ArticlePDF Available

Sociosexuality from Argentina to Zimbabwe: A 48-Nation Study of Sex, Culture, and Strategies of Human Mating

April 2005
Behavioral and Brain Sciences 28(2):247-75; discussion 275-311

Source
PubMed

Authors:

Brunel University London

The Sociosexual Orientation Inventory (SOI; Simpson & Gangestad 1991) is a self-report measure of individual differences in human mating strategies. Low SOI scores signify that a person is sociosexually restricted, or follows a more monogamous mating strategy. High SOI scores indicate that an individual is unrestricted, or has a more promiscuous mating strategy. As part of the International Sexuality Description Project (ISDP), the SOI was translated from English into 25 additional languages and administered to a total sample of 14,059 people across 48 nations. Responses to the SOI were used to address four main issues. First, the psychometric properties of the SOI were examined in cross-cultural perspective. The SOI possessed adequate reliability and validity both within and across a diverse range of modem cultures. Second, theories concerning the systematic distribution of sociosexuality across cultures were evaluated. Both operational sex ratios and reproductively demanding environments related in evolutionary-predicted ways to national levels of sociosexuality. Third, sex differences in sociosexuality were generally large and demonstrated cross-cultural universality across the 48 nations of the ISDP, confirming several evolutionary theories of human mating. Fourth, sex differences in sociosexuality were significantly larger when reproductive environments were demanding but were reduced to more moderate levels in cultures with more political and economic gender equality. Implications for evolutionary and social role theories of human sexuality are discussed.

National level of sociosexuality related to operational sex ratio across 48 nations of the International Sexuality Description Project.

…

Miller et al.). Universal (sex-similar) systems of mechanisms afford enduring pair-bonding (shaded boxes and arrows) while producing dating outcomes (from short-term to nonenduring pair-bonds) as by-products (solid black arrow). The number of these depends, in part, on time until an emotionally close pairbond and whether that bond is maintained. Other mechanisms afford relationship repair (e.g., protest, despair) and even permanent detachment when there is sufficient sustained negative or insufficient positive affect. Then, the process can begin anew. Variability in mechanism parameter settings (e.g., relative levels of neuromodulators) as a result of experiential, maturational, and biological factors produce emergent within and between-subject diversity in mating outcomes over time.

…

. Means and standard deviations of sociosexuality across 48 nations of the International Sexuality Description Project (controlling for sex of participant) Nation Mean SD

…

. Are sex differences in sociosexuality associated with the demanding nature of local environments? r to z test for

…

Figure R2. Relationship status and sex differences in sociosexuality.

…

Figures - uploaded by David Schmitt

Content may be subject to copyright.

Content uploaded by David Schmitt

Content may be subject to copyright.

Over a decade ago, Simpson and Gangestad (1991) intro-

duced a self-report measure of human sexuality called the

Sociosexual Orientation Inventory (SOI; see also Ganges-

tad & Simpson 1990; Simpson 1998). Originally designed to

capture variability in the willingness to have sex outside of

a committed pair-bond, the SOI contains numerous ques-

tions about human mating behavior, romantic fantasies, re-

lational emotions, and attitudes toward casual sex. Re-

sponses to the SOI are typically used to differentiate people

along a single strategic dimension of human mating called

sociosexuality (Simpson & Gangestad 1991). Those who

score relatively low on this dimension are said to possess

a restricted sociosexual orientation – they tend toward

monogamy, prolonged courtship, and heavy emotional

investment in long-term relationships. Those residing at

the high end of sociosexuality are considered more unre-

stricted in mating orientation, they tend toward promiscu-

ity, are quick to have sex, and experience lower levels of

romantic relationship closeness

(Simpson & Gangestad

1991).

Since its introduction, the SOI has become an increas-

ingly popular tool for measuring individual differences in

basic human mating strategies (Hebl & Kashy 1995; Jones

1998; Seal et al. 1994; Simpson 1998; Simpson et al. 2004;

Stephan & Bachman 1999; Wright & Reise 1997). Indeed,

it appears to have become the measure of choice when at-

tempting to relate human mating strategies to other sex-re-

lated phenomena (Allen 2000; Bleske-Rechek & Buss 2001;

Clark 2004; Gangestad & Thornhill 1997; Isaacson 2001;

Reise & Wright 1996; Schmitt 2005; Seal & Agostinelli

1994; Simon 1997, Simpson et al. 1999; Townsend &

Wasserman 1988). Despite its widespread use, very little is

known about the cross-cultural utility of the SOI, with only

a handful of studies directly measuring sociosexuality out-

side the United States (e.g., Bailey et al. 2000). This is es-

BEHAVIORAL AND BRAIN SCIENCES (2005) 28, 247–311

Printed in the United States of America

247

Sociosexuality from Argentina

to Zimbabwe: A 48-nation

study of sex, culture, and strategies

of human mating

David P. Schmitt

Department of Psychology, Bradley University, Peoria, IL 62625

dps@bradley.edu

http://www.bradley.edu/academics/las/psy/schmitt.html

http://schmitt.socialpsychology.org/index.htm

Abstract: The Sociosexual Orientation Inventory (SOI; Simpson & Gangestad 1991) is a self-report measure of individual differences

in human mating strategies. Low SOI scores signify that a person is sociosexually restricted, or follows a more monogamous mating strat-

egy. High SOI scores indicate that an individual is unrestricted, or has a more promiscuous mating strategy. As part of the International

Sexuality Description Project (ISDP), the SOI was translated from English into 25 additional languages and administered to a total sam-

ple of 14,059 people across 48 nations. Responses to the SOI were used to address four main issues. First, the psychometric properties

of the SOI were examined in cross-cultural perspective. The SOI possessed adequate reliability and validity both within and across a di-

verse range of modern cultures. Second, theories concerning the systematic distribution of sociosexuality across cultures were evaluated.

Both operational sex ratios and reproductively demanding environments related in evolutionary-predicted ways to national levels of so-

ciosexuality. Third, sex differences in sociosexuality were generally large and demonstrated cross-cultural universality across the 48 na-

tions of the ISDP, confirming several evolutionary theories of human mating. Fourth, sex differences in sociosexuality were significantly

larger when reproductive environments were demanding but were reduced to more moderate levels in cultures with more political and

economic gender equality. Implications for evolutionary and social role theories of human sexuality are discussed.

Keywords: culture; gender; mating; reproduction; sex differences; sex roles; sexual strategies; sociosexuality

David P. Schmitt received his Ph.D. from the Uni-

versity of Michigan in 1995 and is currently an Associ-

ate Professor of Psychology at Bradley University. He

is Founding Director of the International Sexuality

Description Project (ISDP), a collaborative research

group of over 100 scientists from around the world. One

of the aims of the ISDP is to uncover the ways in which

gender, personality, and culture influence basic human

mating strategies. The second wave of the ISDP is cur-

rently underway and will investigate the personality and

cultural predictors of sexual aggression, domestic vio-

lence, and risky behavior related to HIV/AIDS.

pecially unfortunate because a cross-culturally validated

measure of human mating strategies would help to address

many recent developments in evolutionary psychology and

social role theory (e.g., Gangestad & Simpson 2000; Wood

& Eagly 2002).

In the current study, the SOI was translated from English

into 25 additional languages and administered to samples

from 48 nations as part of the International Sexuality De-

scription Project (ISDP; Schmitt et al. 2003a). The result-

ing ISDP database on sociosexuality was used to address

four main issues. First, the psychometric properties of the

SOI were examined in cross-cultural perspective. Second,

theories concerning the distribution of sociosexuality across

cultures were evaluated (Belsky et al. 1991; Gangestad &

Simpson 2000; Pedersen 1991). Third, the cultural univer-

sality of sex differences in sociosexuality was investigated

(Buss & Schmitt 1993; Symons 1979; Trivers 1972). Fourth,

theories concerning the degree of sexual differentiation in

sociosexuality across cultures were examined (Buss &

Barnes 1986; Eagly & Wood 1999; Kasser & Sharma 1999;

Wood & Eagly 2002). Because national profiles of men’s

and women’s sociosexual mating strategies would have lit-

tle value if the SOI were not reliable and valid across cul-

tures, a review of the psychometrics of the SOI will be ad-

dressed first.

1. Psychometrics of the SOI

Simpson and Gangestad (1991) conducted several studies

to evaluate the psychometric qualities of the SOI. In their

original study, 204 women and 202 men from Texas A&M

University completed an initial pool of 11 items related to

a conceptual definition of sociosexuality. Using principal

axis factor analysis, Simpson and Gangestad documented

that seven of these items – the seven items eventually in-

cluded in the SOI – formed a coherent unitary factor

structure. They also found that the resulting seven-item

SOI scale possessed adequate levels of internal reliability

(a  0.73). In an unpublished study, Simpson and Gan-

gestad (1989) documented the high temporal reliability of

sociosexuality (r 0.94) over a two-month test–retest

period.

Simpson and Gangestad (1991) evaluated the validity of

the SOI using standard construct validation techniques

(Cronbach & Meehl 1955). In a validation study involving

144 romantic couples, participants were asked to complete

the SOI along with other measures concerning their rela-

tionships. Simpson and Gangestad found that sociosexual-

ity was related to how early in the relationship the couple

had engaged intercourse, and SOI scores converged with

established measures of human sexuality in predictable

ways. For example, those who scored as more restricted on

the SOI tended to score higher on Lund’s Commitment

Scale (Lund 1985), Rusbult’s Investment Scale (Rusbult

1980), and Rubin’s Love Scale (Rubin 1970). Data from

these same couples also showed that the SOI possessed dis-

criminant validity, in that sociosexuality was unrelated to sex

drive per se. In additional studies, Simpson and Gangestad

(1991) have shown that the SOI predicts whether someone

has engaged in sex with more than one partner in a given

time period – a key conceptual element of the unrestricted

or promiscuous mating strategy (Baker & Bellis 1994;

Barash & Lipton 2001; Schmitt et al. 2001b).

Although the majority of SOI validation research has

proven highly supportive in American samples, it remains

unclear whether the psychometric soundness of the SOI

generalizes to other languages and cultures. Do the seven

items of the SOI form one coherent dimension within all

cultures, or do sociosexual attitudes and behaviors become

disconnected in certain regions of the world? Can socio-

sexuality be accurately gauged using self-report methods

across all cultures, or does the validity of the SOI fluctuate

across language, geography, ethnicity, history, politics, eco-

nomics, or religion? Do responses to the SOI correlate with

similar psychological and physical attributes across differ-

ent cultures, or does sociosexuality take different anthro-

pometric forms around the world? If the SOI were proven

reliable and valid across cultures, this could have impor-

tant implications for advancing our understanding of the

links between culture and human mating strategies. How-

ever, no study has evaluated the psychometrics of the SOI

in a language other than English, nor has the SOI been ad-

ministered to samples from non-Western cultures. In the

current study, the reliability and validity of the SOI were

evaluated across 26 languages and 48 nations, including

multiple cultures from North America, South America,

Western Europe, Eastern Europe, Southern Europe, the

Middle East, Africa, Oceania, South/Southeast Asia, and

East Asia.

2. Sex differences in sociosexuality

On average, men tend to possess more positive attitudes to-

ward casual, low-investment sex than women do (Carrol et

al. 1985; Fisher et al. 1988; Hendrick et al. 1985; Oliver &

Hyde 1993; Townsend 1995; Wilson 1987). Men also report

that they fantasize about having sex with multiple partners

more than women do (Ellis & Symons 1990; Malamuth

1996), and men behaviorally seek short-term mateships

more than women do (Blumstein & Schwartz 1994;

Eysenck 1976; Laumman et al. 1994; Wiederman 1997).

Experimental tests have further confirmed that men are

more likely than women to consent to sex with a stranger

when approached in a community setting (Clark & Hatfield

1989), even when the stranger is “vouched for” by a partic-

ipant’s same-sex friend (Clark 1990).

This pervasive pattern of sexual differences – across atti-

tudes, fantasy, and behavior – implies that men should be

higher or more unrestricted on sociosexuality than women.

Indeed, the direct evidence on this point is unequivocal, at

least in United States. In every study published to date,

American men report higher levels of sociosexuality than

American women based on responses to the SOI. What re-

mains unknown is whether sex differences in sociosexuality

persist beyond the borders of the United States, especially

across non-Western cultures. Perhaps some cultures are

so generally restrained in sexual matters that sex differ-

ences in sociosexuality have become muted. Other cultures

could be sexually unrestrained to the point that a ceiling

effect occurs, and both sexes “max out” on the SOI. Cer-

tainly, sociosexual sex differences will vary to some degree

across cultures, and this variability itself may be of inter-

est. Nevertheless, there are strong theoretical reasons why

men are expected to sexually think, feel, and behave in a

more unrestricted manner than women do across most

cultures.

Schmitt: Sociosexuality from Argentina to Zimbabwe

248 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

2.1. Parental investment theory

From an evolutionary perspective, sex differences in socio-

sexuality should be consistently observed across human cul-

tures, in part because of fundamental differences in the

evolved reproductive strategies of men and women. Ac-

cording to parental investment theory (Trivers 1972), the

relative proportion of parental investment – the time and

energy devoted to the care of individual offspring – varies

across the males and females of different species. In some

species, males tend to provide more parental investment

than females (e.g., the Mormon cricket; Gwynne 1984). In

other species, females possess the heavy-investing parental

burdens (Alcock 1993; Trivers 1985).

Importantly, Trivers (1972) noted that sex differences in

obligatory parental investment burdens are systematically

linked to the processes of sexual selection in ways that may

influence reproductive strategies or sociosexual orienta-

tions. Namely, the sex that typically invests less in offspring

normally shows a greater eagerness to engage in mating, in-

curs greater costs through more intense intrasexual mating

competition, and is intersexually less discriminating in mate

choice than the heavier-investing parent (Andersson 1994;

Bateson 1983; Clutton-Brock 1991; Maynard Smith 1977).

In short, the lesser-investing sex is usually more unre-

stricted in sociosexual orientation than the heavier-invest-

ing sex. In support of parental investment theory applying

to humans, numerous studies have shown that men possess

a greater eagerness to engage in mating (Baumeister et al.

2001; Clark & Hatfield 1989; Schmitt et al. 2003b); men in-

cur greater costs through more intense intrasexual mating

competitions (Alexander & Noonan 1979; Archer & Loyd

2002; Daly & Wilson 1988); and men’s mate preferences are

less discriminating than women’s, especially in the context

of promiscuous or short-term mating (Buss & Schmitt

1993; Kenrick et al. 1990; Regan 1998a; 1998b; Regan &

Berscheid 1997; Simpson & Gangestad 1992).

Human males also experience lower levels of minimum

parental investment in offspring than females do. That is,

men are not obligated to invest as much as women do in

parenting to produce viable progeny (Symons 1979).

Women must incur the differential costs of internal fertil-

ization, placentation, and gestation to reproduce. All female

mammals carry additional investment burdens associated

with lactation. In humans, lactation can last several years in

a foraging environment (Kelly 1995), years during which it

is harder for women to reproduce and invest in additional

offspring. Men are minimally required to do much less to

reproduce. This investment differential is perhaps most

conspicuous in the case of mate poaching, where a man can

mate with a married woman and then have the woman’s

husband provide extended paternal investments to the

child (Schmitt & Buss 2001). For a woman seeking to poach

a married man, a similar shift of her minimum investment

burdens to his wife would have been unattainable in our an-

cestral past. Of course, many men do invest heavily in their

own children, but in all known cultures women spend much

more time and effort in actively raising children than men

do (Low 1989; Munroe & Munroe 1997; Quinn 1977).

According to parental investment theory, these asymme-

tries in men’s and women’s parental investment levels

should lead women to have less to gain in reproductive out-

put by engaging in indiscriminate, short-term sex with large

numbers of partners (see Bjorklund & Shackelford 1999;

Geary 1998; Hinde 1984). Indeed, the differences between

men’s and women’s potential reproductive benefits from

unrestricted, promiscuous mating may be substantial. Con-

sider that one man can produce as many as 100 offspring by

repeatedly mating with 100 women in a given year, whereas

a man who is monogamous will tend to have only one child

with his partner during that same time period. In evolu-

tionary currencies, this represents a strong selective pres-

sure on men’s mating strategies to favor at least some unre-

stricted desires for multiple partners (Barash & Lipton

2001; Buss & Schmitt 1993). Of course, 100 instances of

only one-time mating between a man and 100 women

would rarely, if ever, produce precisely 100 offspring

(Fletcher & Stenswick 2003). However, this selective pres-

sure remains potent because a man mating with 100 women

over the course of a year – particularly repeated matings

when the women are nearing ovulation and are especially

interested in short-term mating (Gangestad 2001) – would

likely have significantly more offspring than a woman mat-

ing repeatedly with 100 men over the course of a year. His-

torically, this appears to have been the case with those men

having especially large numbers of mating partners greatly

outreproducing their intrasexual (and intersexual) contem-

poraries (Betzig 1986). This is also true among foraging cul-

tures, the vast majority (over 80%) of whom practice some

form of polygyny (Murdock 1967), whereby high-status

men who mate with multiple partners have greater repro-

ductive success than those who do not (though see Low

1988). Whether a woman mates with 100 men or is monog-

amously bonded with only one man, she will still tend to

produce only one child in a given year. The potential re-

productive benefits from desiring promiscuous or multiple

mating, therefore, appear to be much higher for men than

women (Symons 1979).

According to sexual strategies theory (Buss & Schmitt

1993), women can reap some evolutionary benefits from

multiple mating (see also Gangestad 2001; Hrdy 1981;

Shackelford & LeBlanc 2001). In Amazonian cultures that

believe in partible paternity, for example, a woman can re-

ceive the extended benefits of protection and resources

from multiple men by mating with them while she is preg-

nant (Beckerman & Valentine 2002). A woman who en-

gages in multiple mating can also obtain immediate re-

sources, secure a child if her current long-term mate is

infertile, and gain access to high-quality genes by short-

term mating with a man who is not her husband (Fedorka

& Mousseau 2002; Greiling & Buss 2000; Smith 1984).

Many married women appear to desire physically attractive

men when having affairs, affairs that may be one source for

attractive men to especially benefit from short-term mating

by having their offspring raised by cuckolded husbands

(Schmitt & Shackelford 2003). In short, women can repro-

ductively benefit from promiscuous or multiple short-term

mating. The pivotal sex difference in sociosexual mating

psychology is not that women are solely designed for long-

term monogamy. Instead, women possess all the hallmarks

of having evolved a short-term mating strategy (Gangestad

2001; Schmitt et al. 2001a). It is, however, a strategy based

on selectively desiring men of high status, dominance, and

genetic quality (Gangestad & Thornhill 1997; 2003; Thorn-

hill & Gangestad 2003). Men’s short-term strategy, in con-

trast, is focused on more indiscriminate desires that lead to

obtaining numerous sex partners in high-volume quantity

(Schmitt et al. 2003b).

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 249

A clear implication of parental investment theory and

sexual strategies theory is the following hypothesis: Men

should possess more unrestricted sociosexual orientations

than women across human cultures. There have been cross-

cultural studies that show men possess less restrictive mate

preferences than women (Buss 1989; Buunk et al. 2002)

and desire multiple short-term sex partners more than

women do (Schmitt et al. 2003b). However, sex differences

in sociosexuality have never been directly tested across

large numbers of cultures. Indeed, no studies have been

conducted where the sociosexuality of both men and women

were assessed in non-Western cultures. One of the objec-

tives of the present study was to evaluate whether sex dif-

ferences in sociosexuality are robust across the broad range

of human cultures represented in the ISDP. Finding uni-

versal sex differences in sociosexuality would support

parental investment theory (Trivers 1972), as well as other

evolutionary perspectives on human mating (Alexander &

Noonan 1979; Buss & Schmitt 1993; Gangestad & Simpson

2000; Hinde 1984; Symons 1979; Wilson 1987).

3. Cultural influences on sociosexuality

In addition to differences between men’s and women’s so-

ciosexuality, human mating strategies also appear to vary

across different forms of human culture (Broude 1983;

Broude & Greene 1976; Ember 1974; Hartung 1985;

Jankowiak et al. 2002; Lancaster 1989; Low 2000; Whiting

& Whiting 1975). Perhaps the most well-documented links

between culture and human mating strategies are those in-

volving marriage systems. For example, Low (1990) has

shown that tribal cultures with higher pathogen stress are

more likely to have polygynous marriage systems (see also

White & Burton 1988). Monogamous mating systems, in

contrast, are relatively absent in high-pathogen environ-

ments. Marlowe (2003) recently demonstrated that monog-

amy is especially prevalent in cultures with low levels of

pathogens and when men contribute more calories to the

local diet. Indeed, anthropologists have suggested that

many aspects of tribal culture – particularly warfare, kin-

ship, residence, and inheritance patterns – are systemati-

cally related to marriage systems, as well as to rules govern-

ing premarital sex, adultery, jealousy, divorce, postpartum

sex taboo, and incest avoidance (Frayser 1985; Pasternak et

al. 1997).

Theories that link cultural variation with the monogamy–

promiscuity dimension of sociosexuality have also been pro-

posed. For example, Pedersen (1991) has postulated that

the relative number of men versus women in a given cul-

ture should affect sociosexual attitudes and behavior.

Chisholm (1996; 1999a) has argued that high mortality

rates in local cultures should be associated with more

promiscuous mating strategies. Gangestad and Simpson

(2000) have theorized that demanding reproductive envi-

ronments should increase the desire and pursuit of bi-

parental, monogamous partnerships. To date, however, no

study has examined sociosexuality across multiple cultures

in a way that would reveal direct links between cultural en-

vironments and the dimension of sociosexuality. In the cur-

rent study, three theories concerning culture and its effects

on sociosexuality were evaluated: sex ratio theory (Peder-

sen 1991), developmental-attachment theory (Belsky et al.

1991; Chisholm 1996), and strategic pluralism theory

(Gangestad & Simpson 2000).

3.1. Sex ratio theory

Operational sex ratio can be defined as the relative balance

of marriage-age men versus marriage-age women in the lo-

cal mating pool (Pedersen 1991; Secord 1983), although

other formulations have been proposed (Clutton-Brock &

Parker 1992; Hardy 2002; Parker & Simmons 1996). When

computing operational sex ratios, marriage age is usually

treated as between 15 and 49 years (Guttentag & Secord

1983). Sex ratios are considered high when the number of

men significantly outsizes the number of women in a local

culture. Conversely, sex ratios are considered low when

there are relatively more women than men in the mating

market. In most cultures women tend to slightly outnum-

ber men, largely because of men’s greater mortality rate

(Daly & Wilson 1988). Nevertheless, significant variation

often exists in sex ratios across cultures and within cultures

when viewed over historical time (Guttentag & Secord

1983; Lazarus 2002).

Pedersen (1991) argued that a combination of sexual se-

lection theory (Darwin 1871) and parental investment the-

ory (Trivers 1972) leads to a series of predictions concern-

ing the effects of sex ratios on human mating strategies.

According to sexual selection theory, when males desire a

particular attribute in potential mating partners, females of

that species tend to respond by competing in the expression

and provision of that desired attribute. Among humans,

Pedersen had the insight that when sex ratios are especially

low (i.e., there are many more women than men), men be-

come an especially scarce resource that women must com-

pete for with even more intensity than normal (see also

Guttentag & Secord 1983). When combined with the

parental investment notion described earlier in which men

tend to desire promiscuous sex (Buss & Schmitt 1993;

Symons 1979; Trivers 1972), this leads to the following hy-

pothesis: Cultures with lower sex ratios (i.e., more women

than men) should possess higher levels of sociosexuality

(i.e., more promiscuity). The logic of Pedersen’s theory is

that in cultures with many more women than men, men are

scarce and can afford to demand from interested women

that men’s greater desires for promiscuous sex be fulfilled.

As a result of these mating market forces, the culture as a

whole should become more unrestricted in sociosexual ori-

entation (see also Barber 2000; Ember 1974).

Conversely, when sex ratios are high and men greatly

outnumber women, men must enter into more intense

competition for the limited number of potential female

partners (Bateman 1948). Women’s preferences for long-

term monogamous relationships become the key desires

that must be responded to if men are to remain competi-

tive in the courtship marketplace. In this case, Pedersen’s

(1991) logic suggests that cultures with higher sex ratios

(i.e., more men than women) should possess lower levels of

sociosexuality (i.e., should be more monogamous). In this

article, Pedersen’s series of insights will be referred to as

“sex ratio theory.”

Using data from sex ratio fluctuations over time within

the United States, Pedersen (1991) marshaled a compelling

case for a causal link between sex ratios and human mating

strategies (see also Guttentag & Secord 1983). For exam-

ple, high sex ratio fluctuations have been historically asso-

ciated with increases in monogamy, as evidenced by lower

divorce rates and men’s greater willingness to invest in their

children. Low sex ratios have been historically associated

Schmitt: Sociosexuality from Argentina to Zimbabwe

250 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

with indexes of promiscuity, such as an increase in divorce

rates and a reduction in what he termed female “sexual coy-

ness.” National sex ratios were related to sociosexuality

across the 48 nations of the ISDP, enabling sex ratio theory

to be evaluated from a cross-cultural perspective.

3.2. Developmental-attachment theory

Several combinations of life history theory (Low 1998) and

attachment theory (Bowlby 1969) have suggested that cer-

tain critical experiences during childhood play a role in the

development of human mating strategies (Belsky 1999;

Draper & Harpending 1988; MacDonald 1997). Perhaps

most prominent among these is a lifespan model developed

by Belsky et al. (1991). According to this model, early social

experiences adaptively channel children down one of two

reproductive pathways. Children who are socially exposed

to high levels of stress – especially insensitive or inconsis-

tent parenting, harsh physical environments, and economic

hardship – tend to develop insecure attachment styles.

These children also tend to physically mature earlier than

those children who are exposed to less stress. According to

Belsky and his colleagues, attachment insecurity and early

physical maturity subsequently lead to the evolutionary-

adaptive development of what is called an “opportunistic”

reproductive strategy in adulthood (i.e., unrestricted socio-

sexuality). An opportunistic strategy, it is argued, will lead

to higher levels of fitness in high-stress reproductive envi-

ronments. In cultures with inconsistent or stressful social

relations, therefore, children adaptively respond by devel-

oping the more viable reproductive strategy of unrestricted

sociosexuality.

Conversely, those children exposed to lower levels of

stress and less environmental hardship tend to be more

emotionally secure and to physically mature later. These

children are thought to develop a more “investing” repro-

ductive strategy in adulthood (i.e., restricted sociosexuality)

that pays higher evolutionary dividends in low-stress envi-

ronments. All children come equipped with the potential

for unrestricted or restricted sociosexuality, in this view, and

psychological adaptations that are sensitive to local envi-

ronments influence sociosexual desires and behaviors in

adaptive ways. Although the causal mechanisms that influ-

ence sociosexuality are most prominently located within the

family, this model also suggests that certain aspects of cul-

ture may be related to sociosexual variation. Namely, this

model leads to the following hypothesis: In cultures where

families are under more stress and have fewer resources, so-

ciosexual levels should be higher than in cultures with lower

stress and ample resources.

A closely related theory has been proposed by Chisholm

(1996; 1999a). Chisholm argues that local mortality rates –

presumably related to high stress and inadequate resources

– act as cues that contingently shift human mating strate-

gies in evolutionary-adaptive ways (see also Weinrich 1977).

In cultures with high mortality rates and unpredictable re-

sources, the optimal mating strategy is to reproduce early

and often, a strategy related to insecure attachment, short-

term temporal orientations, and unrestricted sociosexuality

(Chisholm 1999b). In cultures that are physically safe and

have abundant resources, mortality rates are lower and the

optimal strategy is to invest heavily in fewer numbers of off-

spring. In safer environments, therefore, one should pur-

sue a long-term mating strategy associated with more re-

stricted sociosexuality. This theory leads to the following

basic hypothesis: Cultures with higher mortality rates, ear-

lier reproduction, and more prolific reproduction should

have higher levels of sociosexuality than cultures with low

mortality, late reproduction, and limited reproduction. Col-

lectively, the Belsky et al. (1991) and Chisholm (1996;

1999a) theories will be referred to as a “developmental-at-

tachment theory” of sociosexuality. To test this theory, var-

ious indexes of familial stress, economic resources, mortal-

ity, and fertility were related to sociosexuality across the 48

nations of the ISDP.

3.3. Strategic pluralism theory

In direct contrast to developmental-attachment theory,

Gangestad and Simpson (2000) have proposed strategic

pluralism theory. According to strategic pluralism theory,

humans possess a menu of alternative mating strategies that

they can follow (see also Buss & Schmitt 1993; Gross 1996;

Simpson & Orina 2003; Thiessen 1994). Which strategy in-

dividuals follow depends on the condition of local environ-

ments. When local environments are demanding and the

difficulties of rearing offspring are high, the adaptive need

for biparental care increases. Because both men and

women are needed to successfully raise viable offspring in

more demanding environments, Gangestad and Simpson

argue that the importance of fidelity and heavy family in-

vestment should correspondingly increase: “In environ-

ments where male parenting qualities are needed and val-

ued, women should be less likely to engage in short-term

mating and extra-pair mating. In response to this, men

should devote greater effort to parental investment”

(p. 585). If true, this would suggest the following hypothe-

sis: In cultures with more demanding environments (e.g.,

higher stress, fewer resources, higher mortality), sociosex-

ual levels should be lower (i.e., people should be more

monogamous).

Conversely, in cultures where biparental care is less nec-

essary for successful child-rearing, Gangestad and Simpson

(2000) expect that monogamy would be less prevalent.

They postulate that in cultures with lower stress and ade-

quate resources, human psychological adaptations should

facultatively cause sociosexuality to increase (i.e., people

should be more promiscuous). Gangestad and Simpson rea-

son that in ancestral environments when biparental care

was not as crucial, men could have afforded to channel

more of their reproductive effort into short-term mating

and unrestricted sociosexuality. Women also could have

benefited from short-term mating (via access to high-qual-

ity genes; Gangestad 2001) given the collateral reduction in

their dependence on a long-term male’s resources and in-

vestment. In this study, various indexes of environmental

demand (e.g., life expectancy, gross domestic product per

capita, human development) and reproductive difficulty

(e.g., low birth weights, child malnutrition, infant mortal-

ity) were related to sociosexuality across the 48 nations of

the ISDP.

4. Culture and sex differences in sociosexuality

As noted earlier, it follows from most evolutionary theories

of human sexuality anchored in the theory of parental in-

vestment that men should score higher than women on so-

ciosexuality (Buss & Schmitt 1993; Symons 1979; Trivers

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 251

1972). However, there may be certain aspects of culture

that influence our evolved psychology in ways that accen-

tuate or attenuate sex differences in sociosexuality. Just as

the degree of sexual differentiation in body size is influ-

enced by local diet and altitude (Gaulin & Boster 1985; Ju-

rmain et al. 2000; Wolfe & Gray 1982), the degree of sexual

differentiation in sociosexuality may vary with local ecolog-

ical conditions. At times, this variability may be functional

and reflect psychological adaptations specifically designed

to moderate sex differences in response to particular eco-

logical conditions.

4.1. Strategic pluralism theory

An implication of strategic pluralism theory (Gangestad &

Simpson 2000) is that women’s sociosexuality should be

more sensitive than men’s to the demands and stressors of

local environments. In demanding environments that re-

quire biparental care, women’s sociosexuality facultatively

shifts, and they become much more restricted. Only some

men react to women’s sociosexual shifts, however, and be-

come restricted themselves. Other, more robust men were

“able to carry out short-term tactics successfully at all times,

regardless of the environmental factors to which women

were responding” (Gangestad & Simpson 2000, p. 586).

Therefore, women – as a group – should be more respon-

sive than men are to environmental influences on sociosex-

uality.

Several findings would seem to confirm the notion that

women’s sexuality is more responsive to environmental fac-

tors. For example, Barry and Schlegel (1984) examined the

186 preindustrial societies of the Standard Cross-Cultural

Sample and found on nearly all measures of sexual behav-

ior that adolescent women were more variable than adoles-

cent men. If sexual behaviors are adaptively responsive to

local ecological conditions in natural environments, there-

fore, the responsiveness appears to be greater for women.

Baumeister (2000), in a massive review of the literature on

sex and sexuality, found that the effects of acculturation, ed-

ucation, politics, religion, and family life on sexual attitudes

and behaviors were all more potent among women than

men. He concluded that “men’s sexuality revolves around

physical factors, in which nature is predominant and the so-

cial and cultural dimension is secondary. For women, social

and cultural factors play a much larger role” (Baumeister

2000, p. 368).

Similarly, strategic pluralism theory postulates that

women’s sociosexuality should be highly dependent on en-

vironmental demands across cultures, but men’s sociosexu-

ality should be less correlated with environmental harsh-

ness. If true, this leads to the following hypothesis: The

demanding nature of local environments should be more

closely correlated with women’s sociosexuality than with

men’s. Moreover, because men tend to be more oriented to-

ward short-term mating in general (Buss & Schmitt 1993;

Schmitt et al. 2002), the following hypothesis also can be

derived: The size or magnitude of the difference between

men and women should be smaller in nondemanding envi-

ronments. This is because in nondemanding environments

women’s sociosexual shifts bring them closer to men’s nor-

mally higher levels of unrestricted sociosexuality. Men may

shift as well, but their overall average on sociosexuality will

not shift as prominently. These hypotheses were evaluated

by correlating various indicators of environmental demands

with men’s sociosexuality, women’s sociosexuality, and the

effect size of sex differences in sociosexuality across cul-

tures.

4.2. Social structural theory

Even if sex differences in the willingness to have uncom-

mitted sex were found to be culturally universal, the differ-

ences may not result from adaptations to sociosexuality per

se. It could be the case that sex differences in sociosexual-

ity are a side effect of some other evolved sex difference –

such as sex differences in physical size (Gaulin & Boster

1985), sex differences in general sex drive (Baumeister et

al. 2001), or perhaps the external location of human male

genitalia (Gagnon & Simon 1973). It also could be that cer-

tain sexual trends pervade all cultures because of sociohis-

torical factors that are relatively unrelated to the evolution-

ary biology of men and women (Harris 1993; MacKinnon

1988).

According to the social structural or “biosocial” theory of

Eagly and Wood (1999; see also Wood & Eagly 2002), the

minds of men and women are not likely to contain sex-dif-

ferentiated adaptations that are specifically designed to

produce universal sex differences in sexuality per se. In-

stead, Eagly and Wood assume that “differences in the

minds of men and women arise primarily from experience

and socialization” (p. 414). Thus, when men and women ap-

pear to differ, it is because they have received dissimilar so-

cialization throughout development – particularly those ex-

periences and expectations associated with a society’s

bifurcated sex roles and manifest degree of patriarchy (Buss

& Barnes 1986; Eagly 1987; LaFrance et al. 2003; Maccoby

1998; Reiss 1986).

Eagly and Wood’s (1999) social structural account is still

an evolutionary theory of human mating, in that it views

men’s evolutionary history as hunters and meat providers

(among other selective factors) as having led to men’s

greater size, strength, and speed. In contrast, women’s evo-

lutionary history of giving birth and prolonged lactation,

among other selective factors, are thought to have led to

women’s prominence in child rearing. These and other

evolved physical features of men and women, it is argued,

ultimately lead to divisions of labor (see Joseph 2000) and,

in socioeconomically complex societies, to patriarchal as-

pects of culture (see also Lerner 1986; Smuts 1995). Ac-

cording to Eagly and Wood (1999), it is these divisions of la-

bor and the regular emergence of patriarchy (including

political, economic, and sexual forms of controlling women)

that, in turn, give rise to sex role ideologies and social ex-

pectations that are the more proximate causes of psycho-

logical differences between the sexes.

Wood and Eagly (2002) recently extended this line of

reasoning and offered a compelling rationale for why some

cultures have more bifurcated or “traditional” sex role ide-

ologies, whereas other cultures have more flexible or “pro-

gressive” ideologies. They argue that in some cultures the

value of men’s hunting skills, their ability to wage war, and

the need for women to stay close to children, among other

features of culture, are especially acute. In these cultures,

the local ecological and social conditions give rise to an

economy that favors men’s skills of production and, as a re-

sult, provides men with social and political power that of-

ten culminates in patriarchy and more traditional sex role

ideologies. Polygyny and warfare are frequently – though

Schmitt: Sociosexuality from Argentina to Zimbabwe

252 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

not always – associated with this cluster of cultural attri-

butes (Divale & Harris 1976; White & Burton 1988), and

the advent of agriculture, industrialization, and greater cul-

tural complexity may further exacerbate this more tradi-

tional form of sex role socialization (Korotayev & Kazankov

2003; Wood & Eagly 2002).

In many other cultures, however, women contribute a

relatively greater proportion of calories to the local diet,

have greater resource control and political power (e.g., as a

consequence of matrilocal residence and matrilineal de-

scent), and have greater reproductive freedom through in-

creased contraception, the benefit of nursemaids, and other

factors (Barry & Yoder 2002; Murphy 2003; Whyte 1978).

In these cultures, the local ecological and social conditions

give rise to an economy that favors women’s skills of pro-

duction more heavily and ultimately culminates in a certain

degree of gender egalitarianism and more progressive sex

role socialization. In essence, Wood and Eagly (2002) argue

that humans have evolved to be exquisitely sensitive to lo-

cal economic and sociopolitical circumstances and respond

by varying the degree of sex role socialization in ways that

may influence sex differences in human mating psychology.

The existence of sex roles is still an evolved feature of hu-

man psychology from this perspective (see also Alexander

2003), but the degree of disparity in sex role socialization

(and the degree of patriarchy) can vary in important and

systematic ways across cultures.

From this social structural perspective, sex differences in

sociosexuality – when they do exist – ultimately result from

evolved features of human psychology that sometimes lead

to patriarchy and sexual divisions of labor. More proxi-

mately, this perspective views sex differences in sociosexu-

ality as flowing from the disparate sex role socialization that

results from patriarchy and divisions of labor (Eagly &

Wood 1999), “sex differences in social behavior arise from

the distribution of men and women into social roles within

a society” (Wood & Eagly 2002, p. 701). This social struc-

tural perspective can be used to generate the following hy-

pothesis: In cultures with traditional sex role ideologies

(where women are more constrained in terms of economics,

politics, and reproductive freedom), sex differences in so-

ciosexuality should be larger. Again, this is because men

and women have experienced bifurcated sex roles and so-

cial constraints in these societies, with women experiencing

sociosexually restricted sex roles and patriarchal social con-

straints (Buss & Barnes 1986; Sprecher et al. 1987). Within

cultures that possess more progressive sex role ideologies –

where women have more access to money, power, and the

ability to make their own reproductive decisions – women

are allowed to explore a wider array of social roles. Indeed,

both men and women enjoy less burdensome and gender-

constraining social structures in cultures with modern sex

role ideologies (Williams & Best 1990), and “when men and

women occupy the same specific social role, sex differences

would tend to erode” (Eagly & Wood 1999, p. 413). Thus,

sex differences in sociosexuality should be smaller, or per-

haps even absent, in cultures with more progressive sex role

ideologies (where women have more equitable amounts of

economic, political, and reproductive freedom).

It is important to note that the primary objective of so-

cial structural theory was to explain the origins of sex dif-

ferences in human mate preferences, not sex differences in

sociosexual mating strategies per se (Eagly & Wood 1999;

Johannesen-Schmidt & Eagly 2002; Wood & Eagly 2002).

Nevertheless, the founding logic of social structural theory

clearly leads to the preceding predictions, with progressive

sex role cultures expected to exhibit smaller sex differences

than traditional sex role cultures. It also should be noted

that some of these predictions were first proposed over 15

years ago, in what was termed the structural powerlessness

hypothesis (Buss & Barnes 1986). For example, in the con-

text of mate preferences, Buss and Barnes (1986) predicted

that “men and women who have been subjected to less tra-

ditional sex role socialization will not show this [mate pref-

erence] sex difference as strongly as will those raised more

traditionally” (p. 569), and “sex differences in [mate] pref-

erences should diminish as the power balance in society ap-

proaches equity between sexes” (p. 569). Others have used

similar theorizing to make predictions concerning women’s

relative status and the degree of sexual differentiation in a

culture (Kasser & Sharma 1999), though not always with

supportive results (Fletcher 2002). In the current study, so-

cial structural theory was evaluated by correlating various

indicators of gender equality (political and economic), re-

productive freedom, and sex role ideology with the magni-

tude of sex differences in sociosexuality across cultures.

5. Method

5.1. Samples

The research reported in this target article is a result of the

International Sexuality Description Project (ISDP; Schmitt

et al. 2003b), a collaborative effort of over 100 social, be-

havioral, and biological scientists. The full range of the

ISDP originally comprised 56 nations. However, in eight of

these nations either the SOI was not administered (i.e., In-

dia, Jordan, and South Africa), or too few participants fully

completed the SOI (i.e., fewer than 25 men or fewer than

25 women; Chile, Cyprus, Indonesia, Malaysia, and Tanza-

nia). At least 25 men and 25 women were needed to achieve

the necessary statistical power for evaluating sex differ-

ences in sociosexuality (when setting b  .80, a  .05, and

looking for an effect moderate to large in size; Cohen 1988).

As seen in Table 1, a total of 48 nations from the ISDP

were used in the present analyses. Three nations were sam-

pled from North America. The Canadian national sample

included three independent, English-speaking subsamples

from the Canadian provinces of Ontario, Alberta, and

British Columbia, and one French-speaking subsample

from Quebec. The French-speaking participants were ad-

ministered the ISDP survey as translated/back-translated

into French. The translation and back-translation proce-

dures will be addressed later. Thirteen subsamples were ob-

tained from the United States. This included at least one

subsample from the states of New York, Illinois, Kentucky,

South Carolina, Florida, Alabama, Texas, New Mexico,

Idaho, California, and Hawaii. The subsamples from main-

land United States consisted of 66% European-American

(non-Hispanic), 10% African-American, 8% Hispanic-Amer-

ican, 5% Asian-American, 2% Native-American, and 9%

who either identified themselves as “other” or did not spec-

ify their origin. The North American world region was also

represented by a sample from Mexico, mainly general com-

munity members who volunteered for the study. Commu-

nity samples in the ISDP tended to be related to colleges

and universities (e.g., many were employed by the local ed-

ucational institutions), and so should not be considered as

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 253

completely independent of the college-related limitations

of most ISDP national samples.

Four nations were sampled from the world region of

South America, including Peru, Bolivia, Argentina, and

Brazil. Eight nations from Western Europe were sam-

pled as part of the ISDP, including one sample each from

Finland, the Netherlands, Belgium (Flanders region),

France, and Switzerland (German-speaking region). Mul-

tiple subsamples were collected from the United Kingdom

(including Northern Ireland and multiple England sam-

ples), Germany, and Austria. The subsamples from England,

Germany, and Austria included both college students and

general community members. Eleven nations from Eastern

Europe were sampled in the ISDP: one sample each from

Estonia, Latvia, Lithuania, Poland, the Czech Republic,

Slovakia, Ukraine, Romania, Serbia, Croatia, and Slovenia.

The ISDP had five nations sampled to represent the world

region of Southern Europe: Portugal, Spain, Italy, Malta,

and Greece.

Three national samples from the Middle East world re-

gion were included in the ISDP: Turkey, Lebanon, and Is-

rael. Five nations from Africa were sampled as part of the

ISDP, including college students from Botswana, the Dem-

ocratic Republic of the Congo, Ethiopia, Morocco, and

Zimbabwe. Three nations from Oceania were sampled for

the ISDP; they included two subsamples from Australia

(one from eastern Australia containing college students and

one from western Australia that included both college stu-

dents and community members), one sample from New

Zealand, and one sample from Fiji. Two nations from

South/Southeast Asia were part of the ISDP, including na-

tional samples from Bangladesh and the Philippines. Four

national samples from East Asia were included: one sample

each from Hong Kong (now a part of the People’s Repub-

lic of China), Taiwan (Republic of China), and Japan; and

two subsamples were accumulated from the Republic of

(South) Korea.

Overall, this collection of national samples represented a

diverse array of ethnic, geographic, and linguistic cate-

gories. In total, SOI scores from the ISDP represent 6 con-

tinents, 10 islands (Malta, Fiji, New Zealand, the Philip-

pines, Hong Kong, Japan, Taiwan, Hawaii, Ireland, and

Britain), 26 languages, and 48 nations (see Table 1). Most

samples were recruited as volunteers, some received course

credit for participation and others received a small mone-

tary reward for their participation. All samples were ad-

ministered an anonymous self-report survey, most surveys

were returned via sealed envelope and/or the usage of a

drop-box. Return rates for college student samples tended

to be relatively high (around 95%), although this number

was lower in some cultures. Return rates for community

samples were around 50%. Further details on the sam-

pling and assessment procedures within each of the world

regions and national samples are provided elsewhere

(Schmitt et al. 2003a; 2003b) and are available from the au-

thor on request.

5.2. Procedure

All collaborators were asked to administer an anonymous

nine-page survey to at least 100 men and 100 women. Some

nations, such as the United States and Canada, contained

numerous convenience samples, and so the national sam-

ple size was much larger than 200. All participants were

provided with a brief description of the study, including the

following written instructions:

This questionnaire is entirely voluntary. All your responses will

be kept confidential and your personal identity will remain

anonymous. No identifying information is requested on this

survey, nor will any such information be added later to this sur-

vey. If any of the questions make you uncomfortable, feel free

not to answer them. You are free to withdraw from this study at

any time for any reason. This series of questionnaires should

take about 20 minutes to complete. Thank you for your partic-

ipation.

The full instructional set provided by each collaborator var-

ied, however, and was adapted to fit the specific culture and

type of sample. Details on incentives and cover stories used

across samples are available from the author.

5.3. Measures

5.3.1. Translation procedures.

Researchers from nations

where English was not the primary language were asked to

conduct a translation/back-translation procedure and ad-

minister the ISDP measures, including the SOI, in their na-

tive language. This process typically involved the primary

collaborator translating the measures into the native lan-

guage of the participants, and then having a second person

back-translate the measures into English. Differences be-

tween the original English and the back-translation were

discussed, and mutual agreements were made on the most

appropriate translation. This procedure tries to balance the

competing needs of making the translation meaningful and

naturally readable to the native participants while preserv-

ing the integrity of the original measure and its constructs

(Brislin 1980), and it is generally regarded as an “etic” ap-

proach to cross-cultural psychology (Church 2001).

As seen in Table 1, this process resulted in the survey be-

ing translated into 26 different languages. Samples from

Ethiopia, Fiji, Hong Kong, Morocco, and the Philippines

were administered the survey in English, but certain terms

and phrases were annotated to clarify what were thought to

be confusing words for the participants. The translation of

the ISDP survey into the Flemish dialect of Dutch used

only a translation procedure, because this involved mainly

word variant changes from the original Dutch. Finally, pi-

lot studies were conducted in several testing sites, in part to

clarify translation and comprehension concerns.

5.3.2. Demographic measure. Each sample was first pre-

sented with a demographic measure entitled “Confidential

Personal Information.” This measure included questions

about sex (male, female), age, sexual orientation (hetero-

sexual, homosexual, bisexual), current relationship status

(married, cohabiting, dating one person exclusively, not

currently involved with anyone), and current socioeco-

nomic status (upper, upper-middle, middle, lower-middle,

lower).

5.3.3. Sociosexual Orientation Inventory (SOI). The SOI

is a seven-item self-report survey designed to measure a

single strategic dimension – restricted versus unrestricted

sociosexuality (Simpson 1998; Simpson & Gangestad 1991).

The first three items of the SOI are intended to capture

overt behavioral expressions of sociosexual variation. Item

1 is: “With how many different partners have you had sex

(sexual intercourse) within the past year?” Item 2 is: “How

many different partners do you foresee yourself having sex

Schmitt: Sociosexuality from Argentina to Zimbabwe

254 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

with during the next five years? (Please give a specific, re-

alistic estimate).” Item 3 is: “With how many different part-

ners have you had sex on one and only one occasion?”

Open-ended blanks are provided after each of the first

three questions of the SOI. The fourth item was designed

to measure covert sociosexual behavior: “How often do

(did) you fantasize about having sex with someone other

than your current (most recent) dating partner?” This item

was followed by an 8-point scale, ranging from 1 (never) to

8 (at least once a day).

Items 5, 6, and 7 were designed to measure sociosexual

attitudes. Item 5 is: “Sex without love is OK.” Item 6 is: “I

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 255

Table 1. Sample size and language of administration for men and women who completed the

Sociosexual Orientation Inventory across 48 nations of the International Sexuality Description Project

Nation Men Women Total Language

Argentina 110 136 246 Spanish

Australia 183 265 448 English

Austria 173 225 398 German

Bangladesh 73 59 132 Bangla

Belgium 129 285 414 Dutch (Flemish)

Bolivia 70 56 126 Spanish

Botswana 94 115 209 English

Brazil 39 49 88 Portuguese

Canada 335 626 961 English/French

Congo, D.R. 91 50 141 French

Croatia 101 100 201 Croatian

Czech Rep. 76 104 180 Czech

Estonia 61 84 145 Estonian

Ethiopia 107 68 175 English

Fiji 66 53 119 English

Finland 28 72 100 Finnish

France 47 54 101 French

Germany 229 379 608 German

Greece 39 154 193 Greek

Hong Kong 90 94 184 English

Israel 139 170 309 Hebrew

Italy 92 108 200 Italian

Japan 125 86 211 Japanese

Latvia 77 78 155 Latvian

Lebanon 106 120 226 English

Lithuania 40 38 78 Lithuanian

Malta 104 119 223 English

Mexico 90 100 190 Spanish

Morocco 60 74 134 English

Netherlands 94 111 205 Dutch

New Zealand 104 152 256 English

Peru 87 91 178 Spanish

Philippines 94 118 212 Englisha

Poland 214 381 595 Polish

Portugal 99 131 230 Portuguese

Romania 100 106 206 Romanian

Serbia 92 95 187 Serbian

Slovakia 55 70 125 Slovak

Slovenia 44 41 85 Slovenian

South Korea 189 289 478 Korean

Spain 81 157 238 Spanish

Switzerland 57 95 152 German

Taiwan 114 88 202 Mandarin

Turkey 190 188 378 Turkish

Ukraine 100 100 200 Ukrainian

United Kingdom 121 275 396 English

United States 948 1,707 2,655 English

Zimbabwe 96 90 186 English

Total ISDP sample 5,853 8,206 14,059 26 languages

Note:

 some English items were annotated for greater comprehension.

can imagine myself being comfortable and enjoying ‘casual’

sex with different partners.” Item 7 is: “I would have to be

closely attached to someone (both emotionally and psycho-

logically) before I could feel comfortable and fully enjoy

having sex with him or her.” All three attitudinal items were

followed by 9-point scales ranging from 1 (I strongly dis-

agree) to 9 (I strongly agree). Responses to item 7 are re-

verse-coded so that higher scores indicate higher or more

unrestricted sociosexuality.

According to Simpson and Gangestad (1991), items 5, 6,

and 7 are highly correlated and should be merged to form

a single “attitudinal” score. This attitudinal score is then

combined with the first four SOI items to form the total

SOI composite measure. However, each item of the SOI

composite measure is first weighted using the following for-

mula: (5  item 1)  (1  item 2 [with a cap on item 2 of

30])  (5  item 3)  (4  item 4)  (2  mean of items

5, 6, and 7)  total SOI composite measure. Again, using

this formula produces an SOI composite measure such that

higher scores are associated with unrestricted sociosexual-

ity. Higher, unrestricted individuals tend to have had sex

with more partners in the previous year (item 1), foresee

having sex with more partners in the next five years (item

2), engage in more one-night stands (item 3), fantasize

more about having sex with someone other than their cur-

rent dating partner (item 4), and possess more permissive

attitudes toward uncommitted sexual relations (items 5, 6,

and 7).

5.3.4. Time Known measure. The Time Known measure

originally used by Buss and Schmitt (1993) was adapted for

use in the ISDP. The Time Known measure asked partici-

pants to rate on a 6-point scale ranging from 3 (definitely

yes) to –3 (definitely not) the degree to which “If the con-

ditions were right, would you consider having sexual inter-

course with someone you viewed as desirable if . . .” they

had known that person for varying amounts of time ranging

from five years to one hour.

5.3.5. Mate Poaching Inventory. The Mate Poaching In-

ventory originally used by Schmitt and Buss (2001) was

adapted for use in the ISDP. Two items were of relevance

to the present study. First, the Mate Poaching Inventory

asked participants to rate on a 7-point frequency scale rang-

ing from 1 (never) to 4 (sometimes) to 7 (always) the degree

to which “Have you ever attempted to attract someone who

was already in a relationship with someone else for a short-

term sexual relationship with you?” Participants from

Lebanon and Poland received a version of this measure in

which they were asked if they had attempted to attract away

an already-mated partner for a new long-term mating rela-

tionship. A second item asked, “While you were in a ro-

mantic relationship, if others attempted to obtain you as a

short-term sexual partner, how successful have they been (if

others have never tried, skip this question)?” Participants

rated this item on a 7-point scale ranging from 1 (not at all

successful) to 4 (moderately successful) to 7 (very success-

ful). Again, Lebanese and Polish samples received long-

term versions of this mate poaching question.

5.3.6. Archival measures. Several archival data sets were

used in this article. National sex ratios were obtained from

the United Nations Statistics Division (2001). Three data

sets were used to evaluate the cross-cultural convergent va-

lidity of national sociosexuality scores. These include data

from the World Values Survey (Inglehart et al. 1998), data

from the International Social Survey Program (Widmer et

al. 1998), and data from the Global Sex Survey (SSL Inter-

national 2001).

Several variables were used to evaluate the level of envi-

ronmental demand. The percentage of low-birth-weight in-

fants and the prevalence of child malnutrition were from

the UNICEF Global Database (United Nations Population

Division 2001). Infant mortality rates were obtained from

the United Nations Statistics Division (2001). Teen preg-

nancy rates and fertility rates were obtained from United

Nations Development Programme (2001). In all cases

higher scores on these variables indicate higher levels of en-

vironmental demand.

The mean age at marriage for women was obtained from

the World’s Women 2000 Report (United Nations Statistics

Division 2001). Life expectancy and gross domestic prod-

uct per capita were obtained from the United Nations De-

velopment Programme (2001). The Human Development

Index, as reported in the United Nations Development

Programme (2001), is defined as the achievement of a na-

tion in basic human capabilities, including health, longevity,

education, and a decent standard of living. Data on human

development were obtained from the United Nations De-

velopment Programme (2001). For these variables, higher

scores are indicative of lower levels of environmental de-

mand.

Several variables were used to evaluate the level of polit-

ical and economic gender equality across cultures. The

Gender Empowerment Measure – a United Nations statis-

tic based on the level of political and economic equality of

men and women within a nation – was obtained from the

United Nations Human Development Programme (2001).

The percentage of women in parliament and the percent-

age of women in ministerial positions were obtained from

the United Nations Statistics Division (2001). The Gender-

Related Development Index reflects the degree to which

men and women differ in the achievement of basic human

capabilities, including health, longevity, education, and a

decent standard of living and was obtained from the United

Nations Human Development Programme (2001). The

Women’s Wage Equality measure was obtained from the

World’s Women 2000 Report (United Nations Statistics Di-

vision 2001). In all cases, higher scores on these measures

indicate a greater degree of political and economic gender

equality.

Several variables were used to evaluate the level of rela-

tional and reproductive freedom of women across cultures.

The percentage of women-headed households and the per-

centage of women in unions who use contraception were

obtained from the World’s Women 2000 Report (United

Nations Statistics Division 2001). The divorce rate across

cultures was obtained from the United Nations Human De-

velopment Programme (2001). In all cases, higher scores

these measures indicate a greater degree of relational and

reproductive freedom for women.

Several variables were used to evaluate the degree of tra-

ditional versus progressive sex role ideology across cultures.

Direct measures of women’s and men’s sex role ideologies

were obtained from Williams and Best (1990). High scores

on the Sex Role Ideology measure (Williams & Best 1990)

indicate more progressive views on the roles of men and

women in society; low scores indicate more traditional

Schmitt: Sociosexuality from Argentina to Zimbabwe

256 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

views. Direct measures of hostile sexism were obtained

from Glick et al. (2000). High levels on the Hostile Sexism

measure reflect more negative attitudes toward women and

may be indicative of greater cultural patriarchy (Sakalli-

Ugurlu & Beydogan 2002). An index of cultural masculin-

ity (i.e., more traditional beliefs about women’s roles in the

family, the workplace, and society) was obtained from Hof-

stede’s (2001) classic IBM study of attitudes and values.

6. Results

6.1. Does the SOI psychometrically measure a single

dimension within cultures?

Because the SOI contains several open-ended items, it is

somewhat susceptible to extreme scores. In the present

study, the upper 1% of full-scale scores were eliminated

from further analyses (i.e., scale scores above 180; see

Rosenthal & Rosnow 1991). The extreme scores were dis-

persed evenly across world regions, and most extreme scor-

ers were men.

To evaluate whether sociosexuality consists of one basic

dimension, all seven items of the SOI were subjected to a

principal axis factor analysis within all nations of the ISDP.

In their original validation research using a sample from the

United States, Simpson and Gangestad (1991) reported

that SOI items tended to form a single factor, and that the

first unrotated factor of the SOI accounted for 39.2% of the

variance. As seen down the first data column of Table 2,

similar levels of variance were accounted for within each of

the 48 ISDP nations, and 44.50% of the variance was ac-

counted when analyzing the entire sample. In addition, a

single factor was suggested for most nations based on eigen-

values above 1 and according to the scree criterion (Cattell

1966). When additional factors were suggested, these sec-

ondary factors typically provided less than 10% of addi-

tional variance to the factor solution. In sum, it was a cul-

tural universal for responses to the SOI to form a single

factor, at least across the different forms of culture repre-

sented in the ISDP.

6.2. Is the SOI psychometrically reliable within

cultures?

The internal reliability of the SOI was evaluated across the

48 nations of the ISDP in two ways. First, raw scores on the

seven individual items of the SOI were evaluated according

to Cronbach’s alpha coefficient. As seen down the second

data column of Table 2, the level of internal reliability was

adequate across most nations. The levels were somewhat

lower than would be expected in the Democratic Republic

of the Congo, Mexico, and Slovakia. However, in most cases

these levels increased when the second method was used to

evaluate internal reliability. The overall level of alpha relia-

bility across all participants of the ISDP was 0.77.

The third data column of Table 2 contains the alpha co-

efficients of the five weighted items composing the SOI

scale. As noted earlier, items 5, 6, and 7 of the SOI are com-

bined when computing composite SOI scores, and all items

are further weighted according to a specific formula (see

Simpson 1998; Simpson & Gangestad 1991). These

weighted internal reliabilities were also adequate across

most cultures. Data from the Democratic Republic of the

Congo were still problematic (a  0.19). The overall level

of weighted reliability across all participants was 0.65, ap-

proaching the level reported originally by Simpson and

Gangestad (1991; a  0.73).

6.3. Is the SOI psychometrically valid within cultures?

To evaluate the validity of the SOI within each of the 48 na-

tions, scores from two other measures included in the ISDP

were used. First, scores from the Time Known measure de-

veloped by Buss and Schmitt (1993) were used to evaluate

within-culture convergent validity. One of the Time Known

items asks, “If the conditions were right, would you con-

sider having sexual intercourse with someone you viewed as

desirable if you had known that person for 1 month?” The

partial correlation (controlling for sex) between consenting

to sex after one month and responses to the SOI within al-

most every nation was significant and positive. Exceptions

included the Democratic Republic of the Congo, Ethiopia,

and Ukraine. The sample from Bangladesh was not admin-

istered the Time Known measure.

Another avenue for evaluating within-culture convergent

validity was to compare the SOI results with responses to

the Mate Poaching Inventory developed by Schmitt and

Buss (2001). This measure asked two questions highly rel-

evant to the SOI. The first question asks how frequently the

participant has attempted in the past to poach (i.e., roman-

tically attract) another person’s romantic partner. As seen in

Table 2, the correlation between the frequency of mate

poaching attempts and sociosexuality was positive in almost

every culture. The only exception to this trend was the

Ukraine. Notably, the convergent validity correlation within

the Democratic Republic of the Congo was significant,

r(92) 0.30, p  .01.

A second question from the Mate Poaching Inventory

asks whether the person had ever been poached away from

a past partner (i.e., had ever been induced to be unfaithful).

It would be expected that those who have an unrestricted

sociosexuality would be more likely to go along with a mate

poach. As seen in the last column of Table 2, correlations

between sociosexuality and going along with a poaching at-

tempt on oneself were positive and significant in almost

every culture. Overall, the within-culture construct validity

evidence of the SOI was highly favorable.

Additional analyses reported elsewhere indicate that so-

ciosexuality is reliably associated with certain physical attri-

butes across cultures (Schmitt 2002; 2003). For example,

men’s self-ratings of physical attractiveness are cross-cul-

turally correlated with unrestricted sociosexuality – a find-

ing that confirms portions of strategic pluralism theory

(Gangestad & Simpson 2000). Among women, this rela-

tionship is less robust across cultures (Schmitt 2002). So-

ciosexuality also appears related to facial symmetry in men,

an attribute closely linked to attractiveness (Schmitt 2003).

In addition, within nearly every major region of the world,

taller women report more promiscuous thoughts, feelings,

and behaviors than shorter women (Schmitt 2003b). This

finding may relate to sex ratio theory (Pedersen 1991), in

that men’s preference for shorter women (Ellis 1992) may

cause taller women to have fewer options in the mating

marketplace. Taller women, as a result, may be forced to

fulfill men’s desires for promiscuous sex to garner a mating

relationship. Alternately, taller women may possess higher

levels of testosterone, which in turn is linked with more un-

restricted sociosexuality (Clark 2004; Udry et al. 1986). Re-

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 257

Schmitt: Sociosexuality from Argentina to Zimbabwe

258 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

Table 2. Psychometric properties of the Sociosexual Orientation Inventory within 48 nations of the

International Sexuality Description Project

Factor Structure Internal Reliability Convergent Validity Correlations

% Variance Weighted Consent Attempted Went Along

Nation of First Factor Alpha Alpha to Sex a Poach with Poach

Argentina 45.36 0.75 0.66 0.39*** 0.40*** 0.48***

Australia 46.00 0.78 0.62 0.48*** 0.38*** 0.39***

Austria 45.12 0.77 0.58 0.37*** 0.40*** 0.43***

Bangladesh 36.49 0.61 0.60 — 0.19** 0.29***

Belgium 41.47 0.75 0.59 0.43*** 0.40*** 0.36***

Bolivia 45.31 0.71 0.69 0.27** 0.15* 0.02

Botswana 42.01 0.75 0.69 0.43*** 0.41*** 0.44***

Brazil 51.58 0.80 0.69 0.46*** 0.48*** 0.49***

Canada 42.92 0.76 0.64 0.43*** 0.36*** 0.31***

Congo, D.R. 24.11 0.36 0.19 0.14 0.30** 0.18*

Croatia 42.14 0.72 0.69 0.39*** 0.30*** 0.18**

Czech Rep. 43.88 0.74 0.61 0.28*** 0.48*** 0.50***

Estonia 47.66 0.76 0.56 0.39*** 0.51*** 0.35***

Ethiopia 31.96 0.52 0.36 0.14 0.27*** 0.31***

Fiji 37.68 0.69 0.63 0.17* ——

Finland 41.07 0.70 0.33 0.30** 0.42*** 0.30**

France 43.94 0.75 0.61 0.42*** 0.50*** 0.47***

Germany 46.01 0.78 0.57 0.39*** 0.32*** 0.40***

Greece 43.10 0.76 0.61 0.42*** 0.24*** 0.18**

Hong Kong 43.19 0.73 0.70 0.49*** 0.45*** 0.46***

Israel 43.52 0.76 0.62 0.44*** 0.45*** 0.36***

Italy 54.82 0.81 0.77 0.36*** 0.53*** 0.35***

Japan 38.69 0.70 0.62 0.29*** 0.42*** 0.41***

Latvia 42.00 0.70 0.63 0.40*** 0.35*** 0.48***

Lebanon 52.93 0.80 0.75 0.48*** 0.33*** 0.14*

Lithuania 39.23 0.66 0.62 0.27** 0.24* 0.30**

Malta 45.60 0.78 0.65 0.54*** 0.42*** 0.33***

Mexico 35.02 0.48 0.57 0.33*** 0.36*** 0.14

Morocco 51.93 0.77 0.74 0.34*** 0.17* 0.15

Netherlands 40.72 0.73 0.52 0.37*** 0.20** 0.08

New Zealand 49.54 0.80 0.65 0.44*** 0.33*** 0.23***

Peru 50.85 0.80 0.70 0.23*** 0.22** 0.23**

Philippines 50.46 0.79 0.73 0.34*** 0.47*** 0.32***

Poland 52.86 0.83 0.71 0.40*** 0.30*** 0.16***

Portugal 45.66 0.78 0.68 0.29*** 0.29*** 0.25***

Romania 52.59 0.79 0.78 0.43*** 0.41*** 0.44***

Serbia 45.88 0.72 0.72 0.41*** 0.53*** 0.58***

Slovakia 37.93 0.31 0.54 0.39*** 0.46*** 0.37***

Slovenia 48.83 0.82 0.71 0.22* 0.28** 0.12

South Korea 41.58 0.75 0.56 0.43*** 0.33*** 0.21**

Spain 47.92 0.81 0.71 0.50*** 0.39*** 0.29***

Switzerland 42.01 0.74 0.58 0.35*** 0.34*** 0.36***

Taiwan 40.75 0.73 0.63 0.43*** 0.56*** 0.62***

Turkey 50.59 0.80 0.63 0.39*** 0.39*** 0.34***

Ukraine 58.48 0.86 0.82 0.11 0.10 0.02

United Kingdom 49.71 0.80 0.68 0.50*** 0.41*** 0.28***

United States 49.63 0.82 0.67 0.50*** 0.37*** 0.29***

Zimbabwe 44.52 0.75 0.72 0.29*** 0.31*** 0.35***

Total ISDP sample 44.50 0.77 0.65 0.44*** 0.37*** 0.31***

Note: ISDP  International Sexuality Description Project. *  p  .05, **  p  .01, ***  p  .001, —  measure not adminis-

tered.

gardless, the cross-cultural consistency of these and other

anthropometric findings suggests that self-reported re-

sponses to the SOI represent more than simple response

styles associated with sexual self-presentation. Instead, the

SOI is apparently tapping aspects of human mating objec-

tively connected to theoretically relevant physical attributes

(Gangestad 2001; Schmitt 2002; 2003), and it does so in ro-

bust ways across a broad range of human cultures.

6.4. Is the SOI psychometrically valid across cultures?

Table 3 contains the estimated means and standard devia-

tions of sociosexuality across the 48 nations of the ISDP. Es-

timated means were obtained using a factorial ANCOVA

with sociosexuality as the dependent variable, nation as the

independent variable, and sex of participant as a covariate.

Across all 48 nations, the correlation between men’s and

women’s mean levels of sociosexuality was significant, r(46)

0.56, p  .001. Sex was entered as a covariate because

some samples contained more men than women, whereas

others contained more women than men, and it was ex-

pected that sex would have a significant within-nation asso-

ciation with sociosexuality. The estimated means in Table 3,

therefore, represent the overall national level of sociosexu-

ality within each of the ISDP samples after controlling for

the confounding effects of sex-linked sociosexual variabil-

ity.

To evaluate the validity of the national SOI profiles pre-

sented in Table 3, mean levels of sociosexuality were corre-

lated with other measures completed by ISDP samples. For

example, responses to the one month time interval of the

Time Known measure were used to compute national Time

Known averages for each nation (after controlling for sex

within each nation). These national averages significantly

correlated with national SOI scores, r(45) 0.79, p 

.001. Thus, as national SOI profiles increase, so do national

tendencies to be quick to consent to sex. Similarly, the na-

tion-level standard deviations of responses to the Time

Known measure were significantly correlated with the na-

tional standard deviations of the SOI, r(45) 0.35, p 

.05. This suggests that the tendency for people within a na-

tion to be clustered or scattered around the sexual average

of their nation is robust across sex-related measures.

Relating national SOI scores to mate-poaching experi-

ences also provided evidence that the national sociosexual-

ity averages in Table 3 were valid. For example, the corre-

lation between a nation’s SOI and a nation’s average

frequency of making mate poaching attempts was positive,

r(45) 0.54, p  .001. The correlation between a na-

tion’s SOI level and a nation’s average reporting of being

successfully poached away from a past partner was positive

as well, r(45) 0.47, p  .001. Overall, comparing re-

sponses available from within the ISDP database strongly

supported the validity of the nation-level SOI results.

A final avenue for evaluating the validity of SOI scores

presented in Table 3 was to compare nation-level sociosex-

uality with data from external sources. The World Values

Study (WVS; Inglehart et al. 1998) is based on representa-

tive samples from 43 countries, 27 of which overlap with the

nations of the ISDP (Argentina, Austria, Belgium, Brazil,

Canada, Czech Republic, Estonia, Finland, France, Ger-

many, Italy, Japan, Latvia, Lithuania, Mexico, Netherlands,

Poland, Portugal, Romania, Slovakia, Slovenia, South Ko-

rea, Spain, Switzerland, Turkey, United Kingdom, and the

United States). The WVS asked participants the extent to

which they agree or disagree with various statements, some

of which were potentially related to sociosexuality. The

statements used in the present analyses included whether

participants believed: Marital fidelity is relatively unimpor-

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 259

Table 3. Means and standard deviations of sociosexuality across

48 nations of the International Sexuality Description Project

(controlling for sex of participant)

Nation Mean SD

Argentina 40.74 28.38

Australia 37.29 23.87

Austria 45.73 31.23

Bangladesh 19.67 17.59

Belgium 32.82 18.81

Bolivia 40.90 32.87

Botswana 27.02 26.78

Brazil 37.93 31.51

Canada 34.52 22.58

Congo, D.R. 32.43 22.68

Croatia 42.98 26.46

Czech Rep. 37.52 24.01

Estonia 39.95 29.73

Ethiopia 26.55 25.63

Fiji 38.58 33.34

Finland 50.50 32.47

France 36.67 23.03

Germany 39.68 24.95

Greece 32.38 16.21

Hong Kong 22.90 16.85

Israel 40.95 26.98

Italy 34.37 26.75

Japan 24.10 18.51

Latvia 43.93 25.44

Lebanon 28.57 25.26

Lithuania 46.10 30.68

Malta 31.27 24.17

Mexico 35.69 23.29

Morocco 39.31 36.85

Netherlands 39.34 25.07

New Zealand 47.69 31.52

Peru 34.59 30.35

Philippines 32.10 28.58

Poland 34.21 25.39

Portugal 29.55 18.37

Romania 32.16 29.87

Serbia 38.72 24.08

Slovakia 34.90 24.55

Slovenia 46.26 25.71

South Korea 22.21 14.80

Spain 33.72 20.64

Switzerland 39.13 22.30

Taiwan 19.22 17.64

Turkey 36.06 31.38

Ukraine 32.27 27.03

United Kingdom 40.17 29.27

United States 37.05 25.77

Zimbabwe 22.66 26.07

Total ISDP sample 35.31 26.05

tant, individuals should have complete sexual freedom,

sometimes marital affairs are justified, sometimes prostitu-

tion is justified, and sometimes divorce is justified. These

five items were collapsed to form a WVS Sexual Permis-

siveness attitude scale. The Sexual Permissiveness scale had

adequate internal reliability (a  0.76). Importantly, the

correlation between national SOI scores from the ISDP

and the Sexual Permissiveness scores from the WVS was

statistically significant, r(24) 0.34, p  .05. This find-

ing provides external convergent validity for the nation-

level SOI scores of the ISDP.

The International Social Survey Program (Smith 1992;

Widmer et al. 1998) is based on representative samples

from 24 countries, 16 of which overlap with the nations of

the ISDP (Australia, Austria, Canada, Czech Republic,

Germany, Israel, Italy, Japan, Netherlands, New Zealand,

Philippines, Poland, Slovenia, Spain, United Kingdom, and

the United States). The percentage of people in each nation

who responded “not wrong at all” to the question, is “sex be-

fore marriage wrong?”, significantly correlated with na-

tional SOI scores in the ISDP, r(14) 0.69, p  .01. This

finding provided evidence of external convergent validity

for the nation-level SOI scores of the ISDP.

The Global Sex Survey (SSL International 2001) is based

on convenience samples from 28 countries, 20 of which

overlap with the nations of the ISDP (Australia, Canada,

Croatia, Czech Republic, France, Germany, Greece, Hong

Kong, Israel, Italy, Japan, Mexico, Netherlands, New

Zealand, Poland, Spain, Taiwan, Turkey, United Kingdom,

and the United States). The Global Sex Survey asked par-

ticipants at what age they started having sex, their lifetime

total number of sexual partners, and the frequency with

which they have sexual intercourse. These three items were

collapsed to form a Global Sex Survey index of unrestricted

sexuality. The Unrestricted Sexuality scale had adequate in-

ternal reliability (a  0.82). Importantly, the correlation be-

tween national SOI scores from the ISDP and the Unre-

stricted Sexuality scale of the Global Sex Survey was

statistically significant, r(18) 0.77, p  .001. This find-

ing provided external convergent validity for the nation-

level SOI scores of the ISDP.

The psychometric properties of the SOI appeared to be

adequate in cross-cultural perspective. Within nearly all

cultures, the SOI comprised a single dimension, was inter-

nally reliable, and demonstrated convergent validity. Across

cultures, national mean-level scores on the SOI also ap-

peared valid. National sociosexuality scores significantly

correlated with other sex-related measures from within the

ISDP and with external indexes of permissive or unre-

stricted sociosexual attitudes and behaviors. In total, this

pattern of results provides a reasonable foundation for mov-

ing to the next issue, the reasons why sociosexuality varies

across cultures.

6.5. Why do nations differ in sociosexuality?

From an evolutionary perspective, there may be several

reasons why cultures have different mating tendencies

(Frayser 1985; Low 2000; Marlowe 2003; Mealey 2000;

Pasternak et al. 1997). Hypotheses from three main theo-

ries concerning sociosexual variation across cultures were

tested here: sex ratio theory (Pedersen 1991), developmen-

tal-attachment theory (Belsky et al. 1991), and strategic

pluralism theory (Gangestad & Simpson 2000).

6.5.1. Sex ratio theory. According to the sex ratio theory

(Pedersen 1991), higher sex ratios (i.e., more men than

women) should be associated with lower sociosexuality (i.e.,

more monogamy). Data on sex ratio levels across the ISDP

were obtained from the United Nations (United Nations

Statistics Division 2001). As predicted, sex ratios were sig-

nificantly correlated with national sociosexuality levels in

the negative direction, r(46) 0.45, p  .001. As shown

in Figure 1, it appeared that much of this variation was

caused by the low sex ratios and high sociosexual levels ev-

ident in the Baltic nations of Estonia, Latvia, and Lithua-

nia. Low sex ratios in the Baltics are not surprising given the

high rates of male suicides and deaths from accidents within

these nations (e.g., Neumayer 2003). However, even with-

out the Baltic nations, the correlation between sex ratio and

sociosexuality was significant, r(43) 0.38, p  .01.

These findings are consistent with the view that cultures

with more women than men possess mating systems driv-

en, via the powers of sexual selection, by men’s evolved de-

sires for unrestricted, promiscuous sex. Figure 1 also shows

that cultures with more men than women (e.g., Hong Kong,

Bangladesh, and Taiwan) tend to be low on sociosexuality.

In these cultures, according to sex ratio theory, the mating

system is driven by women’s more potent desires for long-

term, monogamous mating (see also Guttentag & Secord

1983).

6.5.2. Developmental-attachment theory. According to

developmental-attachment theory, cultures with high fa-

milial stress, low economic resources, and high mortality

rates should possess higher levels of sociosexuality because

stressful sociocultural features lead children along a trajec-

tory of insecure attachment, early puberty, and short-term

mating (Belsky et al. 1991; Chisholm 1996). Table 4 con-

tains the intercorrelations of several sociocultural variables

that provide tests of the developmental-attachment per-

spective. For example, prevalence of low-birth-weight in-

fants, child malnutrition, and infant mortality can all be

seen as indexes of deleterious familial stress (Goldstein &

Peckham 1976; Power & Li 2000) and each should posi-

Schmitt: Sociosexuality from Argentina to Zimbabwe

260 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

Figure 1. National level of sociosexuality related to operational

sex ratio across 48 nations of the International Sexuality Descrip-

tion Project.

tively correlate with sociosexuality. However, as seen in

Table 5, these indexes of familial stress were negatively as-

sociated with sociosexuality in every case. These were pre-

dictive failures for developmental-attachment theory.

Economic resources were indexed by gross domestic

product per capita and the human development index

(United Nations Human Development Report 2001). Ac-

cording to developmental-attachment theory, these vari-

ables should be negatively associated with sociosexuality.

Instead, these variables were positively associated with so-

ciosexuality, with more resources and greater human in-

vestment being associated with higher rates of short-term

mating. Finally, national life expectancy rates (an index of

low mortality) should be associated with lower sociosexual-

ity according to the developmental-attachment view. How-

ever, life expectancy was positively correlated with socio-

sexuality, r(45) 0.38, p  .01. As life expectancies in-

creased and mortality rates decreased, sociosexuality

tended to go higher, not lower as predicted by develop-

mental-attachment theory.

Three other variables may be of interest for evaluating

this perspective on sociosexuality. In the view of develop-

mental-attachment theory, family stress, low resources, and

early mortality are cultural precedents to a reproductive

trajectory that includes early puberty, early reproduction,

and more prolific reproduction (Belsky et al. 1991;

Chisholm 1996). International data on early reproduction

were available in the form of teen pregnancy rates and

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 261

Table 4. Intercorrelations among sociocultural variables used to predict national levels of sociosexuality

Sociocultural Variables OSR LBW CMP IMR GDP HDI LE TPR MAM FR

Operational Sex Ratio

Operational Sex Ratio (n  47) —

Familial Stress

Prevalence of Low Birth Weight (n  44) .64

***

—

Child Malnutrition Prevalence (n  20) .47

.80

***

—

Infant Mortality Rate (n  47) .36

.65

***

.81

***

—

Economic Resources

Gross Domestic Product (n  46) .05 .45

.46

.69

***

—

Human Development Index (n  46) .27 .63

***

.83

***

.96

***

.82

***

—

Mortality

Life Expectancy (n  46) .18 .50

***

.60

.91

***

.71

***

.92

***

—

Early and Prolific Reproduction

Teen Pregnancy Rate (n  45) .38

.72

***

.64

.84

***

.69

***

.82

***

.81

***

—

Mean Age at Marriage for Women (n  41) .09 .50

***

.53

***

.52

***

.75

***

.69

***

.47

***

.65

***

—

Fertility Rate (n  46) .45

.59

***

.75

***

.92

***

.58

***

.87

***

.85

***

.75

***

.40

—

Note: OSR  operational sex ratio; LBW  prevalence of low birth weight; CMP  child malnutrition prevalence; IMR  infant

mortality rate; GDP  gross domestic product; HDI  Human Development Index; LE  life expectancy; TPR  teen pregnancy

rate; MAM  mean age at marriage for women; FR  fertility rate;

 p  .05,

 p  .01,

***

 p  .001.

Table 5. How do sociocultural variables relate in predicted ways to national levels of sociosexuality?

Predictions Based on Predictions Based on Observed Observed

Sociocultural Variables Developmental-Attachment Theory Strategic Pluralism Theory Correlation

Familial Stress

Prevalence of Low Birth Weight Positively Associated with SOI Negatively Associated with SOI 0.51***

(n  44)

Child Malnutrition Prevalence (n  20) Positively Associated with SOI Negatively Associated with SOI 0.64***

Infant Mortality Rate (n  47) Positively Associated with SOI Negatively Associated with SOI 0.38***

Economic Resources

Gross Domestic Product (n  46) Negatively Associated with SOI Positively Associated with SOI 0.22

Human Development Index (n  46) Negatively Associated with SOI Positively Associated with SOI 0.39*

Mortality

Life Expectancy (n  46) Negatively Associated with SOI Positively Associated with SOI 0.38**

Early and Prolific Reproduction

Teen Pregnancy Rate (n  45) Positively Associated with SOI Negatively Associated with SOI 0.36**

Mean Age at Marriage for Women Negatively Associated with SOI Positively Associated with SOI 0.28*

(n  41)

Fertility Rate (n  46) Positively Associated with SOI Negatively Associated with SOI 0.31*

Note: *  p  .05, **  p  .01, ***  p  .001.

women’s mean age at marriage. Contrary to expectations,

both indexes of early reproduction were associated with

lower sociosexuality rates across cultures. Data on more

prolific reproduction were indexed by the total fertility rate

across cultures. Again, contrary to theoretical expectations,

higher fertility rates were associated with lower sociosexu-

ality, r(44) 0.31, p  .05. Overall, developmental-at-

tachment theory failed the statistical tests implemented in

this research.

6.5.3. Strategic pluralism theory. A third evolutionary ex-

planation of national variation in sociosexuality comes from

strategic pluralism theory (Gangestad & Simpson 2000). As

noted earlier, Gangestad and Simpson theorize that in cul-

tures with very harsh and difficult environmental con-

ditions, biparental care becomes a critical component in

raising viable offspring. As the need for biparental care

increases, Gangestad and Simpson theorize that monoga-

mous mateships become more important as well. Conse-

quently, sociosexuality should be lower (i.e., more monog-

amous) in nations with difficult child-rearing environments.

The indicators of familial stress noted earlier – prevalence

of low-birth-weight infants, child malnutrition, and infant

mortality – can all be seen as indexes of difficult child-rear-

ing environments. In support of strategic pluralism theory,

all three of these variables correlated negatively with socio-

sexuality, as seen Table 5.

Similarly, the indexes of economic resources and mortal-

ity also reflect environmental difficulty. In support of strate-

gic pluralism theory, as resources diminish and environ-

ments become deadlier, sociosexual levels become more

monogamous (see Table 5). Finally, teen pregnancy, early

marriage, and prolific fertility may be related to difficulty in

raising successful offspring. Although these factors are only

loosely related to environmental demands, as these vari-

ables increased sociosexuality reliably decreased. Overall,

strategic pluralism theory was confirmed by the statistical

tests implemented in this research and should be viewed,

alongside sex ratio theory, as an empirically supported evo-

lutionary perspective on sociosexual variation across cul-

tures.

6.6. Are sex differences in sociosexuality cross-

culturally universal?

Most evolutionary perspectives on animal mating hypothe-

size that males and females are designed to follow somewhat

different reproductive strategies (e.g., Trivers 1972). In hu-

mans, because men tend to be the lesser-investing parent of

our species, they have more to gain than women do from in-

discriminately engaging in short-term sex with numerous

partners (see also Alexander & Noonan 1979; Buss &

Schmitt 1993; Hinde 1984; Symons 1979; Wilson 1987).

A clear implication of human mating theories anchored

in parental investment theory (Trivers 1972) is that men

should generally be more unrestricted than women across

human cultures. This hypothesis was tested in the current

study by directly comparing men’s and women’s mean lev-

els of sociosexuality within the 48 nations of the ISDP. As

seen in Table 6, sex differences in sociosexuality were sta-

tistically significant for all cultures of the ISDP. Evolution-

ary theories that predict universal sex differences in socio-

sexuality are, therefore, consistent with the findings of the

ISDP.

As noted earlier, the SOI uses open-ended responses to

certain questions, making it somewhat susceptible to distri-

butional skew. To address this issue, median tests were per-

formed to determine whether median sex differences in

sociosexuality mirror the results of mean-level sex differ-

ences. As seen in Table 7, in every culture the median man

was significantly higher on sociosexuality than the median

woman, though in Slovakia this difference was only mar-

ginally significant, c

(1, N  125)  3.54, p  .06. Inter-

estingly, in two cultures, Botswana and Latvia, the median

tests for sex differences displayed higher levels of signifi-

cance than the t-tests comparing men’s and women’s

means.

Listed in Table 7 are Mann-Whitney U analyses (with

corresponding z-tests) for differences between men’s and

women’s distributions on sociosexuality. These analyses

help to determine whether, regardless of extreme values

that can affect mean-level averages, men and women dis-

play significantly different variability along the entirety of

their distributions. These key distributional tests docu-

mented that men’s and women’s sociosexual distributions

were significantly different in every nation of the ISDP.

Similar to mean and median statistics, the Mann-Whitney

U tests support the view that men’s and women’s sociosex-

ual profiles reflect an evolutionary history of sex differences

in parental investment.

Finally, one criticism of the SOI is its use of behavioral

questions. It is possible that, although men and women fun-

damentally differ in sexual desire (Schmitt et al. 2003b),

they do not differ in manifest sexual behavior. To address

this issue, Table 7 includes the significance of sex differ-

ences in the both the behavioral items of the SOI (items 1

through 4) and the attitudinal items (items 5 through 7). As

seen in the right column of Table 7, independent t-tests in-

dicated that men and women significantly differed in both

sociosexual behaviors and attitudes across all but three cul-

tures. The one exception to behavioral sex differences oc-

curred in Latvia (though this sex difference was marginally

significant), and the two exceptions to the pervasive trend

in attitudinal sex differences occurred in Slovakia and

Switzerland (again, this was marginally significant). It ap-

pears, therefore, that sex differences in sociosexuality

largely transcend both behavioral and attitudinal features of

human mating psychology across cultures.

Perhaps more important than any form of statistical sig-

nificance, however, is the strength or magnitude (d) of sex-

ual differentiation across cultures. The d statistic represents

the size of the difference between men’s and women’s means

expressed in pooled standard deviation units (Cohen 1988).

As seen in the right column of Table 6, for most nations the

size of the sociosexual difference between men and women

was moderate to large. The largest sex differences were ob-

served in Morocco (d  1.24), Ukraine (d  1.24), Bolivia

(d  1.20), Greece (d  1.18), and the Philippines (d 

1.16). The smallest sex differences were found in Latvia (d

 0.30), Botswana (d  0.39), Germany (d  0.48), Switzer-

land (d  0.49), and the Democratic Republic of the Congo

(d  0.51). In a meta-analysis of commonly reported sex dif-

ferences in sexuality, Oliver and Hyde (1993) concluded that

most sex differences are only small (d  .20) to moderate (d

 .50) in magnitude. These ISDP results, therefore, place

sex differences in sociosexuality (overall d  0.74) among

the largest and culturally most robust ever documented in

the domain of sex and human mating.

Schmitt: Sociosexuality from Argentina to Zimbabwe

262 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 263

Table 6. The significance and magnitude of sex differences in sociosexuality across 48 nations of the

International Sexuality Description Project

Men Women

Nation MSDMSD t d

Argentina 55.52 31.57 30.10 19.35 7.73*** 0.90***

Australia 46.52 25.02 30.73 20.81 7.17*** 0.66**

Austria 55.89 36.75 38.66 23.93 5.60*** 0.55**

Bangladesh 31.10 18.46 11.80 8.16 7.42*** 1.09***

Belgium 39.68 21.00 26.80 16.24 6.80*** 0.69**

Bolivia 61.47 31.18 21.92 18.94 8.31*** 1.20***

Botswana 33.56 28.68 23.06 24.21 2.86** 0.39*

Brazil 53.96 39.14 27.13 17.32 4.13*** 0.82***

Canada 44.33 25.72 27.30 18.18 11.88*** 0.75**

Congo, D.R. 41.16 25.74 29.55 12.94 2.98** 0.51**

Croatia 57.35 28.76 32.15 16.29 7.61*** 0.95***

Czech Rep. 48.96 28.58 29.49 15.78 5.83*** 0.81***

Estonia 51.51 33.58 31.83 23.53 4.15*** 0.66**

Ethiopia 37.88 28.43 18.89 14.59 5.09*** 0.74**

Fiji 54.30 36.78 25.26 18.55 5.19*** 0.87***

Finland 64.03 38.72 41.60 27.75 3.14*** 0.69**

France 45.88 22.52 30.66 21.21 3.48*** 0.66**

Germany 46.36 29.26 34.44 20.81 5.78*** 0.48*

Greece 43.43 18.30 24.32 13.28 7.26*** 1.18***

Hong Kong 29.88 19.45 19.21 11.86 4.52*** 0.63**

Israel 53.99 28.20 31.71 21.56 7.76*** 0.83***

Italy 51.73 28.57 21.39 14.58 9.65*** 1.13***

Japan 32.47 19.96 20.72 13.52 4.76*** 0.63**

Latvia 49.42 23.61 41.68 26.68 1.90* 0.30*

Lebanon 43.90 26.62 17.21 15.78 9.27*** 1.06***

Lithuania 60.44 35.87 35.25 16.40 3.95*** 0.82***

Malta 40.56 28.58 25.17 16.56 4.99*** 0.64**

Mexico 49.04 27.06 25.99 11.08 7.82*** 0.99***

Morocco 65.58 37.15 20.06 21.32 8.80*** 1.24***

Netherlands 50.51 30.47 31.56 14.90 5.78*** 0.76**

New Zealand 60.42 33.53 38.79 26.93 5.67*** 0.69**

Peru 51.68 35.56 21.23 12.03 7.72*** 1.00***

Philippines 51.24 33.03 17.95 10.53 10.32*** 1.16***

Poland 44.29 28.96 26.90 20.75 8.44*** 0.68**

Portugal 41.27 20.21 21.32 10.76 9.63*** 1.09***

Romania 48.64 33.33 19.48 16.80 7.99*** 0.98***

Serbia 48.99 23.81 31.89 21.28 5.16*** 0.71**

Slovakia 44.27 31.75 28.52 13.54 3.74*** 0.64**

Slovenia 59.45 27.01 36.45 17.99 4.59*** 0.89***

South Korea 30.52 15.70 16.22 10.98 11.71*** 0.97***

Spain 46.08 23.97 25.17 14.47 8.31*** 1.01***

Switzerland 45.25 26.61 34.26 18.20 3.02** 0.49*

Taiwan 28.42 20.50 14.24 7.62 6.17*** 0.80***

Turkey 54.16 35.44 21.71 14.58 11.58*** 1.03***

Ukraine 50.79 28.92 17.36 8.65 11.06*** 1.24***

United Kingdom 57.38 34.71 29.60 21.91 9.53*** 0.95***

United States 48.03 29.63 29.24 20.56 19.07*** 0.73**

Zimbabwe 34.80 31.22 13.98 12.11 5.92*** 0.80***

Total ISDP Sample 46.67 29.68 27.34 19.55 46.32*** 0.74**

Note: ISDP  International Sexuality Description Project. For t-values, *  p  .05, **  p  .01, ***  p  .001. For d values, * 

small effect size, **  moderate effect size, ***  large effect size (Cohen, 1988).

Schmitt: Sociosexuality from Argentina to Zimbabwe

264 BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2

Table 7. Sex differences in sociosexual medians, distributions, behaviors, and attitudes across 48 nations of the

International Sexuality Description Project

Nonparametric Tests Facets of Sociosexuality

Medians Mann-Whitney U Behaviors Attitudes

Nation 2 ztt

Argentina 36.02*** 7.26*** 7.31*** 6.38***

Australia 44.71*** 7.33*** 6.24*** 8.90***

Austria 11.98*** 5.10*** 5.23*** 5.15***

Bangladesh 51.80*** 7.18*** 6.39*** 6.54***

Belgium 44.89*** 7.24*** 5.96*** 6.67***

Bolivia 57.72*** 7.81*** 8.05*** 5.86***

Botswana 14.15*** 3.49*** 2.44* 4.67***

Brazil 8.44** 3.74*** 4.13*** 2.50**

Canada 143.63*** 12.01*** 11.84*** 5.89***

Congo, D.R. 6.35** 2.97** 2.75** 2.10*

Croatia 37.36*** 6.98*** 7.15*** 5.80***

Czech Rep. 21.24*** 5.59*** 5.38*** 5.40***

Estonia 12.80*** 4.38*** 3.62*** 5.44***

Ethiopia 29.79*** 5.76*** 5.04*** 2.54**

Fiji 18.19*** 5.04*** 5.08*** 3.12**

Finland 7.03** 3.27*** 2.84** 3.71***

France 7.87** 3.95*** 3.03** 4.21***

Germany 26.80*** 5.73*** 5.41*** 5.41***

Greece 19.62*** 5.81*** 6.83*** 5.83***

Hong Kong 11.51*** 4.43*** 4.09*** 4.10***

Israel 48.65*** 8.13*** 7.35*** 5.90***

Italy 53.45*** 8.54*** 8.63*** 11.75***

Japan 11.10*** 4.66*** 3.84*** 6.05***

Latvia 5.49* 2.73** 1.45

†

3.80***

Lebanon 48.40*** 8.66*** 8.12*** 10.96***

Lithuania 6.28** 3.60*** 3.76*** 2.55**

Malta 17.84*** 4.22*** 4.59*** 5.13***

Mexico 46.63*** 7.55*** 7.43*** 6.02***

Morocco 58.00*** 7.81*** 8.13*** 8.78***

Netherlands 22.25*** 5.44*** 5.63*** 3.57***

New Zealand 26.37*** 5.72*** 5.47*** 4.71***

Peru 39.66*** 7.63*** 6.90*** 9.15***

Philippines 85.80*** 10.04*** 9.35*** 10.96***

Poland 63.08*** 8.45*** 7.98*** 8.34***

Portugal 50.92*** 8.42*** 8.68*** 8.20***

Romania 45.92*** 8.14*** 6.90*** 11.81***

Serbia 22.83*** 5.61*** 4.43*** 6.31***

Slovakia 3.54

†

3.69*** 3.65*** 0.30

Slovenia 17.95*** 4.83*** 4.30*** 3.56***

South Korea 75.35*** 10.96*** 10.47*** 10.03***

Spain 29.69*** 7.28*** 7.56*** 7.36***

Switzerland 4.56* 2.71** 2.96** 1.91

†

Taiwan 44.99*** 7.61*** 4.95*** 9.70***

Turkey 84.18*** 10.53*** 10.50*** 12.27***

Ukraine 77.65*** 10.03*** 9.47*** 19.39***

United Kingdom 56.90*** 8.55*** 9.05*** 8.70***

United States 269.23*** 18.76*** 17.73*** 19.09***

Zimbabwe 27.95*** 6.34*** 5.78*** 3.89***

Total ISDP Sample 1,690.74*** 46.27*** 43.04*** 42.44***

Note: ISDP = International Sexuality Description Project.

†

 p  .10, *  p  .05, **  p  .01, ***  p  .001.

Within the constraints of the current methodology and

sampling limitations, it can be concluded from these results

that sex differences in sociosexuality are a cultural univer-

sal, supporting the basic tenets of parental investment the-

ory (Trivers 1972). In addition, based on an ANOVA with

sex of participant and nation as independent variables and

sociosexuality as the dependent variable, the overall partial

eta-squared effect size of sex was very large (h

 0.15; Co-

hen 1988), more than double the moderate effect size of na-

tion (h

 0.06). Culture has an important influence on so-

ciosexuality, but biological sex is the larger and stronger

predictor of human mating strategies across the nations of

the ISDP.

6.7. Why do nations differ in the magnitude of sex

differences in sociosexuality?

Even though sex differences in sociosexuality appear to be

culturally universal (at least across the spectrum of modern

ISDP nations), and in some ways sex differences are

stronger than the measurable effects of culture, this does

not mean that sex differences must be the result of evolved

reproductive strategies. It could be that sociosexual sex dif-

ferences are a by-product of some other force that happens

to permeate all human cultures, such as patriarchy, religion,

or some other sociohistorical influence (Harris 1993). It

also could be that sex differences in sociosexuality are the

direct result of some biological difference between men

and women, but the difference does not involve psycholog-

ical adaptations to sociosexuality per se.

6.7.1. Social structural theory. According to social struc-

tural theory (Eagly & Wood 1999; Wood & Eagly 2002),

men and women are not designed to differ in sociosexual-

ity. Instead, pancultural sex differences in sociosexuality

likely stem from ubiquitous differences in the way men and

women fulfill social roles. Eagly and Wood (1999) argue

that the intensity and rigidity of social roles can vary across

cultures, due in part to the local ecology and its influence

on the value of women’s economic, political, and relational

contributions (see also Low 1989). An important implica-

tion of this perspective is that in cultures where women are

more severely constrained in terms of economic, political,

and relational-independent social roles (i.e., cultures with

traditional sex-role ideologies; see Williams & Best 1990),

sex differences in sociosexuality should be larger. Within

cultures that possess more modern or progressive sex-role

ideologies – where women have more access to money,

power, and the ability to make their own reproductive de-

cisions – women are allowed to explore a wider array of so-

cial roles. Consequently, sex differences in sociosexuality

should be smaller, or perhaps even absent, in cultures

where either gender can take on the role of the other (see

also Buss & Barnes 1986).

Table 8 contains the intercorrelations of several socio-

cultural indicators of gender equality, relational freedom,

and sex-role ideology. In most cases, when women have

greater access to political power and resources (e.g., Gen-

der Empowerment Measure), they also tend to have more

relational and reproductive freedom (e.g., head their own

household, use contraception when married, and divorce

more freely). These findings replicate several results from

previous anthropological studies (e.g., Pasternak et al. 1997;

Pearson & Hendrix 1979), though some studies of prein-

dustrial cultures have failed to find robust links between all

indicators of women’s status and sexual freedom. Whyte

(1978), for example, found only female-centered social

structures (e.g., matrilineality and matrilocality) were asso-

ciated with more sexual equality and freedom. Other fac-

tors, such as the degree of warfare in a culture, were not

linked as expected to sexual equality. In this study, women’s

economic and reproductive freedom is generally associated

with progressive sex role ideologies, low patriarchy (i.e., low

hostile sexism), and low levels of cultural masculinity, pre-

cisely as predicted by social structural theory (Eagly &

Wood 1999; Wood & Eagly 2002).

Table 9 contains the correlations between sociocultural

indicators of gender equality and sex differences in socio-

sexuality. As seen in the right side of Table 9, the Gender

Empowerment Measure was negatively correlated with

magnitude of sex differences in sociosexuality, r(32) 

.56, p  .001. As predicted by social structural theory, in-

creased gender equity was associated with the erosion of

large sex differences in human mating strategies. Signifi-

cant associations also were found between sex differences

in sociosexuality and the percentage of women in parlia-

ment, the percentage of women in ministerial positions, the

percentage of women-headed households, and divorce

rates across cultures. Although direct measures of sex role

ideology were not significantly associated with the magni-

tude of sex differences, the low sample sizes from Williams

and Best (1990) and Glick et al. (2000) precluded ample

power for fully evaluating this prediction. Overall, it ap-

peared that when women gain more sociopolitical and re-

lational freedom, sex differences in sociosexuality shift from

large magnitudes to more moderate magnitudes of effect.

Even though sex differences in sociosexuality were more

moderate in progressive cultures, it is not exactly clear from

social structural theory what form this shift in sociosexual-

ity should have taken. Is it the case that men and women

are naturally restricted (Hazan & Diamond 2000), with sex

roles in certain cultures causing large sex differences by

promoting unrestricted sociosexuality in men – perhaps us-

ing male promiscuity as a means of patriarchal oppression?

Are women designed to be more promiscuous than men

(Sherfey 1966), with sex roles in certain cultures somehow

reversing our naturally polyandrous mating system? Are

men naturally more promiscuous than women (Symons

1979), with sex roles in certain cultures minimizing sex dif-

ferences by accentuating unrestricted sociosexuality in

women? Or is it the case both men and women are natu-

rally unrestricted (Barash & Lipton 2001), with sex roles in

certain cultures causing large sex differences by suppress-

ing women’s innate tendency toward sexual promiscuity?

To address these questions, additional theorizing is needed.

6.7.2. Strategic pluralism theory. According to strategic

pluralism theory (Gangestad & Simpson 2000), women are

designed to facultatively shift their mating strategies de-

pending on certain qualities of the local environment. In

demanding environments that necessitate high levels of bi-

parental care (e.g., cultures with high stress, few resources,

and high mortality), women are hypothesized to become

more sociosexually restricted. In nondemanding environ-

ments, women are able to expend additional effort on short-

term mating, in part to gain access to genetically valuable

males (Gangestad 2001; Simpson & Orina 2003), and so

women’s sociosexuality should increase or become more

Schmitt: Sociosexuality from Argentina to Zimbabwe

BEHAVIORAL AND BRAIN SCIENCES (2005) 28:2 265

unrestricted in nondemanding environments. Although

some men’s sociosexuality is thought to react to women’s so-

ciosexual shifts to a certain degree, other men were “were

able to carry out short-term tactics successfully at all times,

regardless of the environmental factors to which women

were responding” (Gangestad & Simpson 2000, p. 586).

As a result, women’s sociosexuality should be highly de-

pendent on environmental demands, but men’s sociosexu-

ality – as a group – should be somewhat less correlated with

environmental stressors, resource levels, and mortality

rates. Importantly, because men tend to be more oriented

toward short-term mating in general (Buss & Schmitt

1993), the size or magnitude of the difference between men

and women should be larger in demanding environments

where women shift away from men’s higher levels of unre-

stricted sociosexuality

These hypotheses were evaluated by correlating various

indicators of environmental demand with men’s sociosexu-

ality, women’s sociosexuality, and with the effect size (d) of

the sex difference in sociosexuality across cultures. As seen

in Table 10, sex differences in sociosexuality were related as

predicted to several indicators of environmental demand.

For example, as the prevalence of low birth weights in-

creased, the difference between men and women margin-

ally increased, r(42)

Sociosexuality from Argentina to Zimbabwe: A 48-Nation Study of Sex, Culture, and Strategies of Human Mating

Abstract and Figures