Effects of Context and Relative Rank on Mate Choice and Affiliation Ratings

Evolutionary Psychology
www.epjournal.net – 2009. 7(3): 449-462
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Original Article
Effects of Context and Relative Rank on Mate Choice and Affiliation Ratings
P. Lynne Honey, Department of Psychology, Grant MacEwan College, Edmonton, Canada. Email:
HoneyL@MacEwan.ca (Corresponding author).
Chad D. Coulombe, Department of Psychology, Grant MacEwan College, Edmonton, Canada.
Abstract: Female dominance has not often been studied as a factor in mate choice and
other social interactions. When it has been examined, there have been a number of
conflicting findings. The present study was designed to clarify interpretations of a study
conducted by Brown and Lewis (2004) that found that men prefer subordinate women in a
workplace context. We presented participants with information about the relative rank of
physically attractive targets, in two very different contexts (work-related and recreational).
We found that the context in which rank cues are presented has an impact on affiliation
ratings, but that cues of rank do not affect mate choice ratings. Future studies of effects of
dominance must take into account the context in which they are presented, and recognize
that rank may not be a sufficient indicator of dominance for the purpose of mate choice by
both men and women.
Keywords: dominance, mate choice, attraction, sex differences
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Introduction
Social dominance in humans results when individuals within a group vary in their
ability to control resources and achieve status or prestige (Hawley, 1999). There are a
number of traits that are correlated with or predict social dominance, including physical
size, aggression, masculinity, and rank, but also persuasion, cooperation, and friendship.
The roles of these various interactive traits have not been fully elucidated, that is there is no
obvious evidence that any one of these traits is a necessary component of social dominance,
and it is likely that there are varying ways in which socially dominant traits or attributes
may combine to allow an individual to compete with others in the group for access to
valued items or opportunities.
Academic study of dominance and dominant behavior has focused primarily on
dominance by men. There is good reason for researchers to focus on the expression and
perception of dominance in men. Across many species, men tend to be larger and more

---

Context, Rank, and Mate Choice
aggressive than women (see Alexander, Hoogland, Howard, Noonan, and Sherman, 1979
for a review). Competition among men can be fierce and has direct fitness consequences,
leading to sexual selection for traits associated with defeating other men due to access
gained to women as well as direct female choice (Darwin, 1871). Within humans, where
moderate sexual dimorphism exists for traits like body size (e.g., Gustafsson and
Lindenfors, 2009) and physical aggression (see Campbell, 1999 for a review), we find
significant effects of dominance in how men are perceived by women, and for reproductive
success (Egan and Angus, 2004; Hopcroft, 2006; Perusse, 1993). When it comes to rank,
famous examples of exceptionally dominant men who monopolize the reproduction of
many women seem to point to such dominance as an adaptive male characteristic. Betzig
(1986) describes the “Genghis Khan effect,” which seems appropriate given that Khan and
his male relatives are theorized to be the ancestors of approximately 16 million modern
humans (Zerjal et al., 2003). Additionally, nearly 20% of men in northwestern Ireland are
likely to have descended from a single warlord from the 5th century (Moore, McEvoy,
Cape, Simms, and Bradley, 2006) which further emphasizes the potential effects that male
rank can have on reproductive success.
Interest in male dominance, however, does not explain why female dominance has
been relatively ignored. Women vary a great deal in traits like aggression, assertiveness and
other personality measures associated with dominance (see Hyde, 2005). Certainly there
are differences in rank among women, where some women have greater access to resources
than others or have authority over others. In an extreme example, it has been demonstrated
that adolescent women in an institutional setting establish a dominance hierarchy through
aggressive and sometimes violent behavior but the highest ranked women tended to neither
aggress nor experience aggression from others. They also claimed more “territory” within
the institution and had more opportunities to solicit attention from men (Deutsch, Esser,
and Sossin, 1978). Although women in a less-structured environment may demonstrate
different forms of leadership, or become leaders for different types of tasks compared to
men, women do take leadership roles and demonstrate dominance (Eagly and Karau,
1991).What fitness consequences might be associated with such differences among
women?
In a comprehensive review, Ellis (1995) outlined a variety of ways in which
dominance may be linked to reproductive success among women in non-human mammal
species. Some obvious measures of reproductive success include number of offspring born,
and number that survive infancy. Other, more indirect correlates of reproductive success or
potential include age of onset of sexual activity, reproductive suppression (or harassment)
of other women, number of partners or copulations, and length of reproductive lifespan.
Among non-primates (rodents in particular), Ellis (1995) documented that female
dominance is strongly associated with reproductive success, and among some primates
there appears to also be a predictive relationship between rank and reproductive success.
Interestingly, reproductive success measured by number of offspring who survive is more
likely to be tied to mothers’ rank than is absolute number of offspring born (as in Pusey,
Williams, and Goodall, 1997). In a recent study of effects of maternal rank among baboons,
sons of high-ranking mothers had lower levels of circulating stress hormones than did sons
of lower-ranking mothers (Onyango, Gesquiere, Wango, Alberts, and Altmann, 2008). It
should be noted that the mothers rank at the time of conception was key, rather than her
rank at the time of measurement. In fact, the men in this study were 4-6 years past the point
Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -450-

---

Context, Rank, and Mate Choice
of maternal care, and many had outlived their mothers. Further evidence of such rank
effects on offspring is seen among hyenas, where dominant women’ high androgen levels
in late gestation predict dominant behaviors among their offspring, compared to the
offspring of lower-ranked women who also had lower gestational androgen levels (Dloniak,
French, and Holekamp, 2006). Although it is not the case that female dominance is
consistently predictive of reproductive success, neither is male dominance (Ellis, 1995).
Based on Ellis’ review, it does appear, however, to be one important factor among non-
human mammals.
Results among humans are not entirely clear. Some studies on effects of female
dominance for mate choice indicate that female dominance has no effect on male choice
(Fletcher, Tither, O’Loughlin, Friesen, and Overall, 2004; Sadalla, Kenrick, and Versure,
1987). When cues of dominance are portrayed as “masculine” traits, then women who
possess those traits are perceived as less attractive (Keisling and Gynther, 1993). Buss
(1981) has demonstrated, however, that men and women express dominance in different
ways. According to Buss, men are more likely to express dominance in an egoistic or
agentic manner, where personal goal attainment is enhanced. Femen are more likely to
express dominance in a more prosocial or communal manner, where both personal and
group interests are furthered. Additionally, men seem to prefer or respect agentic
expressions of dominance, and women seem to respect prosocial expressions. This is
consistent with a meta-analysis (Eagly and Karau, 1991) that reviewed gender differences
in the emergence of leadership. Thus, conceptualizing female dominance in masculine
terms may be counterproductive to understanding effects of female dominance. To add
further texture to the topic, it has been demonstrated that female dominance may interact
with physical attractiveness, such that dominance may play a greater role in reproductive
success when a female is not highly attractive (Gutierres, Kenrick, and Partch, 1999;
Kenrick, Neuberg, Zierk, and Krones, 1994). Once again, when we examine the literature
on male dominance, cues about male dominance may not always be preferred by women.
For example, we find that men perceived as dominant via cues of facial hypermasculinity
seem to be less preferred for long-term relationships (Little, Jones, Penton-Voak, Burt, and
Perrett, 2002) which may be due to attributions about character and parenting potential
(e.g., Perrett et al., 1998). Such attributions are not surprising, given that such men describe
themselves as investing less in parenting and relationship maintenance (Waynforth, 2002).
If female rank has any influence on reproductive success, its effects may differ
across individuals or environmental conditions. In a study that determined that men
sometimes attend to female dominance cues, Brown and Lewis (2004) presented
photographs and descriptive paragraphs to participants, and asked them to evaluate and rate
the targets on a variety of measures. The authors found that men were more likely to find a
subordinate female attractive for long-term relationships. Interestingly, they found no effect
of male rank on female choice, nor did rank appear to impact affiliation ratings (willingness
to socialize with the target) by same- or opposite-sex participants. There are some potential
confounds to the interpretation of these data.
Primarily, the context in which Brown and Lewis’ (2004) descriptions were framed
may have influenced responding. Paragraphs were written to describe individuals in one of
three levels of dominance as indicated by rank, and those ranks were relative to the subject
in the experiment. For example, the subject was to imagine that the target was his or her
assistant, coworker, or supervisor, depending on the experimental condition. Placing these
Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -451-

---

Context, Rank, and Mate Choice
descriptors in a workplace setting may invoke power dynamics that will alter mate choice
decisions in a manner that is not reflective of attraction, per se. Although the authors were
careful to indicate that there were no rules in this particular workplace against employee
fraternization, participants may still have been more cautious about such relationships. A
second potential problem relates to the photographs used in the study. The authors were
careful to use one male and one female photograph that had been selected by independent
raters to be roughly equal in both age and attractiveness, but no information was provided
as to how attractive or how old those photos were perceived to be. Thus, it is possible that
participants were using cues about age, particularly for judging the suitability of female
targets for long-term relationships. Further, other researchers have found that dominance
interacts with attractiveness for judgments of female targets such that cues of subordination
tend to be more appealing in physically attractive targets (Gutierres et al., 1999; Kenrick et
al., 1994), so effects should be interpreted accordingly. Without knowledge of the
approximate age or attractiveness of the photos, we cannot fully interpret Brown and
Lewis’ (2004) results.
In the present study, our goal was to replicate the basic methodology used by
Brown and Lewis (2004), while attempting to remove the potential confound of “workplace
romance” and use photographs that were selected for similarity, but also for cues of age and
attractiveness that would be generally appealing to our undergraduate sample. We
hypothesized that women would demonstrate a greater preference for a higher-ranked man
when that man was not perceived to be her immediate supervisor (which would differ from
Brown and Lewis’ results), and that men would be most attracted to a lower-ranked woman
(consistent with Brown and Lewis’ results). Further, we predicted that cues of rank would
have a larger impact on mate choice ratings in the recreational context than in the work-
related context.
Materials and Methods
Participants
A total of 426 undergraduates at Grant MacEwan College (a liberal arts institution
in northwestern Canada) participated in this study for course credit in an introductory
psychology course. From that sample, we used data from 137 male and 271 female
participants. All participants answered questions about their willingness to affiliate with the
target, and we asked participants to answer mate choice questions only if they were
sexually attracted to individuals who are the sex of the target. Eighteen participants
provided answers that were consistent with homosexual orientation, but here we have
reported data only from individuals who provided answers consistent with heterosexual
orientation. It is possible, however, that there are members of our sample who are bisexual,
or did not respond in a manner consistent with their actual sexual orientation.
The focus group that rated the photos (described later) was comprised of 10 male
and 15 female senior undergraduate students (M = 21.8 yrs ± 0.09 SEM) who did not
participate in the study. These students were recruited for voluntary participation from a
senior-level History of Psychology course.


Procedure
Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -452-

---

Context, Rank, and Mate Choice
Participants were randomly assigned to one of 12 experimental conditions in a 2
(Participant Sex) X 2 (Target Sex) X 2 (Work/Recreation Context) X 3 (Relative Rank)
between-subjects design. Our dependent variables were ratings of willingness to affiliate
with the target in both sexual and non-sexual situations. We also included a manipulation
check question, which is described in a later paragraph.

We presented each participant with a paragraph describing a single target individual
(see below), and a photograph. Paragraphs varied by context (work-related versus
recreational) and by relative rank (subordinate, equal or superior to the participant). The
paragraphs we presented were based on those used by Brown and Lewis (2004), but we had
the target work for a different company, or play for a different soccer team, in order to
reduce the likelihood that participants would answer based on potential repercussions of
engaging in a relationship with the target (e.g., getting fired, or asked to leave the team).

High rank/work related/man target example:
Please imagine that you work for a company, and that company has sent you to a
convention to learn about new software for your office. You have been seated next
to John, who works for a different company across town. John works in a position
that would be equivalent to that of your immediate supervisor (the person that you
report to on a daily basis, who is responsible for employee discipline and for
rewarding reliable or creative performance). John seems like a nice man, he has a
good sense of humor, and he is friendly.

Low rank/recreational/woman target example:
Please imagine that you play for a recreational soccer team, where you are a veteran
player, and you are attending a local tournament. You have just met Sarah, who is
friendly and has a good sense of humor. You’ve really enjoyed talking to her about
your most recent game, and about local restaurants. Sarah is a rookie on her team
(she’s only been playing for a few months), and has a lot of great restaurant
suggestions for you.

In the equal rank conditions, targets were described as holding the same job as the
participant, or having equivalent soccer experience as the participant.
We selected photographs in a manner similar to that used by Brown and Lewis
(2004). Thirty portrait photographs (15 male and 15 female) from a stock-photography
website (www.sxc.hu) were presented to a group of raters who evaluated the photos for age
and attractiveness. The raters were all senior students who did not participate in the study.
Based on those ratings, we selected one male and one female photograph. Both photos were
rated as attractive (Male, M = 8.4 ± 0.19 SEM; Female, M = 8.5 ± 0.24 SEM), and were
perceived as within an age range that would be attractive to an undergraduate population
(see Kenrick, Keefe, Gabrielidis, and Cornelius, 1996). The man was perceived to be
approximately 27 years old (M = 26.8 ± 0.9 SEM) and the woman was perceived to be
approximately 25 years old (M = 25.1 ± 1.1 SEM). We chose photos that appeared to be in
their mid-20s so that it would be somewhat believable that these targets could be
supervisors in a work-related setting, without being dramatically older than our typical
participant population.
Participants read the paragraphs, then viewed the photo, and were then asked to
Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -453-

---

Context, Rank, and Mate Choice
give their impressions of the targets. We used the same questions used by Brown and Lewis
(2004), so all participants rated the extent to which they would like to work for the target,
work with the target, or have the target work for them. They were then asked how much
they would enjoy exercising, or going to a party with the target. Opposite-sex participants
also rated the desirability of the target for dating, marriage or a one-time sexual encounter.
Finally, as a manipulation check, participants were asked to rate the relative dominance of
the target using the question “When you compare [the target] to yourself, how much power
or status do you think [the target] has?”. All ratings were provided by the participant
making an X along a line, between options similar to “not at all” and “very.” We then
measured the distance (mm) from “not at all” to the participant’s X. Responses could vary
between 0 and 12.4 mm (thus, 6.2 mm was the absolute midpoint of the scale).
Results
Manipulation check
For analyses of the manipulation check, we performed a 2 (Participant Sex) X 2
(Target Sex) X 2 (Work/Recreation Context) X 3 (Relative Rank) ANOVA. There was a
main effect of rank, F(2, 384) = 30.15, p < 0.0001. Further, there was an interaction of
context with rank, F(2, 384) = 13.91, p < 0.001, and a three-way interaction of context,
rank and target sex, F(2, 384) = 7.52, p = 0.001. Further analyses of these significant
interactions revealed that low-ranked men in the work-related scenario were seen as less
powerful than were women described in the same way, t = -3.16, p = 0.003, and also much
less powerful than equal-ranked men, t(60) = -6.13, p < 0.0001. When men and women
were described as equal-rank in the workplace setting, male targets were rated higher for
power, t(57) = 3.31, p = 0.002. Further, low-ranked men were rated much lower in the
workplace setting than in the soccer setting, t(69) = -4.16, p < 0.0001, but equal-ranked
men were rated more highly in the workplace setting than in the soccer setting, t( 64) =
3.06, p = 0.003. There was no significant difference in ratings for high-rank men and
women in neither the workplace nor the recreational setting, or for men and women
described as “rookies” in the recreational scenario. Thus, participants did perceive that they
would have power over targets described as subordinate, but this effect was somewhat
more pronounced in the workplace setting, and differed by sex in the workplace setting
(Figure 1a). In particular, low-ranking men in the workplace setting were seen as
particularly subordinate, but equal-ranked men were rated higher than the midpoint of the
scale. In the recreational setting, there was less differentiation by rank, although all
“coaches” got higher ratings (Figure 1b).

Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -454-

---

Context, Rank, and Mate Choice
Figure 1a. Mean ± SEM ratings of target person in the workplace context for the question
“When you compare [the target] to yourself, how much power or status do you think [the
target] has?”
12
11
10
9
8
7
Low
6
Equal
High
5
4
3
2
1
0
By Males
By Females
By Males
By Females
Male Target
Female Target

Figure 1b. Mean ± SEM ratings of target person in the recreational context for the question
“When you compare [the target] to yourself, how much power or status do you think [the
target] has?”
12
11
10
9
8
7
Low
6
Equal
High
5
4
3
2
1
0
By Males
By Females
By Males
By Females
Male Target
Female Target

Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -455-

---

Context, Rank, and Mate Choice
Mate-choice ratings
See Figure 2a for data related to long-term mate ratings and Figure 2b for data
related to short-term mate ratings. For analyses of mate choice ratings, we performed 2
(Target Sex) X 2 (Work/Recreation Context) X 3 (Relative Rank) ANOVAs. There was a
main effect of target sex, such that female targets were given higher ratings by male
participants for both long-term, F(1, 203) = 30.83, p < 0.0001, and short-term, F(1, 203) =
85.03, p < 0.0001, mating opportunities, when compared to ratings of male targets by
female participants. Beyond this expected effect, there were no main effects of
work/recreation context or relative rank. Further, there were no simple interactions among
the variables. The only significant interaction was a context X rank X target sex interaction,
F(2, 203) = 3.14, p = 0.045, on ratings of desirability for long-term relationships, such that
there appears to be a different influence of rank cues on long-term mate choice ratings in
the recreational setting, and those effects seem to differ by sex (Figure 2). Further analysis
on data from the recreational context revealed a significant dominance X target sex
interaction, F(2, 114) = 3.459, p = 0.035, and post-hoc contrasts indicate that high or equal-
dominance women were rated more highly than low-dominance women, t(44) = 2.15, p =
0.037, but this pattern did not occur for male targets.

Figure 2a. Mean ± SEM ratings of target person (by opposite-sex participants) as desirable
for a long-term dating relationship, or potentially marriage.
12
11
10
9
8
7
Low
6
Equal
5
High
4
3
2
1
0
Male
Female
Male
Female
Soccer
Office

Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -456-

---

Context, Rank, and Mate Choice
Figure 2b. Mean ± SEM ratings of target person (by opposite-sex participants) as desirable
for a short-term sexual encounter.
12
11
10
9
8
7
Low
6
Equal
5
High
4
3
2
1
0
Male
Female
Male
Female
Soccer
Office

Affiliation ratings
For analyses of affiliation ratings, we performed 2 (Participant Sex) X 2 (Target
Sex) X 2 (Work/Recreation Context) X 3 (Relative Rank) ANOVAs. There were main
effects of context on willingness to work out with the target, F(1, 384) = 35.77, p < 0.0001,
and to go to a party with the target, F(1, 384) = 6.23, p = 0.013. There were no main effects
of relative rank on either affiliation question. There was a significant main effect of target
sex on both working out, F(1, 384) = 22.78, p < 0.0001, and going to a party with the
target, F(1, 384) = 101.66, p < 0.0001, such that female targets were given higher ratings
for both activities. Participant sex had no main effect on the affiliation ratings. We found
several interactions for the affiliation question about willingness to go to a party with the
target. There was a context X rank interaction, F(2, 384) = 3.78, p = 0.024. Post-hoc
analyses revealed that participants seemed to be less willing to go to a party with
subordinates or superiors in the work-related setting, t(173) = 2.285, p = 0.023, but not in
the recreational setting. Subordinate male targets in the work setting received the lowest
ratings for going to a party, which is reflected in a context X rank X target sex interaction,
F(2, 384) = 5.09, p = 0.007. Further, a rank X participant sex interaction, F(2, 384) = 3.31,
p = 0.038, suggests that male participants are more willing to go to a party with equals or
superiors, t(73) = 2.00, p = 0.05, compared to female participants. Results for the “Gym”
question are presented in Figure 3a and results for the “Party” Question are presented in
Figure 3b.
Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -457-

---

Context, Rank, and Mate Choice
Figure 3a. Mean ± SEM ratings of the extent to which participants would enjoy working
out or playing sports with the target.
12
11
10
9
8
7
Low
6
Equal
5
High
4
3
2
1
0
By Males
By
By Males
By
By Males
By
By Males
By
Females
Females
Females
Females
Male Target
Female Target
Male Target
Female Target
Work
Recreation

Figure 3b. Mean ± SEM ratings of the extent to which participants would enjoy going to a
party with the target.
12
11
10
9
8
7
Low
6
Equal
5
High
4
3
2
1
0
By Males
By
By Males
By
By Males
By
By Males
By
Females
Females
Females
Females
Male Target
Female Target
Male Target
Female Target
Work
Recreation

Evolutionary Psychology – ISSN 1474-7049 – Volume 7(3). 2009. -458-

---

Document Outline

• ??
• ??
• ??
• ??

---