Avuncular tendencies and the evolution of male androphilia in Samoan fa'afafine.

Abstract

The kin selection hypothesis for male androphilia holds that genes for male androphilia can be maintained in a population if the fitness costs of not reproducing directly are offset by enhancing indirect fitness. Kin share some proportion of genes identical by virtue of descent. Theoretically speaking, androphilic males can increase their fitness indirectly by allocating altruistic behavior toward kin, which, in turn, allows kin to increase their reproductive success. Research conducted in Independent Samoa has shown that androphilic males (known locally as fa'afafine) report significantly higher avuncular tendencies relative to gynephilic men. Here, we replicate this sexual orientation difference, using a larger, independent sample, suggesting that the documented sexual orientation difference in avuncular tendencies in Independent Samoa is genuine. We also extend previous research by showing that fa'afafine exhibit significantly higher avuncular tendencies even when compared to a more closely matched control group that also lacks direct parental care responsibilities (i.e., gynephilic men with no children). Although the greater avuncular tendencies of fa'afafine relative to gynephilic men are consistent with the predictions of the kin selection hypothesis for male androphilia, further research is needed before deeming male androphilia an adaptation for promoting elevated avuncularity. Likewise, more research is needed before deeming elevated avuncularity in fa'afafine an evolved adaptation for promoting indirect fitness. We discuss these findings in the context of alternative evolutionary explanations for male androphilia (i.e., an evolved by-product of an adaptation).

Full text

Page 1

ORIGINAL PAPER
Avuncular Tendencies and the Evolution of Male Androphilia
in Samoan Fa’afafine
Paul L. Vasey Æ Æ Doug P. VanderLaan
Received: 27 February 2008/Revised: 24 June 2008/Accepted: 24 June 2008
? Springer Science+Business Media, LLC 2008
Abstract
philia holds that genes for male androphilia can be main-
tained in a population if the fitness costs of not reproducing
directly are offset by enhancing indirect fitness. Kin share
some proportion of genes identical by virtue of descent.
Theoretically speaking, androphilic males can increase their
fitnessindirectlybyallocatingaltruisticbehaviortowardkin,
which, in turn, allows kin to increase their reproductive
success. Research conducted in Independent Samoa has
shown that androphilic males (known locally as fa’afafine)
report significantly higher avuncular tendencies relative to
gynephilic men. Here, we replicate this sexual orientation
difference, using a larger, independent sample, suggesting
that the documented sexual orientation difference in avun-
cular tendencies in Independent Samoa is genuine. We also
extend previous research by showing that fa’afafine exhibit
significantly higher avuncular tendencies even when com-
paredtoamorecloselymatchedcontrolgroupthatalsolacks
directparentalcareresponsibilities(i.e.,gynephilicmenwith
no children). Although the greater avuncular tendencies of
fa’afafine relative to gynephilic men are consistent with the
predictions of the kin selection hypothesis for male andro-
philia, further research is needed before deeming male an-
drophiliaanadaptationforpromotingelevatedavuncularity.
Likewise, more research is needed before deeming elevated
avuncularity in fa’afafine an evolved adaptation for pro-
moting indirect fitness. We discuss these findings in the
context of alternative evolutionary explanations for male
androphilia (i.e., an evolved by-product of an adaptation).
The kin selection hypothesis for male andro-
Keywords
Avuncular tendencies ? Kin selection
Male androphilia ? Samoa ? Evolution ?
Introduction
A large body of research indicates that there is a biological
basis for male androphilia1(Mustanski, Chivers, & Bailey,
2002),andfamilialstudiespointtoageneticcomponent(e.g.,
Bailey,Dunne,&Martin,2000;Kendler,Thornton,Gilman,
& Kessler, 2000). At the same time, research demonstrates
that androphilic males reproduce at about one-fifth to one-
tenth the rate of gynephilic males (Bell & Weinberg, 1978;
Hamer & Copeland, 1994; Saghir & Robins, 1973; Yankel-
ovichPartners,1994).Inlightoftheapparentfitnessbenefits
associatedwithmalegynephilia,onewouldexpectgenesfor
male gynephilia to have long replaced those for male an-
drophilia. Despite this prediction, archaeologic evidence
suggests that male same-sex sexual behavior existed during
humanprehistory(e.g.,Nash,2001;Yates,1993).Moreover,
male androphilia seems to occur at similar (albeit, low) fre-
quencies across different cultural and environmental land-
scapes (Whitam, 1983). This situation is perplexing when
considered within the context of natural selection, a process
that favors the persistence of those traits enabling their
bearers to achieve reproductive success. As such, the main-
tenance of a trait that lowers direct reproduction requires
explanation when viewed from a functional perspective.
The kin selection hypothesis has been advanced as one
possible explanatory framework to account for male
P. L. Vasey (&) ? D. P. VanderLaan
Department of Psychology, University of Lethbridge, 4401
University Drive, Lethbridge, Alberta, Canada T1K 3M4
e-mail: paul.vasey@uleth.ca
1Androphilia refers to sexual attraction to adult males, whereas
gynephilia refers to sexual attraction to adult females. The usage and
meaning of homosexual and heterosexual vary cross-culturally, ren-
dering them poor constructs for the type of cross-cultural research
presented here.
123
Arch Sex Behav
DOI 10.1007/s10508-008-9404-3

Page 2

androphilia (Ruse, 1982; Weinrich, 1987; Wilson, 1975).
The hypothesis holds that genes for male androphilia can be
maintained in the population if the fitness costs of not
reproducing directly are offset by enhancing indirect fitness.
From the perspective of kin selection theory, indirect fitness
isameasureofanindividual’simpactonthefitnessofitskin
(who share some identical genes by virtue of descent),
weighted by the degree of relatedness. Theoretically speak-
ing, androphilic males can increase their indirect fitness by
directing altruistic behavior toward kin, which, in principle,
allows kin to increase their reproductive success. In partic-
ular, androphilic men should allocate altruistic behavior to-
wards close kin, because they share more genes in common
with such individuals.
In formulating this theory, Wilson (1975) stated that
‘‘Freed from the special obligations of parental duties, they
[androphilic males] could have operated with special effi-
ciency in assisting close relatives’’ (p. 555). Similarly, Ruse
(1982) commented that ‘‘…the effect is that in being homo-
sexual,offspringbecomealtruistictowardscloserelativesin
order thereby toincrease their ownoverall inclusive fitness’’
(p. 20). Given that what is at issue here is a theory that can
accountforthe originofsame-sex sexualattraction,itseems
reasonable to interpret these statements as indicating that
same-sex sexual attraction, itself, is a prerequisite for the
expressionof elevated kin-direct altruism,notchildlessness.
If so, then male androphiles should exhibit elevated kin-
directed altruism, whereas male gynephiles (childless or
otherwise) should not. Such a pattern would be consistent
with the notion that male androphilia is a specially designed
adaptation for promoting kin-directed altruism.
Bobrow and Bailey (2001) found that androphilic men in
the United States didnotdiffer significantlyfrom gynephilic
men in terms of their avuncular tendencies. In addition, they
found that androphilic men were more estranged than gyne-
philic men from their respective kin, which runs contrary to
the predictions of the kin selection hypothesis for male an-
drophilia. In a similar study conducted in England, Rahman
andHull(2005)alsofoundnoevidencethatandrophilicmen
were more inclined towards avuncularity compared to their
gynephilic counterparts.
Vasey, Pocock, and VanderLaan (2007) compared the
altruistic tendencies of androphilic and gynephilic males in
the non-Western Polynesian island nation of Independent
Samoausingsimilarmethodstothoseemployedbyprevious
researchers working in the US and England (Bobrow &
Bailey,2001;Rahman&Hull,2005).InIndependentSamoa,
androphilic males are referred to as fa’afafine, which means
‘‘in the manner of a woman.’’ Most fa’afafine tend to be
effeminate, but they range from extremely feminine to
unremarkably masculine,although instances of the latter are
rare (Bartlett & Vasey, 2006; Besnier, 2000; Schmidt, 2003;
Vasey&Bartlett,2007).Despitethisheterogeneityingender
role presentation, fa’afafine are, with very few exceptions,
exclusively androphilic, but they do not engage in sexual
activity with each other. Instead, fa’afafine are attracted to,
andengageinsexualinteractionswith,masculinemaleswho
self-identify as ‘‘straight men’’ (Bartlett & Vasey, 2006;
Danielsson, Danielsson, & Pierson, 1978).
In a Samoan cultural context, ‘‘straight men’’ are those
who self-identify as men and are masculine with respect to
gender role presentation. Inclusion in this category is not
contingent on exclusive sexual activity with women. Most
self-identifiedstraightmenaregynephilic,butmayengagein
sexual activity with fa’afafine or other straight men on a
temporary basis, particularly if female sexual partners are
unavailable. Our participants informed us that most straight
men in Samoa have engaged in sexual interactions with
fa’afafine at least once in their lives (see also Croall &
Wunderman, 1999).
In contrast to research conducted in Western settings,
Vasey et al. (2007) found that fa’afafine reported signifi-
cantlyhigheravunculartendenciesrelativetostraightmenin
Independent Samoa. These findings are consistent with the
basic prediction of the kin selection hypothesis for male an-
drophilia,andraisethepossibilitythatandrophilicfa’afafine
may act as ‘‘helper-in-the-nest,’’ caring for nieces and
nephews and, by extension, increasing their indirect fitness.
Although their findings were consistent with the basic
prediction of the kin selection hypothesis for male andro-
philia, Vasey et al. (2007) were careful to stress that their
results did not provide strong evidence in support of the
conclusionthateithermaleandrophiliaorelevatedavuncular
tendencies in fa’afafine represent specially designed adap-
tations resulting from past selection over evolutionary time.
Rather, they suggested that increased avuncular tendencies
by fa’afafine might simply reflect a generalized adaptive
tendency on the part of all biological males to invest in kin,
regardless of their sexual orientations. Some males, such as
the fa’afafine, may, however, be able to exhibit elevated
levels of avuncularity because they have no direct parental
careresponsibilities.IntheVaseyetal.(2007)study,noneof
thefa’afafineparticipantshadchildren(n = 38).Incontrast,
58%ofthegynephilicmen(n = 43)whoparticipatedinthat
study had at least one child (range = 0–4). Unfortunately,
Vasey et al.’s (2007) samples of gynephilic men with, and
without, children were too small to perform the necessary
analyses to properly test this hypothesis.
Inthisarticle,ourgoalwastoascertainwhetherthesexual
orientation difference in avuncular tendencies originally re-
ported in Vasey et al. (2007) could be replicated using a
larger, independent sample. In addition, we extended previ-
ousresearchbycomparingavunculartendenciesinfa’afafine
with two distinct control groups, namely, gynephilic men
with and without children. In conducting these latter com-
parisons, our aim was to test whether fa’afafine’s elevated
Arch Sex Behav
123

Page 3

avunculartendencies, ifpresent,couldbeexplainedinterms
oftheirlackofdirectparentalcareresponsibilities.Ifso,then
avuncular tendencies in fa’afafine should not differ signifi-
cantly from those of gynephilic men without children, but
both of these groups should differ significantly for this
measure from gynephilic men with children.
Method
Participants
All participants were recruited through a network sampling
procedure on the two larger and more populated islands of
IndependentSamoa:UpoluandSavai’i.Anetworksampling
procedure involves contacting initial participants who dis-
play qualities of interest (i.e., status as fa’afafine or gyne-
philic men), then obtaining referrals from them to additional
participantswho,inturn,providefurtherreferrals,andsoon.
Therateofparticipationforallgroupswasgreaterthan90%.
To replicate the study by Vasey et al. (2007), new data
were collected from 56 self-identified fa’afafine and 95 self-
identified straight men that had not been interviewed previ-
ously. These data were collected during two field trips
(September–October 2006, April–June 2007). Sexual ori-
entationwasassessedusingKinseyratingsofsexualfeelings
over the previous year (Kinsey, Pomeroy, & Martin, 1948).
All 56 of these fa’afafine described their sexual feelings
as exclusively androphilic (Kinsey rating = 6). Of the 95
straight men for whom Kinsey ratings were obtained, 82
(86.3%) described their sexual feelings as exclusively gy-
nephilic (Kinsey rating = 0). Seven (7.4%) reported most
sexual feelings toward females, but occasional fantasies
about males (Kinsey rating = 1), and six (6.3%) reported
mostsexualfeelingstowardfemales,butsomedefinitesexual
feelings about males (Kinsey rating = 2).
In order to obtain sufficiently large sample sizes to com-
pare fa’afafine, gynephilic men with no children, and gyne-
philicmenwithatleastonechild,wecombinedthedatafrom
the 56 fa’afafine and 95 gynephilic men in the replication
sample with data from the sample of 38 fa’afafine and 43
gynephilic men interviewed in Vasey et al. (2007). Of the
additional 38 fa’afafine interviewed in Vasey et al., 37
(97.4%) described their sexual feelings as exclusively an-
drophilic(Kinseyrating = 6),andone(2.6%)reportedmost
sexual feelings toward males, but some definite feelings
toward females (Kinsey rating = 4). Of the additional 43
gynephilic men interviewed in Vasey et al., 35 (81.4%) de-
scribed their sexual feelings as exclusively gynephilic, five
(11.6%) reported most sexual feelings toward females,
but occasional fantasies about males (Kinsey rating = 1),
andthree(7%)reportedmostsexualfeelingstowardfemales,
but some definite sexual feelings about males (Kinsey
rating = 2). After combining these two samples, there were
94 fa’afafine, 66 gynephilic men with no children, and 72
gynephilic men with at least one child.
Procedure and Measures
All participants were interviewed using standardized ques-
tionnaires. A Samoan-speaking research assistant was pres-
ent for those interviews for which participants indicated that
they preferred to do the interview in Samoan or for partici-
pantswhoweredeemedbytheresearcherstobeinsufficiently
fluent in English. Questions were read aloud in English by
one of the researchers and in Samoan by a research assistant
when necessary. The questionnaire used in this study was
available in English and Samoan, after being translated and
back-translated by two fluent Samoan-English speakers.
The questionnaire employed in this study was a modified
version of a previously used Kin Selection Questionnaire
(Bobrow&Bailey,2001;Rahman&Hull,2005;Vaseyetal.,
2007). The questionnaire included questions about the fol-
lowing basic biographic information: age, sex, sexual ori-
entation identity (i.e., fa’afafine or ‘‘straight’’ man), highest
levelofeducationreceived(i.e.,post-secondary,highschool,
junior high school, and primary school or less), and annual
income. Data on the participants’ annual incomes were
converted to American dollars (USD). Samoans, both inside
and outside the fa’afafine community, recognize that
fa’afafine are biological males that are socially distinct from
men and women. Nevertheless, for the sake of consistency,
participants were told, prior to answering questions pertain-
ingtotheKinseyscale(Kinseyetal.,1948),thatthecategory
‘‘males’’includedstraightmenand/orfa’afafine,whereasthe
category ‘‘females’’ included women.
Participants were also asked to complete the Avuncular
Tendencies Subscale. This subscale is a measurement in-
strumentcontainingnineitemsandhasbeenusedtoquantify
willingness to allocate resources to nieces and neph-
ewsinpreviousstudies(Bobrow&Bailey,2001;Rahman&
Hull, 2005; Vasey et al., 2007). The nine items of the
Avuncular Tendencies Subscale are as follows: babysitting
for an evening, babysitting on a regular basis, taking care of
the children for a week while their parents are away, buying
toys forthe children, tutoring one of the children ina subject
you know well, helping to expose the children to art and
music, contributing money for daycare, contributing money
forthechildren’smedicalexpenses,andcontributing money
forthechildren’seducation.Foreachitem,participantswere
asked to rate whether they would be willing to exhibit the
behavior towards nieces and nephews that were the children
ofasiblingwithwhomtheywereemotionallycloseandwho
lived nearby. Responses to these items were based on a 7-
point Likert-type scale that ranged from 1 = ‘‘Strongly
Disagree’’to7 = ‘‘StronglyAgree.’’Participants’avuncular
Arch Sex Behav
123

Page 4

tendenciesscoreswerecalculatedasthemeanratinggivento
the nine items.
In keeping with our previous study (Vasey et al., 2007),
ordinal data were analyzed using a general linear model
(GLM) for an ordinal multi-way frequency analysis (see
Vokey, 2003).
Results
Replication Study
Table 1 summarizes results of the replication portion of the
presentstudy.Internalconsistencyreliabilities,standardized
item alpha (a), were computed for both fa’afafine and gy-
nephilic men on the Avuncular Tendencies Subscale. Reli-
abilities were high on this subscale for both groups
(fa’afafine: a = .85; gynephilic men: a = .88). Two-tailed
independent t-tests revealed that fa’afafine and gynephilic
men did not differ in terms of age or annual income. A GLM
showed that fa’afafine and gynephilic men did not differ in
terms of the highest level of education they received
(G1
tendencies than gynephilic men. A Cohen’s d indicated a
moderate effect size difference between fa’afafine and gy-
nephilic men for avuncular tendencies (d = .57).
2= .001, ns). Fa’afafine exhibited greater avuncular
Comparison of Fa’afafine, Gynephilic Men with No
Children, and Gynephilic Men with at Least One Child
Table 2summarizesresultsofthecomparisonsoffa’afafine,
gynephilicmenwithnochildren,andgynephilicmenwithat
least one child. Internal consistency reliabilities, standard-
izeditemalpha(a),werecomputedforfa’afafine,gynephilic
men without children, and gynephilic men with at least one
child (range = 1–12) on the Avuncular Tendencies Sub-
scale. Reliabilities were high for all three groups on this
subscale (fa’afafine: a = .84; gynephilic men with no chil-
dren: a = .83; gynephilic men with at least one child:
a = .91).
Ananalysisofvariance(ANOVA)indicatedamaineffect
of group for age of participant. Post-hoc Fisher’s Least Sig-
nificant Difference (LSD) revealed that fa’afafine were sig-
nificantly older than gynephilic men with no children (p\
.001), but significantly younger than gynephilic men with at
least one child (p\.001). Gynephilic men with at least one
child were significantly older than gynephilic men with no
children (p\.001). An ANOVA indicated a main effect of
group for annual income. Post-hoc Fisher’s LSD revealed
that fa’afafine had significantly higher annual incomes than
gynephilicmenwithnochildren(p\.001),butdidnotdiffer
in this regard from gynephilic men with at least one child.
Gynephilic men with at least one child had significantly
higherannualincomesthangynephilicmenwithnochildren
(p = .002). A GLM showed that fa’afafine and gynephilic
men did not differ in terms of the highest level of education
they received (G2
An analysis of covariance (ANCOVA) was performed
withAvuncularTendenciesasthedependentvariable,group
as the fixed factor, and with age and annual income as
covariates. This analysis revealed a main effect of group.
Post-hoc Fisher’s LSD revealed that fa’afafine had signifi-
cantly higher avuncular tendencies scores relative to gyne-
philic men without children (p = .001; Cohen’s d = .65)
and gynephilic men with at least one child (p = .001; Co-
hen’s d = .56). The two groups of gynephilic men did not
differ significantly from each other (Cohen’s d = .01).
Comparative data on individual Avuncular Tendencies
Subscale items for fa’afafine, gynephilic men with no chil-
dren,andgynephilicmenwithatleastonechildarepresented
in Table 3.
A two-tailed Pearson’s r correlation indicated no signifi-
cant relationship between avuncular tendencies and number
of children parented among gynephilic men (n = 138,
r = .02, p = .81).
Partial correlations between avuncular tendencies and
age, controlling for income, were calculated for each par-
ticipantgroup.Therewasnosignificantrelationshipbetween
thesevariablesforfa’afafine(r = .07,df = 91,p = .52),gy-
nephilic men without children (r = .20, df = 63, p = .12),
2= 1.53, ns).
Table 1 Replication sample: results summary
Fa’afafine (n = 56)Gynephilic men (n = 95) Two-tailed t-test
MSDMSDta
dfp
Age (in years)27.95 5.9627.828.66
\1b
1.84c
3.49d
145.31
ns
Income (USD)6099.119496.57 3661.42 3712.77
65.05 .07
Avuncular tendencies 6.271.045.621.30
135.66.001
aBetween group equality of variances not assumed
bLevene’s test for equality of variances, F = 5.85, p = .017
cLevene’s test for equality of variances, F = 4.08, p = .045
dLevene’s test for equality of variances, F = 5.49, p = .02
Arch Sex Behav
123

Page 5

orgynephilicmenwithatleastonechild(r = -.20,df = 69,
p = .10).
Also, partial correlations between avuncular tendencies
and income, controlling for age, were calculated for each
participant group. There was also no significant relationship
between these variables for fa’afafine (r = -.06, df = 91,
p = .57), gynephilic men without children (r = .16, df =
63, p = .22), or gynephilic men with at least one child
(r = .04, df = 69, p = .73).
Discussion
In contrast to research conducted in Western countries
(Bobrow&Bailey,2001;Rahman&Hull,2005),Vaseyetal.
(2007) showed that androphilic males (fa’afafine) reported
significantly greater avuncular tendencies than gynephilic
males (‘‘straight men’’) in the Polynesian island nation of
Independent Samoa. In this study, we replicated this sexual
orientationdifferenceinavunculartendenciesusingalarger,
independent sample. This replication suggests that the doc-
umented sexual orientation difference in avuncular tenden-
cies in Independent Samoa is genuine and not the result of
sampling bias. This finding is consistent with the kin selec-
tion hypothesis prediction that androphilic males should
direct more altruism toward kin than gynephilic males.
A number of potentially interrelated proximate factors
might account for why the avuncular tendency results ob-
tained for androphilic males (fa’afafine) in this study and
Vaseyetal.(2007)differedfromthoseconductedinWestern
countries (Bobrow & Bailey, 2001; Rahman & Hull, 2005),
despitethefactthatallemployedverysimilarmethodologies.
To begin with, Independent Samoa is a relatively tiny nation
consisting of four populated islands, which are closely
Table 2 Combined sample: results summary
Fa’afafine (n = 94)Gynephilic men with no
children (n = 66)
Gynephilic men with at
least one child (n = 72)
Analysis of variancea
MSDM SDMSDFdfwithin
p
Age (in years)c,d,e
Income (USD)c,e
Avuncular tendenciesb,c,d
29.48 7.3121.884.2233.397.4853.33229
\.001
.001
\.001
5956.367980.222400.50 3517.595636.105096.89 7.42 229
6.20 .925.561.065.55 1.379.18 227
aBetween-groups df = 2 for all analyses
bGroups were compared using ANCOVA with age and income included in the model as covariates
cStatistically significant difference (p\.05) between fa’afafine and gynephilic men with no children
dStatistically significant difference (p\.05) between fa’afafine and gynephilic men with at least one child
eStatistically significant difference (p\.05) between gynephilic men with no children and gynephilic men with at least one child
Table 3 IndividualAvuncularTendenciesSubscaleitemsforfa’afafine,gynephilicmenwithnochildren,andgynephilicmenwithatleastonechild,
controlling for age and income: results summary
Act
Fa’afafine
(n = 94)
Gynephilic men
with no children
(n = 66)
Gynephilic men
with at least one
child (n = 72)
F2, 227
p
MSDM SDM SD
Babysitting for an eveninga
Babysitting on a regular basisa,b
Taking care of the children for a week while their parents are awaya,b
Buying toys for the childrena,b
Tutoring one of the children in a subject you know wellb
Helping to expose the children to art and musicb,c
6.361.355.531.89 5.841.89 4.10.018
5.512.01 4.552.194.902.12 3.64.028
5.71 1.894.772.22 4.81 2.304.96 .008
6.261.44 5.36 1.765.67 1.676.09.003
6.64 0.79 6.36 1.106.00 1.70 4.57 .011
\.001
ns
6.47 1.166.261.13 5.451.93 8.40
Contributing money for daycare
Contributing money for the children’s medical expensesa
Contributing money for the children’s educationa,b
5.731.805.18 1.75 5.17 1.79 2.60
6.491.17 6.08 1.29 6.101.44 2.75
ns
6.67 0.875.951.415.97 1.548.54
\.001
aStatistically significant difference (p\.05) between fa’afafine and gynephilic men with no children
bStatistically significant difference (p\.05) between fa’afafine and gynephilic men with at least one child
cStatistically significant difference (p\.05) between gynephilic men with no children and gynephilic men with at least one child
Arch Sex Behav
123