This study explored how peer relationships impact the human stress response in gay males. It found that:
1) Gay male friends engaged in higher levels of co-rumination (extensive negative problem discussion) when discussing problems compared to a control group.
2) However, gay male friends did not experience rises in the stress hormone cortisol in response to co-rumination.
3) Co-rumination in gay male friendships was related to increased negative affect.
Gender issues in psychology: Interaction Styles in Childhood and Adulthood
How Peer Relationships Impact Stress in Gay Males
1. Effect of Peer 1
Running head: EFFECT OF PEER RELATIONSHIPS
The Effect of Peer Relationships on the Human Stress Response in Gay Males: An Exploratory
Study
Ashleigh Vogle
2. Effect of Peer 2
Abstract
This research addressed the construct of co-rumination in the gay population. Co-
rumination is extensive problem discussion and focusing on negative emotions (Rose, 2002).
Past research looking at same-sex female friendships has indicated that co-rumination is related
to having close relationships as well as being related to an increase in the stress hormone,
cortisol, which can elevate depression and anxiety (Byrd-Craven et al., in press). In addition, past
research has suggested that some gay men may have social styles that are similar to female
friendships (Dorfman et. al, 1995; Schneider and Witherspoon, 2000). This study involved 8
pairs of gay friends (N = 16) and utilized an experimental manipulation that elicited co-
rumination and was representative of the participant’s everyday response to stress. Participant
pairs were randomly assigned to a problem talk condition or a control condition. I found that gay
men engaged in higher levels of co-rumination when they were in a problem-talk condition
compared to those in a control condition. However, gay men did not experience rises in cortisol
in response to co-rumination. Finally, co-rumination in gay men was related to negative affect.
3. Effect of Peer 3
The Effect of Peer Relationships on the Human Stress Response in Gay Males: An Exploratory
Study
Friends play an integral part in our lives. Our friends give us a sense of belongingness,
support us, and provide a sounding board for us. We always have someone that we can count on
even if it’s just to hang out and watch television together, due to friends. Although overall
friends are very beneficial, sometimes there are disadvantages to having friends. Friends can
have a negative impact and even transfer distress to us. However, not all friendships are alike.
Some friends provide social support, and some friends transfer a lot of stress, whereas others
transfer hardly any stress at all.
The trade-offs in friendships are even more pronounced when we look at differences in
female and male friendships. On average, men tend to engage in larger groups of friends and
form bonds by participating in shared activities. Many men enjoy direct competition and prefer
to do activities with friends such as playing a game together or engaging in a physical activity.
Women, on the other hand, tend to, on average, engage in smaller groups of dyads or triads of
friends and form bonds through telling each other secrets and their feelings, self-disclosure.
Women, on average, do not enjoy the direct competition that males do. Women prefer less
competitive activities. Going to a coffee shop and talking for an hour with a close friend would
be something that was more characteristic of a female friendship (Byrd-Craven & Geary, 2007)
Due to the different characteristics of male and female friendships, men and women
emphasize different things within their friendships. What different characteristics are male and
female friendships emphasizing, and how do these different friendship characteristics affect men
and women’s’ well-being? Questions such as this are important. However, most previous
research concerning friendship has looked at heterosexual men and women. Although
4. Effect of Peer 4
heterosexual male and heterosexual female friendships are the most typical, gay men and lesbian
women’s friendships are important as well, and may have important features that impact overall
well-being. This paper will attempt to address differences of female and male friendships, as well
as address gay men’s friendships and how these friendships affect them.
Carbery and Buhrmester (1998) examined the different roles of friends through three
phases of young adulthood. They found that women reported receiving more affection, intimacy,
and guidance/advice from friends than men did. Also, they found that married women remain
more dependent on their friends than married men, despite the shift of focus to their spouse. This
research suggests that women are more dependent on their friends than men. In a two-part study,
Grabill and Kerns (2000) looked at the relationship between attachment styles and intimacy in
friendship. They studied three intimacy characteristics. The three intimacy characteristics were
the following: feeling validated, understood, and cared for by a partner throughout conversations;
self-disclosure; and frequency of response to a partner’s disclosure. In the first study, they found
that women reported more self-disclosure, more responsiveness, and felt responded to at a
greater degree by others significantly more than men. In the second study, they found female
friends engaged in and reported more intimacy than male friends in either instance. Grabill and
Kerns (2000) speculate that their results suggest that individuals who are seeking out intimacy
with others try to fill this void by self-disclosing.
So far, research examining males and females friendships has been documented.
However, the topic of gay men’s friendships has not been addressed. It has been established that
men and women’s friendships are quite different, but where do gay men’s friendships fall? Are
gay men’s friendships a reflection of typical female or male friendships, or do they fall into a
category all of their own? This topic is of particular interest because so little research examining
5. Effect of Peer 5
the characteristics of gay men’s friendships is available. Also, little if any research is available
examining how gay men’s friendships affect their well-being.
Galupo (2007) examined friendship patterns of gay, lesbian, and bisexual individuals.
Galupo found that gay and bisexual men reported having more cross-sex friendships than lesbian
and bisexual women. An exploratory study by Schneider and Witherspoon (2000) examined
friendships of gay, lesbian, and heterosexual male and female youth. Schneider and Witherspoon
found that 50 percent of gay men’s friends are women on average. Additionally, they found that
68 percent of their friends are on average heterosexual. Finally, they found that heterosexual
female youth and gay men talk with friends more often than either heterosexual male youth or
lesbians. This research could suggest that gay men tend to gravitate toward friendships with
heterosexual women. Furthermore, the research suggests that gay men, on average, may have
friendships that are similar to friendships among heterosexual girls and women.
Dorfman et. al (1995) examined the varying types of social support for homosexual and
heterosexual men and women. They found that although older heterosexual men and gay men
possess the same levels of social support, their sources of social support were different.
Heterosexual men relied on support from family for social support; whereas, gay men relied on
friends for their social support. Although this research was done on older gay men, it is important
because it is evidence that gay men are relying on social support from friends more heavily then
heterosexual males.
The general characteristics of heterosexual female friendships have been examined;
however, how are these friendships impacting the well-being of women? Uchino et. al (1999)
suggests that friendships have a positive effect on individuals. They found that individuals who
received high levels of social support had lower blood pressure than individuals who received
6. Effect of Peer 6
lower levels of social support. Also, the studies they reviewed suggested that social support is
related to a stronger immune response. Additionally, many of the studies found that individuals
who possessed high levels of social support also had more robust natural killer responses
compared to individuals who had lower levels of social support systems. Finally, they
acknowledge that stress, among other environmental factors, can lead to the release of cortisol.
Chronic cortisol release can then dampen the immune system. Although social support can be
related to an improvement in our health, can social support from friends be harmful?
Stroud et. al (2002) examined adrenocortical responses in men and women in response to
social rejection stressors and achievement stressors. The social rejection stressor involved a
gradual rejection and exclusion by confederates in the experiment, and the achievement stressor
involved verbal and math challenges. They found that men’s cortisol levels rose significantly
more than women’s in response to the achievement stressors. They also found that women’s
cortisol levels rose significantly more than men’s in response to the social rejection stressors.
Stroud et. al’s research, “suggests that women may not only use interpersonal strategies to cope
with stress but also may show greater physiological responses to interpersonal events” (p. 323).
Rose (2002) looked at a specific aspect of friendships, co-rumination. “Co-rumination
refers to excessively discussing personal problems with a dyadic relationship and is characterized
by frequently discussing problems, discussing the same problem repeatedly, mutal
encouragement of discussing problems, speculating about problems, and focusing on negative
feelings” (p. 1830). Rose examined the levels and the impact of co-rumination in boys’ and girls’
friendship interactions. She found that girls engaged in co-rumination significantly more than
boys did. Also, she found that co-rumination was related to increased friendship quality as well
as an increase in emotional maladjustment.
7. Effect of Peer 7
Rose et. al (2007) examined the relationship between emotional maladjustment and
friendship quality as it relates to co-rumination in girls and boys. They found that co-rumination
predicted feelings of positive friendship quality as well as an increase in feelings of closeness.
However, in girls, co-rumination also predicted increases in anxiety and depression. Co-
rumination did not predict an increase in anxiety and depression in boys. Finally, Byrd-Craven
et. al (2008) examined the relationship between co-rumination and cortisol in same-sex female
friendships. Same-sex friend dyads were randomly assigned to a problem-talk condition in which
the friends discussed problems or to a control condition where they designed a recreation center.
They found that friends in the problem-talk condition engaged in higher levels of co-rumination
than friends in the control condition. Also, they found that co-rumination predicted increases in
cortisol. Co-rumination is an important construct to examine within friendships because although
there are positive characteristics of friendships such as feeling close to your friend, there are
trade-offs as well. Co-rumination is related to an increase in internalizing symptoms, such as
anxiety and depression as well as increases in cortisol levels in women.
How exactly could physiological responses, such as an increase in cortisol, affect
individuals? In a meta-analysis by Dickerson and Kemeny (2004) they write:
Prolonged cortisol activation (produced by frequent exposure to stressors or by failing to
shut down thus response after stressor termination) is associated with a number of
negative biological and health effects, including suppression of aspects of the immune
system (e.g., decreased lymphocyte proliferation and cytokine production); damage to
hippocampal neurons; and the development and/or progression of certain chronic diseases
such as diabetes and hypertension (p. 356).
8. Effect of Peer 8
If co-rumination is shown to increase levels of cortisol while friends are talking about their
problems in a lab setting, then this could suggest that each time when friends are co-ruminating
outside of a lab their cortisol levels are raising. If friends are co-ruminating frequently, then they
may be suffering from prolonged cortisol activation, which could have serious effects on health
outcomes. Additionally, co-rumination has been related to anxiety and depression in women,
which shows that it impacts psychological health as well.
Previous research has suggested that gay men talk to their friends at higher levels, which
are more similar to heterosexual female youth, than lesbians or heterosexual male youth
(Schneider and Witherspoon, 2000). Also, older gay men tend to rely more on friends for social
support than heterosexual men. It is possible that gay men have similar friendship styles
compared to heterosexual females. In this exploratory study, I will be examining the following
hypotheses 1) gay men will engage in higher levels of co-rumination when they are in a problem-
talk condition compared to those in a control condition; 2) gay men will experience rises in
cortisol in response to co-rumination 3) co-rumination in gay men will be related to negative
affect.
Method
Participants
Participants involved in the study attend a large coeducational research university in the
Midwest. The age of the participants was estimated to be between 18 and 23 years old.
Participants were recruited from the psychology participant pool, fliers, and the university’s
Lesbian, Gay, Bisexual and Transgender (LGBT) organization. There were 16 participants in the
study (8 friendship dyads). Friendship dyads were composed of gay men who indicated that they
9. Effect of Peer 9
were close, platonic friends. Participants received course credit or were entered into a raffle to
win $50.
Procedure
Participants completed a Problem Generation questionnaire and the Co-rumination
questionnaire. Participants completed the experiment in approximately 60 minutes. Participants
were randomly assigned to the Problem Talk condition (6 dyads) or the Control condition (2
dyads). Saliva samples were collected from participants after completion of the informed
consent. Participants then completed the Problem Generation questionnaire, and afterwards they
were video recorded for five minutes in a warm-up task of planning a menu. Dyads were not
informed which condition they would be assigned to until after their pre-task cortisol measure
was taken in order to avoid group effects of task anticipation. Participants in the Problem Talk
and Control conditions were video recorded during the seventeen-minute discussion periods.
Participants assigned to the Problem Talk condition each selected one problem from the Problem
Generation questionnaire to discuss. Participants in the Problem Talk condition were asked to
talk about either their problem, their friend’s problem, or both problems in a similar manner they
would talk about their problems together in everyday life. Participants assigned to the Control
condition were asked to design an amusement park. The amusement park was selected as the
activity for the Control condition because it is a neutral cooperative activity that involves no
social dimension. Through this activity, we were able to determine if fluctuations in cortisol were
primarily due to social and inter-personal nature of tasks. After participants completed the
discussion periods they were separated into two areas of the room, given instructions not interact
with each other. Saliva samples were collected fifteen minutes after the task. Finally, participants
completed the Co-rumination questionnaire.
10. Effect of Peer 10
Saliva samples were collected immediately after completion of the informed consent and
15-20 minutes post-task. Saliva samples were collected 15-20 minutes after the task because at
this time cortisol levels are likely at their post-stressor peak. During the time that dyads were
separated into the two areas of the room to prevent further interaction, they looked at home,
garden, travel, furniture, or architecture magazines for 15 minutes.
Determination of Salivary Analytes
Participants were instructed to avoid eating a large meal, caffeine, and nicotine an hour
before they came to the laboratory. These instructions were given to avoid potential confounding
features of HPA responses. Saliva was collected by participants holding 1 x 4 CM absorbent
swabs in their mouths for 1 to 2 minutes. Saliva saturated swabs were stored at -20°C until
shipped overnight on dry-ice to Penn State University. Samples were assayed for cortisol
(enzyme immunoassay) using commercially available regents (Salimetrics, State College, PA)
without modification to the manufacturers recommended protocols. Cortisol levels area reported
in micrograms per deciliter (ug/dL). Cortisol assays have average intra- and inter-assay
coefficients of variations less than 15 percent.
Measures
Problem Generation and Salience questionnaire (Rose et al., 2005)
Participants generated three current problems and wrote short descriptions of them. They
also rated these problems on varying dimensions such as how much the problem bothered them,
how in control they felt in regard to their problem, and if it would be easy to solve the problem.
All of these dimensions were rated on a 5 point Likert scale. This questionnaire was included as
a tool to assess factors that would influence participants’ pre-task cortisol levels.
11. Effect of Peer 11
Co-rumination questionnaire (Rose, 2002)
Participants rated 27 items assessing co-rumination with same-sex friends on a 5-point
scale. Items assessed focusing on negative affect, rehashing problems, mutual encouragement of
problems, discussing problems extensively, and speculating about problems. Items measured a
more extreme form of problem discussion compared to items typically used to measure
normative self-disclosure (e.g., “When we talk about a problem that one of us has, we usually
talk about that problem every day even if nothing new has happened”).
Coding
The system of coding co-rumination of Rose et al. (2005) was adapted for this study.
Both conditions of the seventeen minute interactions were coded on four dimensions of co-
rumination: rehashing problems, speculating about causes and consequences of problems,
focusing on negative affect, and mutual encouragement of problem talk. Coders rated each dyad
on a 5-point Likert scale to the degree in which the dyad was characterized by the dimension.
Two or three coders rated each interaction. For each dimension, inter-rater reliability was high
(range = .80 to .91). The sum of the scores across the four dimensions was the equivalent of the
total observed co-rumination scores.
Results
Analytic Strategy
Cortisol scores were positively skewed. A natural log transformation was used in all
analyses to normalize distributions (Gordis et al., 2006).
For this study, I tested three hypotheses. First, I predicted that gay men would engage in
higher levels of co-rumination when they are in a problem-talk condition compared to those in a
control condition. This hypothesis was supported. A one-way ANOVA revealed that gay men
12. Effect of Peer 12
engaged in significantly higher levels of co-rumination when they were in a problem-talk
condition compared to those in a control condition, F(1,14) = 118.243, p<.001. Second, I
predicted that gay men would experience rises in cortisol in response to co-rumination. A
regression analysis was conducted to determine the effect of co-rumination on post-task cortisol
levels. Post-task cortisol levels were predicted from observed co-rumination. Pre-task cortisol
was used as a covariate. The effect of observed co-rumination was not significant, β = .048, t(1,
15) = .370, p = .717. Thus, this hypothesis was not supported. Finally, I predicted that co-
rumination in gay men will be related to negative affect. A regression analysis was conducted to
determine the effect of co-rumination on negative affect. The effect of co-rumination was
significant, β = .841, t(1, 15) =5.814, p<.001. Please refer to Appendix 1 for means and standard
deviations of all measures.
Discussion
The purpose of this study was to investigate the frequency of co-rumination in gay men
and if increasing levels of co-rumination is related to increasing levels of cortisol. Gay men who
talked about their problems engaged in significantly higher levels of co-rumination than gay men
who did not discuss their problems. There was no significant effect of co-rumination on cortisol
levels. We did not find increased levels of cortisol with increased levels of co-rumination. Co-
rumination was shown to be related to negative affect. In Byrd-Craven et. al (2008) a study with
heterosexual women the mean in the Problem Talk condition was 13.68 (SD = 2.56) and the
mean in the Control condition was 7.65 (SD = 2.77) Additionally, the finding that gay males
engage in co-rumination is consistent with other previous research. Schneider and Witherspoon’s
(2000) research suggests that gay men, on average, may have friendships that are similar to
friendships among heterosexual female youth. They found that gay men talk with friends at
13. Effect of Peer 13
frequencies that are similar to heterosexual female youth, which is significantly more than
lesbians and heterosexual male youth. Additionally, Dorfman et. al, (1995) researched the social
support systems of gay men and found that gay men depend on social support from their friends
significantly more than heterosexual men. Finally, Grabill and Kerns’ (2000) research found that
women seek out intimacy with others through self-disclosure. All of this research supports the
idea that gay men are engaging in co-rumination. Gay men and heterosexual women are both
having more conversations with their friends than lesbians or heterosexual men. Also, gay men,
similar to heterosexual women depend on their friends for social support. Heterosexual women
tend to self-disclose to their friends in an attempt to increase intimacy in the relationship, and
this study could suggest that gay men are perhaps doing the same thing. Possibly they are
disclosing information to their friends in an attempt to increase intimacy as well. It would appear
that there is a considerable degree of overlap in the friendship styles of gay men and heterosexual
women.
Although it appears that gay men have some similar friendship characteristics as
heterosexual women, their cortisol levels are not increasing in a response to the co-rumination.
This is interesting because past research found that increased levels of co-rumination of
heterosexual females led to increases in cortisol (Byrd-Craven et al., 2008; Byrd-Craven et al., in
press). Perhaps gay men did not have responses in cortisol because co-rumination isn’t stressful
to them. Stroud et. al’s (2002) research found that men’s cortisol levels rose significantly more
than women’s cortisol levels when they experienced achievement stressors, and women’s
cortisol levels rose significantly more than men’s cortisol levels when they experienced social
rejection stressors. Maybe gay men didn’t experience a rise in cortisol levels due to something
because of their biological make-up.
14. Effect of Peer 14
Another explanation why there wasn’t a significant increase in cortisol in response to co-
rumination could be due to the small sample size of this study. Perhaps if there were more
participants in this study a better representation of this population would have been accounted for
and an increase in cortisol could be apparent in response to co-rumination. One final reason why
there wasn’t a cortisol increase in response to co-rumination could be due to cortisol levels that
are already raised when participants arrived at the experiment. Participants could have come into
the laboratory with anticipatory stress, due to the nature of the experiment. Researching a special
population in a conservative city most likely put some stress on participants. Also, since this
study was conducted in one session participants were not able to get acclimated to the laboratory.
The mean for the pre-task cortisol was -1.67 and the mean for the post-task cortisol was -1.81
which doesn’t appear to be that great of a difference. Possibly their pre-task cortisol is a
reflection of an already elevated stress response, thus a further increase in cortisol is not possible
or not as likely.
Finally, the finding that co-rumination is related to negative affect is consistent with
previous research on heterosexual girls and women (Rose et al., 2007; Byrd-Craven et al., 2008;
Byrd-Craven et al., in press). This has been posited as a potential link between engaging in co-
rumination and depression and anxiety symptoms (Byrd-Craven et al., in press). This finding
suggests the co-rumination may have costs for gay men as well, and that there is some transfer of
distress. Gay men may be more relationship-focused than heterosexual men, and in addition to
societal issues they face (e.g., discrimination), may be more likely to show internalizing
symptoms. Excessively discussing these problems with a close friend may, ironically, do more
harm than good.
15. Effect of Peer 15
As limitations are present in all research, this study is no exception. One of the biggest
limitations in this study is the small sample size. Doing research which involves special
populations is definitely challenging, and this challenge was present in this study as well. Future
research of this area could include a larger sample size to increase the power of the study. By
increasing the sample size this would allow future researchers to determine if cortisol levels are
increasing in response to increased levels of co-rumination or if they weren’t. Another limitation
of this study includes the session style. If participants were getting anxious because of the
laboratory setting and because of the nature of the study (research involving gay men in a very
conservative city) then this confound could be ameliorated by breaking up the experiment into
two sessions. The first session could be devoted to filling out questionnaires, and the second
session could involve participants being video recorded in the menu planning warm up task as
well as being video recorded in either the problem talk or the control condition. This would
allow participants to become acquainted and feel comfortable in the lab setting. Then, more
accurate cortisol measures could be ensured.
Additionally, future research could examine friendships of gay men and heterosexual
women. Schneider and Witherspoon’s (2000) research revealed that gay men’s friendships were
composed of 68 percent heterosexual women. It would be very interesting to look at the
friendships of gay men and heterosexual women and see if co-rumination is present in their
friendships in addition to examining if co-rumination increased cortisol levels.
16. Effect of Peer 16
References
Byrd-Craven, J., & Geary, D. C. (2007). Biological and evolutionary contributions to
developmental sex differences. Reproductive BioMedicine, 15, 10-20.
Byrd-Craven, J., Geary, D. C., Rose, A. J., & Ponzi, D. (2008). Co-ruminating increases stress
hormones in women. Hormones and Behavior, 53, 489-492.
Byrd-Craven, J., Granger, D.A., & Auer, B. J. (in press). Stress reactivity to co-rumination in
young women’s friendships: The relationship between cortisol and alpha-amylase.
Journal of Social and Personal Relationships.
Carbery, J., & Buhrmester, D. (1998). Friendship and need fulfillment during three phases of
young adulthood. Journal of Social and Personal Relationships, 15(3), 393-409.
Dickerson, S. S., & Kemeny, M. E. (2004). Acute stressors and cortisol responses: A theoretical
integration and synthesis of laboratory research. Psychological Bulletin, 130(4), 355-391.
Dorfman, R., Walters, K., Burke, P., Hardin, L., & Karanik, T. (1995). Old, sad and alone: The
myth of the aging homosexual. Journal of Gerontological Social Work, 24(1/2), 29-44.
Galupo, M. P. (2007). Friendship patterns of sexual minority individuals in adulthood. Journal of
Social and Personal Relationships, 24(1), 139-151.
Grabill, C. M., & Kerns, K. A. (2000). Attachment style and intimacy in friendship. Personal
Relationships, 7, 363-378.
Rose, A. J. (2002). Co-rumination in the friendships of girls and boys. Child Development, 73(6),
1830-1843.
Rose, A. J., Carlson, W., & Waller, E. M. (2007). Prospective associations of co-rumination with
friendship and emotional adjustment: Considering the socioemotional trade-offs of co-
rumination. Developmental Psychology, 43(4), 1019-1031.
17. Effect of Peer 17
Schneider, M. S. & Witherspoon, J. J. (2000). Friendship patterns among lesbian and gay youth:
An exploratory study. The Canadian Journal of Human Sexuality, 9(4), 239-246.
Stroud, L.A., Salovey, P., & Epel, E. S. (2002). Sex differences in stress response: Social
rejection versus achievement stress. Society of Biological Psychiatry, 52, 318-327.
Uchino, B. N., Uno, D., & Holt-Lunstad, J. (1999). Social support, physiological processes, and
health. Current Directions in Psychological Science, 8(5), 145-148.
18. Effect of Peer 18
Appendix 1
Measures Group Mean Standard Deviation
Observed co-rumination Problem Talk condition 12.58 1.49
Observed co-rumination Control condition 4.25 .29
Post-task cortisol Problem Talk condition -1.74 .78
Post-task cortisol Control condition -2.02 .29
Observed negative affect Problem Talk condition 2.83 .39
Observed negative affect Control condition 1.00 .01