1. Introduction

Studies on human mate preferences have reported a number of sex differences. Women typically prefer older partners (e.g., Buss, 1989a, Kenrick & Keefe, 1992, Otta et al., 1999, Waynforth & Dunbar, 1995). They have stronger preferences for resource-acquisition characteristics in a partner whereas men have stronger preferences for physical attractiveness (e.g., Buss, 1989a, Buss, 1990, Buss, 1994, Buss & Barnes, 1986, Feingold, 1990, Feingold, 1991, Feingold, 1992, Hill, 1945, Li et al., 2002, Waynforth & Dunbar, 1995). These sex differences have been attributed to sex-specific reproductive constraints: the minimal investment in reproduction by a female is greater than that of a male owing to the costs of producing large gametes (Bateman, 1948), internal gestation, lactation, and extended parental care (Trivers, 1972). Thus, female reproductive success is constrained by access to the resources necessary to raise costly offspring whereas male reproductive success is constrained by access to fertile females (Trivers, 1972). Consequently, women should prefer partners who demonstrate willingness and ability to invest direct resources in offspring (e.g., partners with resource-acquisition characteristics or older partners with greater accumulated resources) and men should prefer cues to reproductive capacity and fertility (e.g., a healthy, attractive appearance).

High levels of intrasexual variation, however, implicate complexity and trade-offs in human mate preferences (e.g., Smuts, 1989, Smuts, 1991a, Smuts, 1991b, Waynforth & Dunbar, 1995). In addition to providing direct resources through investment of parental care, males can also provide offspring with indirect heritable qualities (Trivers, 1972). Males possessing high heritable fitness are likely to be successful at pursuing short-term mating strategies and, as a consequence, are less likely to provide parental care and investment in long-term relationships (Waynforth, 1999). Thus, females must trade off the importance of obtaining genetic quality versus material resources in a partner (see work of Gangestad & Simpson, 2000, and Waynforth, 2001).

Most studies on human mate preferences have used samples from societies with cash economies and a division of labor in which women have historically been constrained in their participation in the workforce (e.g., Buss & Barnes, 1986, Hrdy, 1997). When women can only secure resources through a partner, they may benefit from choosing a partner with material resources over a partner with indirect heritable qualities. It has been argued that when females can access the resources necessary to raise offspring independently, the importance of male investment of resources in offspring will decrease (Cashdan, 1993, Gangestad, 1993, Low, 1990). Concordant predictions come from the social structural theory, which posits that socialization of the sexes into gender roles is responsible for sex differences in mate preferences (Eagly & Wood, 1999, Wood & Eagly, 2002); that is, men and women attempt to fill specific gender roles (female as homemaker and male as breadwinner). Mate preferences reflect attempts to maximize resources denied to each sex by gender roles. Although the social structural and “adaptive trade-offs” perspectives both predict decreased magnitudes of sex differences in mate preferences with increasing sexual equality, they offer differing explanations for the mechanisms by which these shifts would occur. The perspectives may, however, be compatible and offer proximate-level and ultimate-level explanations, respectively (Wood & Eagly, 2002).

Studies on the effects of female status on mate preferences have yielded conflicting results. Positive relationships have been reported between (a) expected female income and preference rankings for resource-acquisition characteristics (e.g., Townsend, 1989, Wiederman & Allgeier, 1992); (b) female income and importance placed on the potential income of a partner (Buss, 1989b); and (c) female income and requests for resources in lonely-hearts advertisements (Gil-Burmann et al., 2002). In addition, positive assortative mating has been reported for cultural and economic status (Kalmijn, 1994) and educational attainment and socioeconomic origins (Kalmijn, 1991). Conversely, Johannesen-Schmidt and Eagly (2002) reported positive relationships between the extent to which females endorsed the traditional female gender role and preferences for good earning potential and age in a partner. Similarly, Koyama, McGain, and Hill (2004) reported a negative relationship between importance placed on “good earning potential” in a partner and feminist attitudes. Furthermore, in reanalyses of Buss's (1990) data from 37 cultures, Eagly and Wood (1999) found female empowerment to be negatively correlated with female preferences for male earning potential and older partners and Kasser and Sharma (1999) reported a negative relationship between educational equality and female preference for male resource-acquisition characteristics. Two studies have reported positive relationships between female status and preferences for a cue to heritable quality in a partner: women's participation in economies (Gangestad, 1993) and own-rated financial prospects (Koyama et al., 2004) were found to relate positively to preferences for physical attractiveness in a partner.

Evidence for an effect of female resources on preferences for direct investment of resources versus indirect genetic benefits is not conclusive. Female economic status does not appear to be associated with diminished preferences for resource-acquisition characteristics in a partner but is associated with preferences for physical attractiveness. Conversely, attitudes associated with endorsement of a less-traditional gender role are associated with decreased preferences for resource-acquisition characteristics. Gangestad and Simpson (2000) proposed that these discrepancies reflect differences in the measures used. Wealth may not be the same as the power tapped by measurement of attitudes toward sexual equality or cultural measurements of female empowerment. Alternatively, measurement of attitudes may not tap the actual ability of individuals to provide for offspring independently. We argue that assessment of control of resources includes the effects of both access to resources (as previously measured by income) and autonomy (as previously tapped by measures of attitudes), providing a measure of ability to provide for offspring independently without the confounding effects of assortment for wealth. It is predicted that the ability of females to acquire and control the resources necessary to raise offspring will allow preferences to shift toward indirect heritable benefits, as the importance of acquiring resources from a partner decreases. Thus, the aims of the current study were to (a) develop a measure of resource control at the level of the individual and (b) assess relationships between this and female mate preferences. Our study is unique in investigating the effects of control of resources on mate preferences, in examining the effects of autonomy at the level of the individual, and in using a sample more representative of the general population than previous studies.

Previous research have indicated that measures of female control of resources do not covary such that they can be usefully combined to provide a single measure (Low, 1990, Whyte, 1978, Whyte, 1979, Yanca & Low, 2004). Such studies have used measures including control of the fruits of one's labor, control of dwellings, and authority over others in the family and the community (Low, 1990, Whyte, 1978, Whyte, 1979, Yanca & Low, 2004). Thus, we developed a series of questionnaire items designed to assess the extent to which individuals independently acquire and control resources and possess power (i.e., exert authority over others).

To summarize, it was predicted that female resource control will influence the trade-off between preferences for resource-acquisition characteristics and those for indirect genetic benefits. Increasing female control of resources was predicted to be associated with (a) decreased preferences for resource-acquisition characteristics in a partner and (b) increased preferences for physical attractiveness. We predicted that the effects of resource control would differ from those of income and expected positive relationships between income and preferences for resource-acquisition characteristics (because of assortative mating). We assessed preferences for resource-acquisition characteristics and cues to indirect heritable benefits through ranking of partner characteristics. In addition, preferences for age in a partner provided a measure of an indirect preference for accumulated resources.

2. Methods

3. Results

3.2. Resource control

Measures of participants' resource control were entered into a factor analysis. Factors were extracted using principal components analysis and rotated using the standard Varimax rotation with Kaiser normalization. Two factors with eigenvalues >1 were extracted (see Table 1). Variables that loaded highly on Factor 1 (eigenvalue=2.15, accounting for 30.74% of the variance) were financial independence, control of finances, and input in decisions in the home and the workplace. Factor 1 was interpreted as representing financial independence and power. Variables that loaded highly on Factor 2 (eigenvalue=1.34, accounting for 19.19% of the variance) were importance of financial independence and importance of having a career. Factor 2 was interpreted as representing ambition. Participants' scores for each factor were computed using the regression method, such that the mean of each factor was zero and the variance was equal to the squared multiple correlation between estimated and true factor scores.

Table 1. Measures of participants' control of resources: factor loadings, eigenvalues, and percentages of variance for factor analysis on resource control questionnaire responses

	Factor	Eigenvalue	Percentage of variance	Variable	Loading (r)
	Financial independence and power	2.15	30.74	Financial independence	.75
				Control of finances	.6
				Input in decisions in the home	.69
				Input in decisions in the workplace	.69
	Ambition	1.34	19.19	Importance of financial independence	.8
				Importance of having a career	.84

3.3. Preliminary analysis

Spearman's correlations between resource control factors, mate preferences, and possible confounding variables are displayed in Table 2. Bonferroni correction was applied (i.e., p values were multiplied by the number of relationships examined). There were positive correlations between income and age preferences as well as financial independence and power and age preferences. There were negative correlations between ambition and age preferences. This preliminary analysis may suggest that the factors tap different aspects of resource control. All potential confounding variables, however, were found to correlate with at least one of the dependent or independent variables (with the exception of crowding in natal home). Therefore, the effects of these covariates should be controlled before conclusions are made. Crowding was not included in further analyses because it was not related significantly to any other variable.

Table 2. Spearman's correlations between variables

		Ideal partner age	Maximum partner age tolerated	Minimum partner age tolerated	Preference ranking for physical attractiveness	Preference ranking for good financial prospects	Income	Financial independence and power	Ambition
	Own age	.87⁎	.78⁎	.83⁎	.03	−.01	.4⁎	.52⁎	−.16⁎
	Parents' income while growing up	−.11⁎	−.1⁎	−.11⁎	−.01	.01	.07	−.11⁎	−.03
	Crowding	.06	.05	.08	.02	.02	.05	.03	.02
	Self-rated attractiveness	.07	.03	.06	.09⁎	.03	.12⁎	.1⁎	.05
	Ideal relationship type on day of testing	.13⁎	.2⁎	.15⁎	−.02	−.07	.001	.08	−.07
	Ideal partner age		.85⁎	.83⁎	.02	.01	.39⁎	.48⁎	−.4⁎
	Maximum partner age tolerated			.66⁎	.02	−.002	.33⁎	.41⁎	−.17⁎
	Minimum partner age tolerated				.03	.02	.38⁎	.49⁎	−.13⁎
	Preference ranking for physical attractiveness					.12⁎	.01	.04	.05
	Preference ranking for good financial prospects						.04	−.02	.03
	Income							.42⁎	−.08

⁎ P<.01 (with Bonferroni correction).

3.4. Analysis

3.4.1. Age preferences

Ideal partner age and maximum and minimum partner ages tolerated were entered in turn as dependent variables in a hierarchical regression model. For results, see Table 3.

Table 3. Multiple linear regression models with mate preferences as independent variables

		Ideal partner age		Maximum partner age tolerated		Minimum partner age tolerated		Preference for physical attractiveness over financial prospectsa
		β	p	β	p	β	p	β	p
	Own age	.8	<.001	.7	<.001	.73	<.001	−.004	NS
	Marital status	−.03	.049	.003	NS	.02	NS	−.01	NS
	Ideal relationship type	−.005	NS	−.015	NS	.03	.022	.04	NS
	Parents' income while growing up	−.03	.049	−.018	NS	−.008	NS	.04	NS
	Self-rated attractiveness	.03	.013	.007	NS	−.003	NS	.01	NS
	Income	.05	.002	.023	NS	.06	.001	−.184	.006
	Financial independence and power	.023	NS	.016	NS	.05	.002	.15	.01
	Ambition	−.013	NS	−.05	.006	−.013	NS	.001	NS
	Adjusted R2	.68		.52		.62		.009
	F	490.92		252.29		383.64		NA
	p	<.001		<.001		<.001		.07

a Binary logistic regression model.

Ideal partner age was significantly predicted by own income (β=.05, p=.002); that is, wealthier women prefer older partners. There was no effect of resource control on ideal partner age; thus, Prediction 1 was not supported. There was a positive relationship between own age and ideal partner age (β=.8, p<.001), replicating previous findings that age preferences are contingent upon own age (Kenrick & Keefe, 1992). Self-rated attractiveness was positively related to ideal partner age (β=.03, p=.013). Parents' income while growing up (β=−.03, p=.049) and marital status (β=−.03, p=.049) were negatively related to ideal age difference, indicating that women from a wealthier background or who were in a relationship preferred younger partners.

Ambition significantly predicted maximum partner age tolerated (β=−.05, p=.006); that is, women who are ambitious are less willing to tolerate partners much older than them, providing support for Prediction 1. There was a positive relationship between own age and maximum partner age tolerated (β=.7, p<.001), indicating that maximum partner age tolerated increases with own age.

Financial independence and power (β=.05, p=.002) and income (β=.06, p=.001) significantly predicted minimum partner age tolerated; that is, wealthier women or women who are financially independent and powerful are less willing to tolerate younger partners. Thus, Prediction 1 is not supported in this measure. Minimum partner age was also significantly predicted by own age (β=.73, p<.001) and ideal relationship type (β=.03, p=.022); that is, minimum partner age increases with own age and women who are seeking a committed relationship are less tolerant of younger partners.

3.4.2. Preference rankings

The dichotomous variable indicating preference for physical attractiveness versus good financial prospects was entered as the dependent variable in a binary logistic regression model. Financial independence and power significantly predicted this preference [β=.15, Exp(β)=1.2, p=.01]; that is, resource control was associated with preferences for physical attractiveness over good financial prospects, providing support for Prediction 2. Income also significantly predicted this preference [β=−.18, Exp(β)=1.2, p=.006], indicating that wealthier women prefer good financial prospects over physical attractiveness.

4. Discussion

We isolated two resource control factors (financial independence and power and ambition) from female responses to questionnaire items and examined relationships between these and mate preferences in females. Financial independence and power was associated with older minimum partner ages tolerated and preferences for physical attractiveness over good financial prospects in a partner. Ambition was associated with younger maximum partner ages tolerated. Our results suggest resource control to be an important predictor of sex-differentiated mate preferences.

Because no variable loaded onto both factors and because each factor influenced mate preferences in different ways, we conclude that our factors are distinct and tap different aspects of resource control. Financial independence and power may tap general resource control whereas ambition may tap attitudes and desires associated with obtaining resource control. Although neither factor provided support for both predictions independently, our results demonstrate that resource control does lead to shifts in mate preferences in the predicted directions.

We predicted that the effects of female income and resource control on sex-differentiated mate preferences would differ because of the confounding effects of assortative mating associated with wealth and the importance of actual control over resources. There was a positive relationship between income and ideal partner age but no effect of resource control. The effect of financial independence and power on minimum partner age tolerated was in the same direction as that of income. This resource control factor, however, was also associated with preferences for physical attractiveness over good financial prospects, whereas income was associated with the opposite preference. It can be concluded that the effects of resource control on preferences for cues to heritable quality over resources differ from those of income. Concordant effects of financial independence and power and income on minimum partner age tolerated may reflect an unwillingness of financially independent, powerful women to support a younger partner. Alternatively, this may reflect assortment for personality characteristics associated with obtaining independence and power, which may not be associated with younger partners.

We found no effect of resource control on ideal partner age. Ambition, however, was associated with decreased maximum partner age tolerated, indicating that a measure of resource control is associated with decreased age preferences (manifested as lower tolerance of older partners). Our results are consistent with a shift in the trade-off between preferences for cues to investment of direct resources (i.e., maximum partner age tolerated and good financial prospects) and those for cues to indirect heritable benefits (i.e., physical attractiveness). These results are consistent with the hypothesis that constraints on female access to and control of resources contribute to sex differences in preferences for age, physical attractiveness, and resources in a partner.

Our results are concordant with those of studies that have assessed attitudes toward the traditional female gender role, feminist attitudes, and cultural levels of female empowerment (e.g., Koyama et al., 2004, Johannesen-Schmidt & Eagly, 2002, Eagly & Wood, 1999, Gangestad, 1993). Although we predicted that these measures will correlate with resource control, our results implicate the importance of constraints on female ability to access and control resources in sex-differentiated mate preferences and highlight the importance of assessing actual control of resources rather than access to resources (e.g., wealth). Resource control may tap the access to resources associated with wealth or income, as well as the attitudes and power that are associated with feminist attitudes and endorsement of the traditional female gender role.

Parents' income was negatively related to ideal partner age. This may suggest that women from wealthier backgrounds are less interested in older partners, again implicating a role of access to resources in sex-differentiated mate preferences. Self-rated attractiveness was positively related to ideal partner age. This may reflect condition dependence in which women who consider themselves to have a higher mate value are able to demand partners with greater accumulated resources (i.e., older partners). To our knowledge, no previous study has reported condition dependence in age preferences.

The temporal context of the mating strategy pursued is known to influence mate preferences (e.g., Buss & Schmitt, 1993, Gangestad & Simpson, 2000, Little et al., 2002). Therefore, it is possible to argue that any variation in mate preferences associated with resource control may reflect underlying differences in the temporal context of the relationship sought. For example, women who are financially independent from a partner may be single through choice or may choose to pursue only casual relationships, thus using a short-term mating strategy. By including marital status and ideal relationship type on the day of testing in analyses, however, we have demonstrated that resource control influences mate preferences above and beyond the effects of the pursuit of long-term or short-term mating strategies.

Despite the low variation in the socioeconomic status and education of our sample, we were able to access women from broader age and status profiles as compared with previous studies. Results suggesting effects of resource control on mate preferences despite the low demographic variation of the sample are encouraging, and further research should attempt to reach women in the highest and lowest income brackets and with greater diversity in level of education.

In conclusion, our results implicate the role of constraints on women in the expression of sex differences in mate preferences reported in previous studies. Further research will explore the effects of control of resources on magnitudes of sex differences in preferences. Our results are consistent with perspectives of mate preferences based on both adaptive trade-offs and social structures. Further research should attempt to assess the proximate mediating role of perceptions of autonomy and endorsement of the traditional female gender role in the relationships reported in this study.