name-calling. The Homophobic Content Agent Target Scale (HCAT;
Poteat & Espelage, 2005) is comprised of two subscales that assess the extent to which
individuals called others homophobic epithets (perpetration; five items) and were
called homophobic epithets by others (victimization; five items) during the previous 30
days. Response options include never, one or two times, three or four times, five or six
times, or seven or more times.
The first scale, the perpetration scale, measures how many times in the past 30 days
a child has called others homophobic epithets. Item wording was ‘Some kids call each
other names like homo, gay, lesbo, fag, or dyke. How many times in the last 30 days did
you say these words to the following persons? followed by five relationship probes such
as A friend?’ Higher scores indicate higher homophobic perpetration. Poteat and
Espelage (2005) found that the perpetration scale was moderately correlated with the
Illinois bully scale (Espelage & Holt, 2001), (males, r = .61; females, r = .58) and
negatively correlated with the measures of empathic concern (males, r = −.42; females,
r = −.23) and perspective-taking (males, r = −.28; females, r = −.31) (IRI; Davis, 1983).
In the current study, the Cronbach alpha coefficient of .80 indicates strong reliability.
The second scale, the victimization scale, measures how many times in the past 30
days a child was called homophobic epithets by others. Item wording was ‘Some kids
call each other names like homo, gay, lesbo, fag or dyke. How many times in the last
30 days did the following persons say these words to you?’ followed by five relationship probes such as ‘A friend?’. Higher scores indicate higher homophobic victimization. Poteat and Espelage (2005) found that the victimization scale was strongly
correlated (males, r = .62; females, r = .68) with the University of Illinois victimization
scale (Espelage & Holt, 2001). In the current study, an α = .69 was calculated.
General Bullying and Victimization. The University of Illinois aggression scales
(Espelage & Holt, 2001) assess the occurrence of bullying behavior and victimization
by peers. For all items, students are asked to indicate how often in the past 30 days they
have engaged in a specified behavior. Two subscales are focused on in this study. The
first, the bullying scale, contains nine items specifying bullying behaviors including
teasing, social exclusion, name-calling, and rumor spreading such as I teased other
students. Higher scores indicate higher self-reported bullying. For this sample, a
Cronbach alpha coefficient of .86 was obtained. The second subscale, the victimization
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 189
scale, contains four items assessing victimization by peers such as Other students
called me names. Higher scores indicate more self-reported victimization. In the
present study, the Cronbach alpha coefficient was .79 for this scale.
Masculine Attitudes. The traditional masculinity scale of the Adolescent Masculinity
Ideology in Relationships Scale (AMIRS; Chu, Porche, & Tolman, 2005) assesses the
level of traditional masculinity attitudes held by an individual. It can be used to assess
either males or females on their traditional masculinity attitudes. The AMIRS is unique
in that it was created specifically for use with adolescents whereas most measures of
masculinity attitudes are created for use with an adult population (Chu et al., 2005).
Additionally, it was created to specifically assess adolescent masculinity within relationships, which makes it an ideal measure to examine peer contextual effects. One of
the examples of items is: It’s important for a boy to act like nothing is wrong, even
when something is bothering him. The AMIRS has been shown to be moderately
correlated with the Male Role Attitudes Scale (Snell, 1989) (r = .48), suggesting
convergent validity (Chu et al., 2005). Additionally, it has been shown to be negatively
correlated with the Attitudes Towards Women Scale for Adolescents, which is a
measure of unconventional attitudes toward women’s roles and rights (Galambos,
Petersen, Tobin-Richards, & Gitelson, 1985) (r = −.42), providing evidence of discriminant validity. An examination of the scale found that seven items load on the first
factor and represented the traditional masculinity scale, with the remaining five items
representing the non-traditional masculinity scale (Newlin, 2009). An alpha coefficient
of .76 was found in this study on the traditional masculinity scale.
Friendship Nominations. The Ennett and Bauman (1994) approach was utilized to
collect friendship nominations. Students were asked to list up to eight students at their
school with whom they spent the most time and considered their friends, excluding
siblings. These names were converted to the code number of the respective participant
and matched with the corresponding survey data.
Results
Preliminary Data Analyses
Before constructing the Hierarchical linear modeling (HLM) models, preliminary
analyses of demographic differences across study measures were conducted. Differences across homophobic name-calling perpetration and victimization, traditional
masculinity attitudes, and bullying perpetration and victimization were examined.
Gender Differences on Wave 1 Measures. Youth did not differ on levels of bullying
perpetration or levels of bullying victimization by gender, but females reported lower
scores on homophobic name-calling perpetration compared with males, as well as
reported significantly lower scores on homophobic name-calling victimization compared with males. Lastly, females reported significantly lower levels of traditional
masculinity attitudes compared with males (see Table 1).
Racial Differences on Wave 1 Measures. The results revealed a statistically significant
difference in scores of bullying perpetration between races. Post hoc comparisons with
Scheffé’s statistic suggested that African-American youth reported significantly
greater bullying perpetration over both White, biracial, and youth of another ethnicity.
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
© 2014 John Wiley & Sons Ltd
222
181
90
1.33 (.50)
1.51 (.60)
1.32 (.45)
1.25 (.36)
1.24 (.34)
1.52 (.60)
1.46 (.61)
1.38 (.52)
1.41 (.55)
M (SD)
236
257
N
51
134
132
176
1.42 (.60)
1.37 (.46)
243
250
M (SD)
6.62**
8.63**
F
−.66
−.99
t
1.62 (.84)
1.55 (.83)
1.78 (.91)
1.87 (1.12)
1.56 (.79)
1.77 (.94)
1.49 (.70)
1.61 (.81)
1.63 (.89)
M (SD)
1.63 (.87)
1.62 (.83)
M (SD)
Victim
2.15
4.46**
F
−.23
−.12
t
1.60 (.82)
1.73 (.92)
1.67 (.92)
1.33 (.60)
1.46 (.76)
1.81 (.96)
1.80 (.91)
1.71 (.88)
1.62 (.87)
M (SD)
1.79 (.94)
1.54 (.79)
M (SD)
t
1.15
7.91**
F
1.22
3.24**
HCATp
1.49 (.70)
1.37 (.63)
1.37 (.53)
1.39 (.65)
1.30 (.49)
1.50 (.82)
1.47 (.60)
1.50 (.72)
1.35 (.56)
M (SD)
1.51 (.73)
1.34 (.55)
M (SD)
t
2.03
2.58
F
2.55*
2.96**
HCATv
1.86 (.48)
1.98 (.52)
1.99 (.43)
1.74 (.54)
1.91 (.49)
1.95 (.48)
1.97 (.47)
1.95 (.47)
1.91 (.51)
M (SD)
2.09 (.47)
1.77 (.46)
M (SD)
Masc
4.06*
2.89*
F
.86
7.42**
t
Note: Bully = bully perpetration; Victim = bully victimization; HCATp = homophobic name-calling perpetration; HCATv = homophobic name-calling
victimization; Masc = traditional masculinity.
* p < .05, ** p < .01.
Grade
5
6
7
8
Race
White
Black
Biracial/Other
Gender
Male
Female
School
A
B
N
Bully
Table 1. Demographic Differences Across Measures
190
Michelle Birkett and Dorothy L. Espelage
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 191
Differences in traditional masculinity attitudes were also found between races. Post
hoc comparisons suggested that White youth reported significantly lower traditional
masculinity attitudes than African-American, biracial, and youth of another ethnicity.
The analyses did not, however, find any statistically significant differences across race
on bullying victimization, homophobic name-calling perpetration nor homophobic
name-calling victimization (see Table 1).
Grade Differences on Wave 1 Measures. The results revealed a significant difference in
bullying perpetration between grades. Post hoc comparisons with Scheffé’s statistic
suggested that seventh and eighth graders had higher bullying perpetration than fifth
and sixth graders. There was also a difference in bullying victimization between
grades, with post hoc comparisons indicating that eighth graders reported higher
bullying victimization than fifth graders. Homophobic name-calling perpetration also
showed a difference in the mean scores between grades. Post hoc comparisons indicated that seventh and eighth graders perpetrate greater homophobic name-calling than
fifth or sixth graders. And finally, results also revealed differences in the traditional
masculinity attitudes across grades. Post hoc comparisons indicated that seventh and
eighth graders reported higher traditional masculinity attitudes than fifth or sixth
graders. No differences on mean rates of homophobic name-calling victimization were
found across grade (see Table 1).
School Differences Across Measures. Schools were examined for differences on
demographic characteristics and study measures. A chi-square revealed that the proportion of African-American students was greater at school A vs. school B (55% vs.
17% African-American), χ2, (2, N = 493) = 88.43, p < 001. Schools did not differ from
each other on bullying perpetration, bullying victimization, homophobic name-calling
perpetration, or traditional masculinity attitudes. Schools did differ on the amount of
homophobic name-calling victimization that they indicated, with school A reporting a
higher amount than school B (see Table 1).
Correlations Among Variables at Wave 1 and Wave 2. In order to examine study
measures and their relations to each other, a correlation matrix was created with the
study variables at both wave 1 and wave 2 (see Table 2). In order to guard against alpha
inflation, a Bonferroni correction was used to establish the alpha level at .001 (.05/45).
As shown in the table, the correlations between many of the study measures were
significant (ps < .001). Only traditional masculinity attitudes were consistently not
correlated with bullying victimization and homophobic victimization at both wave 1
and wave 2. Also of note, bullying perpetration was highly correlated with homophobic
perpetration (r = .66 at wave 1, r = .60 at wave 2). As shown in Table 1, measures across
wave 1 and wave 2 show strong stability. The stability coefficient was r = .51 for
bullying perpetration, r = .57 for bullying victimization, r = .56 for homophobic
perpetration, r = .60 for homophobic victimization, and r = .52 for traditional masculinity attitudes (ps < .001).
Identification of Peer Groups at Wave 1. Peer group formation was conducted using
wave 1 friendship nomination data. First students were separated by grade, and then
multiple matrices of strongly tied peers (e.g., peers with reciprocated friendships)
were created. The decision to examine only students with strongly tied peers was
because research has shown reciprocated adolescent friendships last longer and are
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
© 2014 John Wiley & Sons Ltd
—
.51*
.22*
.50*
.31*
.22*
1.39 (.57)
1.63 (.87)
1.68 (.89)
1.38 (.69)
1.84 (.53)
1
1.39 (.53)
1.62 (.85)
1.66 (.87)
1.42 (.64)
1.92 (.49)
M(SD)
.13
.57*
.19
.38*
.04
.27*
—
2
.39*
.21*
.56*
.40*
.28*
.66*
.23*
—
3
.33*
.50*
.41*
.60*
.20
.36*
.52*
.56*
—
4
.24*
.03
.32*
.21*
.52*
.25*
.02
.30*
.14
—
5
—
6
.25*
—
7
.60*
.21*
—
8
.53*
.41*
.66*
—
9
.31*
.01
.24*
.23*
—
10
Note: Bully perp = bullying perpetration; Bully vict = bullying victimization; HCAT perp = homophobic name-calling perpetration; HCAT vict =
homophobic name-calling victimization; Masc = traditional masculinity. Values in bold on the lower diagonal indicate measure stability coefficients from
wave 1 to wave 2.
* Sig. at .001.
Wave 1
1. Bully perp
2. Bully vict
3. HCAT perp
4. HCAT vict
5. Masc
Wave 2
6. Bully perp
7. Bully vict
8. HCAT perp
9. HCAT vict
10. Masc
Wave
Table 2. Individual Means, Standard Deviations, and Correlations Across Measures
192
Michelle Birkett and Dorothy L. Espelage
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 193
considered to be closer friends with higher levels of contact (Bukowski & Newcomb,
1984). Additionally, because of the longitudinal nature of this study, the use of
reciprocated friendships increased the stability of peer groups over time. And
finally, preliminary analyses with both reciprocated and unreciprocated ties suggested that higher fit statistics would be obtained by examining only reciprocated
friendships.
Peer groups were constructed within grade due to both the longitudinal nature of the
study and preliminary analyses. These analyses indicated that less than 1% of students
reported reciprocated friendships across grade and that higher fit statistics were
achieved when students were split by grade. Students, who did not have any reciprocated friendships, also known as type 1 isolates, were excluded from analyses as they
did not significantly differ across any demographic measures, and this was a study of
peer effects.
Matrices of reciprocated friendships were then entered into the UCINET 6 and
NetDraw programs (Borgatti, Everett, & Freeman, 1999) where the Girvan–Newman
algorithm (Girvan & Newman, 2002) was used to identify groups of students who
reported having more friendships with each other than with students in other groups.
The Girvan–Newman algorithm is a preferred method for subgroup creation with
these data as it creates groups of students that are highly connected within and less
connected between groups. Additionally, because the resulting groups retain no
overlap or connections to each other—unlike other methods of group detection (see
Gest, Moody, & Rulison, 2007)—this method is appropriate for use in hierarchical
linear modeling. The algorithm’s steps for community detection are: (1) the betweenness of all edges in the network is calculated; (2) the edge with the highest betweenness is removed; (3) the betweenness of all edges affected by the removal is
recalculated; (4) steps 2 and 3 are repeated until no edges remain. The algorithm also
produces a fit statistic called the modularity (Q), for each possible number of peer
groups. The larger the value of Q, the stronger the group structure, therefore the
program user can be confident that the number of peer groups chosen is the best fit
for the data. Networks with good community structure commonly report Qs that
range from .3 to .7 (Du, Feldman, Li, & Jin, 2007). Fifty-nine peer groups were found
across all students. See Tables 3 and 4 for full descriptives and Figure 1 for a graphical representation.
Hierarchical Linear Modeling
Three steps were involved in developing a two-level hierarchical linear model. First, a
fully unconditional (null) model was estimated. Through this, within-group and
between-group components that explained variance in the outcome variable were
calculated. An intra-class correlation (ICC) was then calculated to determine the
amount of variance between groups, as well as a χ2 which revealed if between-group
variation existed. Then multilevel modeling building occurred, in which level 2 variables (i.e., peer group-level variables) were hypothesized to explain level 1 (i.e.,
individual-level) parameter variation on homophobic perpetration. Level 2 variables
were calculated by finding the wave 1 mean scale score for each peer group. In order
to assist with interpretation, level 2 variables were also grand mean centered
(Raudenbush & Bryk, 2002). A level 1 random-intercept model was tested next to
evaluate slope heterogeneity. Heterogeneous slopes would indicate that the relation
between level 1 variables and the outcome variable were different between peer groups.
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
© 2014 John Wiley & Sons Ltd
4
5
0
2
6
1
2–11
5
5
1
1
10
0
2–12
0
7
2
4–13
4
5
0
9
(.70)
36.2%
Eighth
3
1
1
6–11
2
1
2
5
(.60)
41.8%
Fifth
3
2
1
4–13
3
3
0
6
(.66)
35.9%
Sixth
3
2
2
3–9
4
3
0
7
(.71)
21.9%
Seventh
School B
8
3
1
2–11
5
6
1
12
(.77)
28.9%
Eighth
20
31
8
2–13
27
28
4
59
(.60–.77)
31.9%
Total
Note: Peer groups identified as female if >80% of members were female; male if >80% identified as male; and mixed gender if the peer group did not fall
into the two previous categories. Peer groups identified as mostly Black if >80% of members were black; mostly White if >80% of members identified as
White; mixed race if the peer group did not fall into the two previous categories.
9
(.60)
29.0%
11
(.71)
29.6%
Number of peer groups
(Q)
% Reciprocated
Gender
Female
Male
Mixed gender
Race
Mostly Black
Mostly White
Mixed race
Range in size
Seventh
Sixth
Variable
School A
Table 3. Peer Group Descriptive Statistics at School and Grade Level
194
Michelle Birkett and Dorothy L. Espelage
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 195
Table 4. Means and Standard Deviations of Variables at the Peer Group Level,
and at Both Waves
Wave
Wave 1
Bully perpetration
Bully victimization
HCAT perpetration
HCAT victimization
Masculinity
Wave 2
Bully perpetration
Bully victimization
HCAT perpetration
HCAT victimization
Masculinity
Mean
(SD)
1.38
1.60
1.61
1.41
1.92
(.30)
(.44)
(.51)
(.35)
(.27)
1.38
1.62
1.65
1.39
1.81
(.28)
(.42)
(.58)
(.34)
(.31)
HCAT = homophobic-name-calling.
(a)
(b)
Figure 1. Peer Networks of School A and B.
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
196
Michelle Birkett and Dorothy L. Espelage
What follows is an examination of homophobic name-calling perpetration (HCATp)
and the possible individual and peer contextual effects, followed by an analysis of peer
group socialization of HCATp over time.
Examination of the Contextual Effect of Groups. An unconditional model of an individual’s level of homophobic name-calling perpetration (HCATp) at wave 1 was
created (see equation 1).
Level 1: HCATpij = β0 j + rij
(1)
The resulting chi-square analysis rejected the null hypothesis that mean homophobic
perpetration scores across all peer groups were equal, χ2(58, N = 389) = 151.80,
p < .05. An ICC revealed that 20% of the variance of homophobic perpetration existed
between peer groups.
Average peer group level of masculinity attitudes (PG_Masc) entered on the level 2
(see equation 2).
Level 1: HCATpij = β0 j + rij
Level 2: β0 j = γ 00 + g 01 (PG_Masc) + u 0 j
(2)
Peer group traditional masculinity attitudes accounted for 62% of the variance of
homophobic perpetration between peer groups, and a significant portion of variance
remained. Multiple contextual models were examined by adding individual-level scale
scores of bullying behavior, bullying victimization, traditional masculinity attitudes,
and homophobic victimization into the model. The fit of average peer group levels of
masculinity, bullying perpetration and victimization, and homophobic victimization
was also examined. And finally, due to differences across study measures on grade,
gender, and race, these were also entered at the level 1, as well at the level 2 in which
they were entered as peer group averages. Models were built slowly, entering two
variables at a time and retaining a variable if the final estimation of fixed effects
indicated a p value of p < .05, as well as a reduction of model deviance. The best fitting
contextual model with fixed slopes had level 1 variables of bullying perpetration
(bully) and homophobic victimization (HCATv) and a level 2 variable of peer group
traditional masculinity (see equation 3).
Level 1: HCATpij = β0 j + b1j (Bully ij ) + b 2j (HCATv ij ) + rij
Level 2: β0 j = γ 00 + g 01 (PG_Masc) + u 0 j
β1 j = γ 01
(3)
β 2 j = γ 02
This model accounted for 97% of the variance between groups and 47% of the
variance within groups compared with the null model. The model parameters indicated
that beyond having peers with high traditional masculinity attitudes, those who participate in traditional bullying as well as those who are called homophobic names
themselves will be most likely to call others homophobic names. Level 1 slopes were
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 197
then allowed to vary randomly to examine the relationship of homophobic perpetration
with bullying perpetration and homophobic victimization. Results revealed that the χ2
associated with the bullying slope was statistically significant, χ2(57, N = 389) = 95.08,
p < .05. Gender of the peer group was then added as a moderator (see equations 4
and 5).
Level 1: HCATpij = β0 j + b1j (Bully ij ) + b 2j (HCATv ij ) + rij
Level 2: β0 j = γ 00 + g 01 (PG_Masc) + u 0 j
(4)
β1 j = γ 01 + u 0 j
β 2 j = γ 02
Level 1: HCATpij = β0 j + b1j (Bully ij ) + b 2j (HCATv ij ) + rij
Level 2: β0 j = γ 00 + g 01 (PG_Masc) + u 0 j
(5)
β1 j = γ 01 + g 11 (PG_Gender) + u 0 j
β 2 j = γ 02
The results indicated that the gender of the peer group was a significant moderator
of the relationship between bullying and homophobic perpetration, with the relationship being increased in male peer groups. Further, gender of the peer group accounted
for 11% of the variance in slopes across groups (see Table 5 for full parameter
estimates).
Table 5. Parameter Estimates and Model Fit for the Contextual Model of
Homophobic Perpetration
Model 5
Fixed effects
For intercept (β0j)
Intercept (γ00)
PG_Masc (γ01)
For bully slope(β1j)
Bully (γ10)
Gender (γ11)
For HCAT vict (β2j)
HCAT vict (γ20)
Random effects
Bully slope (u0j)
Model fit
Coef.
SE
t
p
1.66
.75
.03
.12
55.15
6.04
.000
.000
.76
.15
.08
.07
9.58
2.07
.000
.043
.47
.07
Variance between
.011**
.098**
Deviance
670.50
6.92
.000
Variance within
.297
Parameters
4
** p < .01.
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
198 Michelle Birkett and Dorothy L. Espelage
Peer Group Socialization. As McPherson et al. (2001) recommended, in order to
examine peer group socialization, this model incorporated longitudinal data from wave
2, fall 2008. Due to the use of longitudinal data, the wave 1 eighth-grade cohort was not
included in these analyses, as that cohort of students had graduated at the time of the
wave 2 data collection. The eighth-grade students were examined, and they did not
differ on any measures of demographics beyond age.
Socialization models were constructed to examine if peer group homophobic namecalling at wave 1 (PG_HCATpT1) predicted individual homophobic perpetration at
wave 2 (HCATpT2), even after controlling for an individual’s original level of
homophobic perpetration at wave 1 (HCATpT1). In the first unconditional model,
differences in peer-level homophobic name-calling scores at wave 2 were examined
(see equation 6).
Level 1: HCATpT2ij = β0 j + rij
(6)
Level 2: β0 j = γ 00 + u 0 j
Differences between peer groups on the homophobic perpetration were found, χ2(38,
N = 212) = 82.73, p < .05. An ICC indicated that 18.4% of the variance of homophobic
perpetration existed between peer groups.
An individual’s wave 1 level of homophobic perpetration entered as a predictor of
their wave 2 homophobic perpetration (see equation 7).
Level 1: HCATpT2ij = β0 j + b1j (HCATpT1ij ) + rij
Level 2: β0 j = γ 00 + u 0 j
(7)
β1 j = γ 01
Individual homophobic perpetration wave 1 scores accounted for 35% of the within
group variance of wave 2 scores, and a significant portion of variance between groups
was left to be explained, χ2(34, N = 212) = 51.70, p < .05.
Random slopes were also specified to determine if peer groups differed in the
extent that wave 1 homophobic perpetration scores predicted their wave 2 scores (see
equation 8).
Level 1: HCATpT2ij = β0 j + b1j (HCATpT1ij ) + rij
Level 2: β0 j = γ 00 + u 0 j
(8)
β1 j = γ 01 + u 0 j
The relationship between wave 1 and wave 2 homophobic perpetration significantly
differed by groups, and a random slopes and intercept model was retained, χ2(34, N =
212) = 76.02, p < .05.
Next, average peer group levels of homophobic perpetration during wave 1 were
entered at the level 2 as a test of peer group socialization over time. A socialization
effect was found, as peer group level of homophobic perpetration at wave 1significantly impacted the intercept of the level 1 equation (see equation 9).
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 199
Level 1: HCATpT2ij = β0 j + b1j (HCATpT1ij ) + rij
Level 2: β0 j = γ 00 + g 01 (P
PG_HCATpT1) + u 0 j
(9)
β1 j = γ 01 + u 0 j
This model indicated that despite controlling for an individual’s homophobic perpetration at wave 1, a significant additional amount of variance was predicted by the
homophobic behaviors of their peer group at wave 1.
And finally, peer group levels of gender, grade, race, and homophobic victimization
at wave 1 were entered as predictors of the intercept and slope, as well as peer group
level of homophobic perpetration as a predictor of the level 1 slope. Of these variables,
only peer group level of homophobic victimization (PG_HCATvT1) was significant, as
it moderated the association between wave 1 and wave 2 homophobic perpetration (see
equation 10).
Level 1: HCATpT2ij = β0 j + b1j (HCATpT1ij ) + rij
Level 2: β0 j = γ 00 + g 01 (P
PG_HCATpT1) + u 0 j
(10)
β1 j = γ 01 + g 11 (PG_HCATvT1) + u 0 j
The final model included an individual’s homophobic perpetration at wave 1, which
was moderated by the amount of homophobic victimization their peer group had
experienced at wave 1. Students within peer groups that experienced high amounts of
victimization at wave 1 had a high association between their past levels homophobic
name-calling and their future levels homophobic name-calling. Peer group level of
homophobic perpetration at wave 1 was also a significant predictor of an individual’s
wave 2 behavior, with highly homophobic peer groups leading to higher individual
levels of homophobic name-calling, after controlling for an individual’s original level
of homophobic name-calling. This model accounted for 95% of the variance between
peer groups on average homophobic perpetration, and the remaining variance of the
intercept parameter became non-significant, χ2(33, N = 212) = 41.86, p > .05 (see
Table 6 for full parameter estimates). This model accounted for 43% of the variance of
the level 1 slope between peer groups, although a significant portion of variance still
existed with the slope parameter, χ2(33, N = 212) = 60.40, p < .05. See Table 7 for a
summary of both final models.
Discussion
These findings are in line with a large body of research that has demonstrated the
strong influence peers exert on adolescents (Kandel, 1978) and the role peer groups
play in the formation and maintenance of aggressive behaviors (Cairns, Cairns,
Neckerman, Gest, & Gariepy, 1988; Espelage et al., 2003, 2007). This study expands
previous work by demonstrating the role adolescent peers play in influencing
homophobic name-calling. Analyses confirmed that not only do adolescent peer groups
differ in their levels of homophobic perpetration but also that peer group levels of
homophobic perpetration influenced individual levels of homophobic perpetration over
time, despite controlling for an individual’s original behavior. These results are both
consistent with Poteat’s (2007) findings and expand his findings to a larger, younger,
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
200
Michelle Birkett and Dorothy L. Espelage
Table 6. Parameter Estimates and Model Fit for the Socialization Model of
Homophobic Perpetration
Model 10
Fixed effects
For intercept (β0j)
Intercept (γ00)
PG_HCATperpT1(γ01)
For HCAT slope (β1j)
HCATPerpT1(γ10)
PG_HCATVict (γ11)
Random effects
HCAT Perp slope (u0j)
Model fit
Coef.
SE
t
p
1.63
.38
.05
.13
33.13
2.87
.000
.007
.49
.09
.60
.24
Variance between
.008
.068*
Deviance
457.43
5.31
.000
2.47
.018
Variance within
.449
Parameters
4
* p < .05.
and more diverse sample. It is clear that even young adolescents can be influenced by
their peer group’s homophobic name-calling behavior over time.
Beyond demonstrating the influence of peers, this study also provides insight into
which individuals and peer groups are mostly likely to perpetrate homophobic namecalling. A particularly interesting finding is that being called homophobic names is
associated with calling others homophobic names. Results showed that those who
perpetrated were likely to be victims themselves. Additionally, this effect remained
over time, as students who were members of peer groups that experienced high levels
of victimization at wave 1 reported significantly increased levels of homophobic
perpetration at wave 2, even after controlling for their original levels of perpetration.
These students perpetrated more homophobic name-calling, seemingly in response to
the name-calling that they experienced. These analyses are likely assessing the use of
homophobic language in the banter that often occurs among adolescent friendship
groups. Although homophobic banter is prevalent in middle schools and high schools
(Poteat & Espelage, 2005), victims of homophobic bullying and name-calling are at
risk for increased mental health problems, poor school outcomes, and increased drug
use (Birkett, Espelage, & Koenig, 2009). Additionally, even when name-calling may be
a playful exchange among friends, the presence and tolerance of homophobic language
in schools may create a hostile environment for LGBT students, and in turn lead to
increased health problems for these youth (Birkett et al., 2009).
This study also points to young males and their peer groups as important intervention targets. Those who participated in traditional bullying were more likely to have
high levels of homophobic name-calling, but this result was only found for male
peer groups. This is not surprising given the documented link between bullying
and homophobic name-calling (Birkett et al., 2009; Poteat & Espelage, 2005;
Rivers, 2004), and the literature that suggests that boys in particular use homophobic
name-calling to assert their dominance over others (Epstein, 2001; Mandel &
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 201
Table 7. Summary of Final Models of Homophobic Perpetration
Predictor variables
Individual—Level 1
Grade
Gender
Race
HCAT perpetration
HCATPerp x PG_HCATVict
HCAT victimization
Bully perpetration
BullyPerp x PG_Gender
Bully victimization
Traditional masculinity
Peer group—Level 2
Grade
Gender
Race
HCAT perpetration
HCAT victimization
Bully perpetration
Bully victimization
Traditional masculinity
HCATPerp at wave1
HCATPerp at wave2
Contextual model
Socialization model
✓
✓a
✓
✓a
✓
✓
✓
Note: All predictor variables are from Wave 1.
a
Level 1 predictor variable that was found to have a random slope. Cross-level predictor
indicated below.
Shakeshaft, 2000; Phoenix et al., 2003). These findings suggest the necessity of interventions that teach young men to assert themselves in ways beyond bullying and
homophobia.
Results also demonstrated that the influence of masculinity attitudes on homophobic
perpetration is complex. Peer group masculinity attitudes were significantly predictive
of an individual’s homophobic perpetration, with students in peer groups that held high
traditional masculinity attitudes perpetrating greater homophobic name-calling.
However, over time, individuals did not change their amount of perpetration based on
their peer group’s masculinity attitudes, as the socializing effect became nonsignificant after controlling for an individual’s original amount of homophobic namecalling. Despite this, results suggest that masculinity attitudes and gender play a
significant role in maintaining behaviors and norms within peer groups. Males who
bully, who have been called homophobic names themselves, and who have friends with
traditional masculinity attitudes and high amounts of homophobic bantering will be
among the most likely to call others homophobic names.
In terms of intervention, these results suggest the importance of teaching youth early
that name-calling, bullying, and homophobia are not an acceptable ways of showing
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
202
Michelle Birkett and Dorothy L. Espelage
authority, power, or masculinity. Schools, teachers, and parents can be a powerful force
encouraging a new dominant narrative of strength, a narrative that encompasses
strength of character, openness to differences, and asking for help when necessary.
Adults can role model to all youth, particularly boys, that true strength comes from
within and not from emasculating others. Additionally, as the experience of victimization may push students to participate in this cycle as a perpetrator, it is particularly
important for interventions to occur at an early age before the onset of victimization
experiences. And finally, these results suggest that interventions which explore the role
that peers play in the development and maintenance of homophobic behavior may be
particularly effective.
This study demonstrated multiple strengths. In particular, the use of social network
analysis and multilevel modeling allowed for a rich analysis of both individual and
peer-group factors of homophobic name-calling (Aboud, 2005), and the nuanced
processes behind how adolescents develop homophobic behavior were examined.
Additionally, the use of a diverse sample of children allowed for the cross-level
modeling of race, gender, and grade differences within results. And finally, this study
examined an important but under-researched area which has implications for a highrisk age group and population.
Despite these strengths, several limitations should be noted. Firstly, the social
network analysis was conducted on only those students enrolled in the study. Although
the response rate was high (90% to 95%), some students not completing the friendshiplisting task meant the full network was not captured. Missing data—particularly with
reciprocated ties—can significantly disrupt the ability to construct a social network
(Neal, 2008). However, in comparison to other studies that have examined the social
networks of school children, our data are comparable if not stronger. The Add Health
data contain some of the largest and most used collections of school-based social
networks; however, schools were included with up to 50% of their data missing
(Bearman, Moody, & Stovel, 1997). Another limitation of the present study is that
students could list a limited number of friendships, which has been suggested to distort
network structure (Holland & Leinhardt, 1973). Additionally, isolates were not able to
be considered given the uncertainty of whether students who had no friendships were
indeed isolates given the incomplete nature of friendship lists. And finally, as we only
examined the phenomena over a short period of time, this study is unable to determine
the influence of peers on name-calling over the course of middle school.
One of the more fascinating findings of this investigation was that perpetrators are
often also the victims of homophobic perpetration themselves. Therefore, future
research should examine if this effect holds true across all individuals. Are there
certain students who are more resilient to their peers’ behavior? Further examining the
characteristics of youth who increase their name-calling behavior and those who do not
will help to advance intervention strategies.
This study highlights the importance of the peer context in understanding behavior
and attitude formation in adolescents. Adolescent peer groups were shown to impact
the formation of homophobic name-calling. In addition to the importance of peers,
these analyses gave insight on the role of gender and masculinity in homophobia. Peer
groups high in traditional masculinity showed increased homophobic name-calling,
and male peer groups showed an increased relationship between homophobic namecalling and bullying behaviors. These results inform intervention by indicating that
homophobic name-calling during early adolescence is strongly influenced by peers and
rooted in gender and masculinity. Interventions for homophobic name-calling can be
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 203
most successful when they target youth early, involve peers, and subvert individual and
peer group notions of traditional masculinity.
References
Aboud, F. E. (1989). Disagreement between friends. International Journal of Behavioral Development, 12, 495–508.
Aboud, F. E. (2005). The development of prejudice in childhood and adolescence. In J. F.
Dovidio, P. Glick, & L. A. Rudman (Eds.), On the nature of prejudice: Fifty years after Allport
(pp. 310–326). Malden, MA: Blackwell.
Aboud, F. E., & Doyle, A. B. (1996). Does talk of race foster prejudice or tolerance in children?
Canadian Journal of Behavioural Science/Revue Canadienne Des Sciences Du
Comportement. Special Issue: Ethnic Relations in a Multicultural Society, 28, 161–170. doi:
10.1037/0008-400X.28.3.161
Bearman, P. S., Moody, J., & Stovel, K. (1997). The Add Health network variable codebook.
Chapel Hill: University of North Carolina.
Birkett, M. A., Espelage, D. L., & Koenig, B. W. (2009). Homophobic bullying and school
climate as moderators of negative outcomes in LGBQ youth. Journal of Youth and Adolescence, 38, 989–1000. doi: 10.1007/s10964-008-9389-1
Borgatti, S. P., Everett, M. G., & Freeman, L. C. (1999). UCINET 5.0 for Windows. Natick, MA:
Analytic Technologies.
Brechwald, W. A., & Prinstein, M. J. (2011). Beyond homophily: A decade of advances in
understanding peer influence processes. Journal of Research on Adolescence, 21, 166–179.
Brown, B. B. (2004). Adolescents’ relationships with peers. In R. M. Lerner, & L. Steinberg
(Eds.), Handbook of adolescent psychology (2nd ed., pp. 363–394). New York: Wiley.
Bukowski, W. M., & Newcomb, A. F. (1984). Stability and determinants of sociometric status
and friendship choice: A longitudinal perspective. Developmental Psychology, 20, 941–952.
Cairns, R. B., Cairns, B. D., Neckerman, H. J., Gest, S. D., & Gariepy, J.-L. (1988). Social
networks and aggressive behavior: Peer support or peer rejection? Developmental Psychology, 24, 815–823. doi: 10.1037/0012-1649.24.6.815
Chu, J., Porche, M., & Tolman, D. (2005). The Adolescent Masculinity Ideology Scale: Development and validation of a new measure. Men and Masculinities, 8, 93–115. doi: 10.1177/
1097184X03257453
Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach. Journal of Personality and Social Psychology, 44, 113–126.
D’Augelli, A. R., & Hershberger, S. L. (1993). Lesbian, gay, and bisexual youth in community
settings: Personal challenges and mental health problems. American Journal of Community
Psychology, 21, 421–448. doi: 10.1007/BF00942151
DeVoe, J. F., & Kaffenberger, S. (2005). Student reports of bullying: Results from the 2001
school crime supplement to the national crime victimization survey (NCES 2005-310). U.S.
Department of Education, National Center for Education Statistics. Washington, DC: U.S.
Government Printing Office.
Du, H., Feldman, M. W., Li, S., & Jin, X. (2007). An algorithm for detecting community
structure of social networks based on prior knowledge and modularity. Complexity, 12,
53–60. doi: 10.1002/cplx.20166
Ennett, S. T., & Bauman, K. E. (1994). The contribution of influence and selection to adolescent
peer group homogeneity: The case of adolescent cigarette smoking. Journal of Personality
and Social Psychology, 67, 653–663. doi: 10.1037/0022-3514.67.4.653
Epstein, D. (2001). Boyz’ own stories: Masculinities and sexualities in schools. In W. Martino,
& B. Meyenn (Eds.), What about the boys? Issues of masculinity in schools (pp. 96–109).
Philadelphia, PA: Open University Press.
Espelage, D. L., & Holt, M. K. (2001). Bullying and victimization during early adolescence:
Peer influences and psychosocial correlates. In R. Geffner, & M. Loring (Eds.), Bullying
behaviors: Current issues, research, and interventions (pp. 123–142). Binghampton, NY: The
Haworth Press. doi: 10.1111/1467-8624.00531
Espelage, D. L., Holt, M. K., & Henkel, R. R. (2003). Examination of peer-group contextual
effects on aggression during early adolescence. Child Development, 74, 205–220. doi:
10.1111/1467-8624.00531
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
204
Michelle Birkett and Dorothy L. Espelage
Espelage, D. L., Green, H., Jr., & Wasserman, S. (2007). Statistical analysis of friendship
patterns and bullying behaviors among youth. In L. Hanish, & P. Rodkin (Eds.), Peer social
networks, new directions for child and adolescent development (Vol. 118, pp. 61–75). San
Francisco, CA: Jossey-Bass. doi: 10.1002/cd.201
Galambos, N. L., Petersen, A. C., Tobin-Richards, M. H., & Gitelson, I. B. (1985). The Attitudes
Toward Women Scale for Adolescents (ATWSA): A study of reliability and validity. Sex
Roles, 13, 343–354. doi: 10.1007/BF00288090
Gavin, L. A., & Furman, W. (1989) Age differences in adolescents’ perceptions of their peer
groups. Developmental Psychology, 25, 827–834. doi: 10.1037/0012-1649.25.5.827
Gest, S. D., Moody, J., & Rulison, K. L. (2007). Density or distinction? The roles of data
structure and group detection methods in describing adolescent peer groups. Journal of
Social Structure, 8. Retrieved from http://www.cmu.edu/joss/content/articles/volume8/
GestMoody
Girvan, M., & Newman, M. E. J. (2002). Community structure in social and biological networks. Proceedings of the National Academy of Sciences, 99, 7821–7826. doi: 10.1073/
pnas.122653799
Hershberger, S., & D’Augelli, A. R. (1995). The impact of victimization on the mental health
and suicidality of lesbian, gay, and bisexual youth. Developmental Psychology, 31, 65–74.
doi: 10.1037/0012-1649.31.1.65
Holland, P. W., & Leinhardt, S. (1973). The structural implications of measurement error in
sociometry. Journal of Mathematical Sociology, 3, 85–111.
Ibarra, H. (1992). Homophily and differential returns: Sex differences in network structure and
access in an advertising firm. Administrative Science Quarterly, 37, 422–447. doi: 10.2307/
2393451
Kandel, D. B. (1978). Homophily, selection, and socialization in adolescent friendships. American Journal of Sociology, 84, 427–436. doi: 10.1086/226792
Kimmel, M. S., & Mahler, M. (2003). Adolescent masculinity, homophobia, and violence:
Random school shootings, 1982–2001. American Behavioral Scientist, 46, 1439–1458. doi:
10.1177/0002764203046010010
Kosciw, J. G., & Diaz, E. M. (2006). The 2005 National School Climate Survey: The experiences
of lesbian, gay, bisexual and transgender youth in our nation’s schools. New York: GLSEN.
Lazarsfeld, P. F., & Merton, R. K. (1954). Friendship as a social process: A substantive and
methodological analysis. In M. Berger, T. Abel, & C. H. Page (Eds.), Freedom and control in
modern society (pp. 18–66). New York: Van Nostrand.
Levant, R., Majors, R., & Kelley, M. (1998). Masculinity ideology among young AfricanAmerican and European-American women and men in different regions of the United States.
Cultural Diversity and Mental Health, 4, 227–236. doi: 10.1037/1099-9809.4.3.227
Mandel, L., & Shakeshaft, C. (2000). Heterosexism in middle schools. In N. Lesko (Ed.),
Masculinities at school (pp. 75–103). Thousand Oaks, CA: Sage.
McKinley, J. (2010). Suicides put light on pressures of gay teenagers. The New York Times.
Retrieved from http://www.nytimes.com/2010/10/04/us/04suicide.html
McPherson, J. M., & Smith-Lovin, L. (1987). Homophily in voluntary organizations: Status
distance and the composition of face-to-face groups. American Sociological Review, 52,
370–379.
McPherson, J. M., Smith-Lovin, L., & Cook, J. M. (2001). Birds of a feather:
Homophily in social networks. Annual Review of Sociology, 27, 415–444. doi: 10.1146/
annurev.soc.27.1.415
Mollica, K. A., Gray, B., & Trevino, L. K. (2003). Racial homophily in newcomers’ social
networks in a culturally diverse setting. Organization Science, 14, 123–136. doi: 10.1287/
orsc.14.2.123.14994
Neal, J. W. (2008). Kracking the missing data problem: Applying Krackhardt’s cognitive social
structures to school-based social network research. Sociology of Education, 81, 140–162. doi:
10.1177/003804070808100202
Newlin, J. L. (2009). Investigation into the construct validity of the adolescent masculinity
ideology in relationships scale. Unpublished manuscript.
Phoenix, A., Frosh, S., & Pattman, R. (2003). Producing contradictory masculine subject
positions: Producing narratives of threat, homophobia and bullying in 11–14 year old boys.
Journal of Social Issues, 59, 179–195. doi: 10.1111/1540-4560.t01-1-00011
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015
Homophobic Name-calling and Peer Groups 205
Poteat, V. P. (2007). Peer group socialization of homophobic attitudes and behavior during
adolescence. Child Development, 78, 1830–1842. doi: 10.1111/j.1467-8624.2007.01101.x
Poteat, V. P., & Espelage, D. L. (2005). Exploring the relation between bullying and homophobic
verbal content: The Homophobic Content Agent Target (HCAT) Scale. Violence and Victims,
20, 513–528. doi: 10.1891/vivi.2005.20.5.513
Poteat, V. P., Espelage, D. L., & Green, H. D., Jr. (2007). The socialization of dominance: Peer
group contextual effects on homophobic and dominance attitudes. Journal of Personality and
Social Psychology, 92, 1040–1050. doi: 10.1037/0022-3514.92.6.1040
Raudenbush, S. W., & Bryk, A. (2002). S, Hierarchical linear models. Newbury Park, CA: Sage.
Rivers, I. (2004). Recollections of bullying at school and their long-term implications for
lesbians, gay men, and bisexuals. Crisis, 25, 169–175.
Rubin, K. H., Bukowski, W., & Parker, J. (2006). Peer interactions, relationships, and groups. In
N. Eisenberg (Ed.), Handbook of child psychology: Social, emotional, and personality development (6th ed., pp. 571–645). New York: Wiley.
Ryan, A. M. (2001). The peer group as a context for the development of young adolescent
motivation and achievement. Child Development, 72, 1135–1150. doi: 10.1111/14678624.00338
Snell, W. E., Jr. (1989). Development and validation of the Masculine Behavior Scale: A
measure of behaviors stereotypically attributed to males vs. females. Sex Roles, 21, 749–767.
doi: 10.1007/BF00289806
Steinberg, L., & Monahan, K. C. (2007) Age differences in resistance to peer influence.
Developmental Psychology, 43, 1531–1543. doi: 10.1037/0012-1649.43.6.1531
Whitley, B. E., Jr. (2001) Gender-role variables and attitudes toward homosexuality. Sex Roles,
45, 691–721. doi: 10.1023/A:1015640318045
© 2014 John Wiley & Sons Ltd
Social Development, 24, 1, 2015