Archives of Sexual Behavior

, Volume 42, Issue 2, pp 203–211

Age Preferences in Dating Advertisements by Homosexuals and Heterosexuals: From Sociobiological to Sociological Explanations


    • Department of SociologyRutgers University
Original Paper

DOI: 10.1007/s10508-012-0031-7

Cite this article as:
Burrows, K. Arch Sex Behav (2013) 42: 203. doi:10.1007/s10508-012-0031-7


Current sociobiological thought suggests that significant components of mate selection are based on indicators that correlate with the ability to produce and support offspring. Theorists have suggested that men tend to be attracted to and marry younger women, while women tend to be attracted to and marry older men. This behavior is referred to as age hypergamy. I complicate this picture by using gay men as a population in which to explore alternative components of mate selection as reflected in our behavior. Analyses of 120 dating advertisements from gay men and heterosexual men and women indicated that there exists a good measure of hypergamic age preference that is comparable to the heterosexual population and that relates to subjects’ gender presentation. Data suggest that the biological-reproductive theory of age hypergamy is incomplete and support a cultural reproduction model of gender role behavior and preference in both heterosexuals and homosexuals.


Mating preferencesHomosexualsMeme theoryDating adsHypergamy


Classic sociobiological thought suggests that some of the significant components of mate selection are based on visual, vocal, and chemical cues that indicate well-being, particularly as they relate to reproducing and supporting offspring (Trivers, 1985; Wilson, 1975). Using this bioreproductive understanding of human mate selection, marriage hypergamy theory posits that women tend to seek men with financial characteristics suggestive of the ability to support a family, characteristics that are frequently present in older rather than younger men. The theory further suggests that men are drawn to the physical characteristics of younger women, in part because these characteristics signify an ability to bear children (Cameron & Collins, 1997). Rather than asserting that either partner consciously evaluates future mates based on childbirth or child support potential, classic sociobiological theory argues that it is deep-seated biological drives that influence mate preference (Nur, 1989; Trivers, 1985; Wilson, 1975). More recent research indicates that mate selection is more complex than originally thought, as in some areas of the world men prefer women with a high hip to waist ratio, which is thought to lead to the bearing of more sons (Manning, Trivers, Singh, & Thornhill, 1999). Modern revisions of this theory better balance biological and conscious decisions, but at the heart of the theory is still biological drives, and it this aspect of the theory that I am exploring in this article. In heterosexual marriage and partnership, the cross-cultural rule for age hypergamy is the concept that labels the common behavior of men marrying younger women and women marrying older men.

Because the research supporting the sociobiological approach predominantly addresses mate selection among heterosexuals only, or homosexuals as a group, however, it leaves room for further cultural discussions of the process by which mate selection takes place. I engage in this discussion by focusing on mate selection among gay couples, building on research in this field that has shown tentative support for age hypergamic patterns in mate selection similar to those of heterosexuals (Rasmussen et al., 1998). The premise of my approach is based upon the idea that when couples are biologically unable to reproduce with each other, hypergamy patterns, if they exist, must have an explanation other than that of pure reproductive ambition.1

The presence of age hypergamy among gay men would raise questions for both biological and sociological explanations of behavior. For sociobiologists, it calls attention to similarities between heterosexual and homosexual mate selection and calls into question its premise as propagation of the species as a significant explanation of age hypergamy. Yet, for social scientists, it raises an equally serious question about how to develop a social explanation for such consistent behavior patterns between apparently diverse populations.

In this study of mate selection among gay men,2 I focused upon the following questions: (1) Is it legitimate to assume that age hypergamy is a common pattern of social behavior among gay men? (2) If age hypergamy is a common behavior among gay men, is age the consistent criterion for mate selection? (3) If age correlates with one or more additional criteria for mate selection, could they provide a basis for a social understanding of age hypergamy among both heterosexuals and homosexuals? (4) Lastly, what might such an understanding look like?

I ultimately argue that it is a sociological understanding based on gender roles and the socialization of gender role behavior that provides a broad-based explanation of age hypergamy among populations that can and cannot reproduce with each other. More specifically, I’ll introduce the cultural memetic theory as an approach towards understanding how age hypergamy becomes the norm among both gay and heterosexual couples.

To examine the presence of a systematic pattern of age difference preference in the intimate partnerships of gay men, I relied on men’s self-identification as having either a more masculine or feminine gender expression or “butch” and “femme.” Focusing on gender role in this way allowed me to attend to the under-explored effect that these gender expressions play in mate selection. I analyzed a sample of personal advertisements collected from eight urban communities in the United States, following the approach of Cameron and Collins (1997), Hou (2002), and Rasmussen et al. (1998) (see also Harrison & Saeed, 1977; Hirschman, 1987; Rajecki, Bledsoe, & Rasmussen, 1991; Thorne & Coupland, 1998). I specifically examined those observations indicating strong association with masculine or feminine gender role characteristics to determine if these groups exhibited similar age hypergamy patterns to those of heterosexual men and women.

Women’s fertile years correlate with youth, therefore making younger women more attractive, reproductively speaking, to a potential mate. Alternatively, men tend to have more resources as they grow older, making older men more attractive to women (Bailey, Gaulin, Agyei, & Gladue, 1994). Homosexual men rate younger partners as more attractive, and therefore as more desirable (Jankowiak & Fischer, 1992). In their study, Bailey et al. (1994) found that preference for younger partners was significantly less pronounced for homosexual men than for heterosexual men, and gay and heterosexual men have been found to have similar age preferences for sexual partners (Silverthorne & Quinsey, 2000). By dividing gay men into groups of more masculine and more feminine, the present study may provide more clarity of gay men’s preferences.

The sociological approach to understanding patterned age differences in relationships proposed in this article considers the gender roles of the partners, regardless of the biological sex composition of the partnership. Testing alternative theories of age hypergamy requires a focus on gender roles and the extent to which they are adopted and transmitted between men and women of all sexual orientations. Researchers operating outside of the sociobiology framework have found that homosexual couples hold similar expectations for dating and courtship behavior as do heterosexual couples (Holmes, Little, & Welsh, 2008; Hou, 2002; Thorne & Coupland, 1998). This article builds upon that work by focusing on the gender roles adopted by the couple.

A major criticism of sociobiology involves the possibility of alternative explanations for social behavior. In particular, social rather than biological and evolutionary theories can account for age hypergamy. For example, age disparity between partners is a phenomenon that may, in fact, be falling out of preference in favor of romantic love, regardless of financial or reproductive health (Shorter, 1976). Age is the characteristic most likely to be similar among married persons, especially among younger couples (Buss, 1985), and a shared cohort presents the possibility of similar perspectives and cultural experiences. Assortative mating—that men tend to marry women judged to be more physically attractive and women seek men considered more financially successful—could be a result of a status characteristic exchange processes unrelated to reproductive drives3 (Hirschman, 1987; Humphreys & Berger, 1981; Webster & Driskell, 1983). Marital homogamy in general, and educational homogamy in particular, in which partners have the same or similar levels of educational attainment, has been shown to be increasingly influential in partner selection (Becker, 1973; Blackwell & Lichter, 2000; Schwartz & Mare, 2005), and is perhaps an even more potent factor in mate selection than religious affiliation (Kalmijn, 1991; Qian, 2008). Increasing marriage homogamy on multiple planes, not only age, is taking place against a backdrop of changing gender roles, including increasing sexual parity in the job market (Oppenheimer, 1988).

Social explanations of age hypergamy focus on gender roles and suggest that selection of marriage partners may be dictated by traditional sex stereotypes or gender roles rather than biology (Urberg, 1979). Age itself may be perceived to be negatively correlated to sexual function: high sexual function and therefore younger partners are preferred by both heterosexual and homosexual men (Posner, 1992). Heterosexual women’s desire for older men balances the younger age preference of men in the marriage market (Posner, 1992). Age differences in heterosexual marriages have been declining since 1900. Husbands were, on average, nearly 5 years older than their wives in 1900, but were just over 2 years older in 2000 (Rolf & Ferrie, 2008). Multiple studies using heterosexual and homosexual personal dating ads (Harrison & Saeed, 1977; Kenrick, Keefe, Bryan, Barr, & Brown, 1995; Rasmussen et al., 1998) find that men stipulate ages younger than their own, while the reverse is true for women. In one sample of 48 personal advertisements written by women, advertisers who sought a younger male partner sought a partner who was, on average, two years younger than herself. For those women who sought an older partner, they sought partners who were, on average, 8.4 years older than herself (Rajecki et al., 1991). Conversely, the 65 male advertisers in the same study specified a partner up to 9.8 years younger (Rajecki et al., 1991). In their study of gay and heterosexual dating ads, Kenrick et al. (1995) found that the age of the sought partner was, in part, related to the age of the seeker. Both heterosexual and homosexual men preferred progressively younger partners as they aged, so that men in their 20s preferred partners no more than 4 years older than themselves, and above 50 years of age, they were only interested in partners who are, at maximum, 4 or 5 years younger than them. The cross-cultural nature of patterned age differences in marriage and partnership is well established, but appears to be strongest among heterosexual men who seek younger partners, and this finding has been replicated in over 50 cultures (Buss, Shackleford, & Leblanc, 2000; Symons, 1979).

It has been well established that sexual preference is established in adolescence or before (Bailey & Zucker, 1995). Substantial research exists on gender role behavior and gender expression among both heterosexuals and homosexuals. A gender role preference can be defined as the desire to adhere to cultural standards regarding appropriate masculine and feminine behavior (Callan & Liddy, 1982), a set of expectations about what behaviors are appropriate for people of a particular sex (Kessler & McKenna, 1978), and they also can reflect the social and personal imperative to achieve a normal identity (Wood, Gosling, & Porter, 2007). Gays and lesbians tend to describe themselves as either masculine or feminine and indicate a preference for masculine or feminine mates (Bailey, Kim, Hills, & Linsenmeier, 1997). In his cross-cultural study of dating advertisements, Hou (2002) stated that “… some gays and lesbians offered and sought specific sexual characteristics as well as sincerity and safety features for their ideal partners, and need heterosexual-style role-play for their life” (p. 27). Among gays and lesbians, a culturally-specific vocabulary has emerged to identify persons who act in accordance with either stereotypical masculine or feminine gender roles, including the terms “butch” for masculine persons and “femme” for feminine persons (Maracek, Finn, & Cardell, 1982). As with heterosexuals, some gay men and women adopt a gender role that is more androgynous in appearance and behavior; these individuals may resist labeling or categorization into a particular gender category (Madson, 2000; Rosenzweig & Lebow, 1992; Valentine, 2007).

In summary, the evidence of age hypergamy in both heterosexual and gay populations is well established, although such evidence is weaker and less robust among homosexuals. These results among homosexuals are problematic for sociobiological explanations of age hypergamy, and uncertainty remains as to the source, or driver, of this behavior. In this article, I add to the literature by providing evidence of the presence of age hypergamy in homosexual males when gender roles are taken into consideration.

I propose that the mechanism that is driving age hypergamy among heterosexuals is similar to the one creating age hypergamy in gay couples, if indeed such patterns exist. Rejecting the null hypothesis for H1, H2 or both is indicative that the transmission of patterned age differences in American gay and heterosexual partnerships is not entirely biological, and that a cultural explanation of gender role behavior and preferences may be a more robust explanation.


Gay men who associate more strongly with the traditional American masculine gender role will seek a partner who is similarly younger as heterosexual men desire.


Gay men who identify with the traditional American feminine gender role will seek a partner who is similarly older as heterosexual females desire.



The final sample consisted of 30 heterosexual male advertisers (M age = 37.7 years), 30 heterosexual female advertisers (M age = 39.9 years), 30 gay male advertisers who self-identified as more masculine (M age = 39.1 years), and 30 gay male advertisers who self-identified as more feminine (M age = 38.9 years). For the entire sample of 120 advertisers, the mean age was 38.6 years. Only 25 % of the participants were female.


The sampling universe consisted of 582 online dating advertisements placed between September and December 2003. Data were collected from a total of eight sources, seven of which were online versions of gay/lesbian/bisexual/transgender (GLBT) specific newspapers. The eighth source, the San Francisco Bay Guardian, is an online version of an alternative weekly newspaper that is not specifically aimed at the gay community. This source was chosen because it included personal ads from gay and lesbian persons as well as heterosexuals.

In 2003, the search site, “Queer Information Network” (QIN), listed all currently active GLBT online newspapers in the U.S.4 was explored for potential source papers. Using a purposive sampling method, all online papers listed on QIN that met the following criteria were selected as initial sampling pools:
  1. (1)

    the paper was published completely, or partially, online;

  2. (2)

    the online version of the paper included personal advertisements written by the advertisers (papers including links to dating search engines such as or were not selected.);

  3. (3)

    with the exception of the San Francisco Bay Guardian, the papers catered specifically to a GLBT audience; and

  4. (4)

    the papers were based out of a medium or large urban center.


These criteria severely limited which online papers were available for sampling—specifically criterion 2 relating to the type of personal advertisements used in this research.

After applying the above criteria, the search yielded a list of eight sites: Bay Windows in Boston; Gay People’s Chronicle of Cleveland, Ohio; Just Out in Portland, Oregon; Philadelphia Gay News; Quest in Madison, Wisconsin; San Francisco Bay Guardian; Washington Blade of Washington, DC; and Windy City Times in Chicago. I used a purposive sampling method to select the individual dating advertisements from the source sites during the sampling period. The first 30 advertisements from each of four categories that met the following criteria were selected:5
  1. (1)

    the advertisement must include the precise age of the advertiser;

  2. (2)

    the advertisement must include the precise desired age range of a dating partner; and

  3. (3)

    for the groups with masculine or feminine characteristics, the advertisements must include at least one gender role type specific word.


Each advertiser selected for the sample was required to list a preferred age range of the desired partner plus their own age. If a person only listed a certain age they were seeking, such as: I am a 30 and am seeking a 35 year old partner, then that ad was thrown out. However, if an ad stated: I am 30 and seeking a partner ages 23–32, that ad was selected for the sample. It was acceptable if a person listed an age range in which both the lowest and highest age in the range was higher (or lower) than the seekers own age.

These observations were distributed evenly among the following categories: men seeking men-masculine gender type (MSM-Masculine), men seeking men-feminine gender type (MSM-Feminine), men seeking women (MSW), and women seeking men (WSM). In total, 455 MSM advertisements appeared in these eight sources over the sampling period. Of this total, approximately 5 % met the sampling criteria. Over the sampling period, 93 WSM advertisements appeared in the San Francisco Bay Guardian, and approximately 32 % of these met the sampling criteria. Sixty-four MSW advertisements were posted during the sampling period in the San Francisco Bay Guardian, and 71 % of these met the sampling criteria. The disparity between the percentage of gay men and heterosexuals who met the sampling criteria was a result of more stringent sampling criteria for gay men, for whom a gender preference was required; heterosexuals needed only to list their own age and an age preference to meet the sampling requirements. Table 1 lists the sampled data of MSM advertisements from each source, as this sample provides the main focus for analysis.
Table 1

Sample selection MSM advertisements


Total number of MSM ads

Percentage of total MSM ads

Number of MSM ads in sample

Percentage of total MSM ads used

Bay Windows





Gay People’s Chronicle





Just Out





Philadelphia Gay News










San Francisco Bay Guardian





Washington Blade





Windy City Times










MSM Men seeking men

In order to test hypergamy effects, MSM ads were classified according to the gender-identification of the male placing the ad. The following words constituted “masculine” identification: Butch, Masculine, Masc., Top, and Daddy, while the following words constituted “feminine” identification: Feminine, Femme, Fem, Bottom, and Submissive. The words “Top,” “Daddy,” “Bottom,” and “Submissive” were chosen as surrogates for masculine and feminine gender roles, although in practice these four words may take on different meanings in the gay community than gender roles per se; for some men, they only specify a sexual preference and, for others, they also extend to having more masculine gender characteristics (top, daddy) or more feminine gender characteristics (bottom, submissive) (Carrier, 1977; Wegesin & Meyer-Bahlburg, 2000; Weinrich et al., 1992).6

Additionally, the first 30 ads that met the sampling criteria among the MSW and WSM ads were selected as control groups to validate that age hypergamy exists among heterosexuals in these online dating advertisements and to compare hypergamy patterns between heterosexual and the homosexual groups in the study. In total, 120 ads were used in the study (19.5 % of all advertisements that appeared during the sampling period), classified into four groups: two study groups (MSM-Masculine and MSM-Feminine) and two control groups (MSW and WSM).

I recorded the age of the advertiser and the specific age range of the desired partner for each advertisement that met the sampling criteria in addition to gender role language used. I performed an ANOVA with Tukey post hoc comparisons across the four study groups.

The key dependent variables were preferred age range of partner (both older and younger maximums) and the key independent variable was whether or not the seeker was a heterosexual male, heterosexual female, butch gay male, or femme gay male. The dependent variables were determined by taking a simple mean of all the maximum ages older and younger that seekers in a particular category (e.g., MSM-Masculine) desired (see Table 2). The other variable included was the age of the seeker.
Table 2

Means and SDs: homosexual and heterosexual advertisers






F (df = 3)

Subgroup differences









Mean age seeker











Years older










ab, ac, ad,

Years younger










ab, ad, bc, cd










Post hoc comparisons were conducted using ANOVA. Significant subgroup differences are as: ab MSM-M (male seeking male-masculine) versus MSM-F (male seeking male-feminine), ac MSM-M versus MSW (male seeking women), ad MSM-M versus WSM (women seeking male), bc MSM-F versus MSW, cd MSW versus WSM

p < .10; ** p < .05; *** p < .01


Data collection procedures were as follows. For every ad that met the criteria (age of seeker, specific age range of partner sought, gender-role language for those seekers in those categories), records were kept of the seeker’s age, lowest age sought, highest age sought, and self-gender-role language used (if appropriate).


The first hypothesis was that gay men who adopt the traditional American masculine gender role would desire a partner who was similarly younger as heterosexual men desire, and the second hypothesis was that gay men who adopt the traditional American feminine gender role would desire a partner who was similarly older as heterosexual females desire. Table 2 presents descriptive data on the two study groups and two control groups. All significance levels must have been at the .05 level to be considered significant.

Regarding the first hypothesis, masculine gay men desired a partner who was almost 13 years younger, on average, than themselves, at p < .01. This preference was strikingly similar to heterosexual men, who desired a partner younger by 14 years. Turning to the mean age older that an advertiser desires, masculine gay men preferred a partner who was less than 1 year older, and heterosexual men preferred a partner who was up to 6 years older, this result was also at .01 significance. Preferences aside, however, most heterosexual men are in relationships in which they are, on average, 3 years older than their spouse (Buss, 1989).Homosexual men’s “real-life” partners are very similar to that of heterosexual men in terms of age difference (Kenrick et al., 1995). In essence, what people, both gay and heterosexual, say they prefer in dating ads does not seem to be consistent with real-life relationships.

In the test of the second hypothesis, feminine gay men preferred a partner only 4 years younger, resembling heterosexual females, who preferred partners younger by 3 years, this result was at .01 significance. In comparison, feminine gay males preferred a partner who was up to 11 years older, and heterosexual females preferred a partner up to 10 years older, which is also at .01 significance. These results provide support for H2, which asserts that feminine gay men and heterosexual women will have similar age difference preferences.

Table 2 provides descriptive statistics and the significance of differences between all pairings. The right hand column denotes significant contrasts between specific gender categories.


The strict biological perspective is unable to incorporate my findings, as it only predicts age hypergamy in populations who can reproduce with each other. However, I am faced with an equally vexing problem: I’ve demonstrated that reproductive fitness is not the only parameter that matters in mate selection, but what is the additional driving force? What factor ties heterosexuals and homosexuals together? Reliance on meme theory is one way to explain age hypergamy in both populations.

This article’s focus on homosexuals with self-identified gender role preferences permits an explicit exploration of the impact of gender role on age hypergamy. Because age hypergamy exists among gay males in which there is no possibility of genetic reproduction between the two partners in the dyad, and in which age hypergamy is itself patterned by gender roles within the individual couple, the sociobiological understanding of age hypergamy is an insufficient explanation for the phenomenon in either gay or heterosexual couples. Therefore, we must seek alternative explanations that can explain the behavior in both heterosexual and homosexual couples. Any explanation must account for patterns based on age and gender performance. I rely upon Dawkins’ (2006) concept of the cultural meme as a basis to understand common gender role behaviors. Dawkins’ concept of a meme and the memetic reproduction of gender roles more fully explains the phenomenon among both gay and heterosexual couples.

A meme is conceptualized to be a unit of information akin to cultural DNA (Distin, 2005) that is reproduced through a form of cultural evolution similar to Darwinian evolution (Dawkins, 2006; Distin, 2005). Similar to the biological ideas of reproductive fitness in which the “most fit” genes survive to be passed down to future generations (Dawkins, 2006), memes are deemed to be more or less fit, with a larger or smaller chance of surviving in the meme pool over several generations (Dawkins, 2006; Distin, 2005).

The language of memes and cultural imitation provides a lens through which to view mate selection behavior among both heterosexual and homosexual couples. Mate selection practices, including the gender role differences on which different selection preferences inhere, may persist in the culture even in the absence of a biological imperative, if the selection practice is a form of memetic imitation of gender role behavior. When considered as a set of instructions, or a blueprint, of cultural behavior (Coker, 2008; Dawkins, 1999), the meme, or cultural practice, being explored in this research is the meme that instructs: men desire to marry younger women, women wish to marry older men. Translated into gender-role language so that it is applicable to those in both homosexual and heterosexual relationships, the same meme can be imitated, or reproduced, with a slight variation: masculine people wish to marry younger (feminine) people, feminine people wish to marry older (masculine) people. Memetic imitation therefore does not require an additional biological mechanism for the behaviors’ reproduction; transmission of cultural practices, including age hypergamy, can be memetic as well as genetic. Memetic reproduction permits an explanation of how the behavior of age hypergamy persists among both heterosexuals and homosexuals.

These results, although they provide tentative support for the research hypotheses, could also support a biological explanation for behavior. Evolutionary functions, such as reproduction, can result in conditional decision rules that can be socialized (Kenrick, Li, & Butner, 2003). In particular, just because a behavior is learned or socialized, does not mean it does not also have a biological basis. For example, Kenrick and Keefe (1992) write:

Although evolutionary hypotheses often focus upon genetic explanations, modern theorists assume that organismic predispositions only unfold in interaction with the environment. Evolved mechanisms require environmental inputs to develop, they require environmental inputs to trigger them, and they may be enhanced or inhibited by other relevant factors in the immediate or the developmental environment. Further these evolved mechanisms interact with cognitive factors like attention, expectation, schema activation, and goal orientation. (p. 91)

Sexual strategies theory attempts to explain, using evolutionary language, the adaptive processes that men and women utilize in navigating the evolutionary process of sexuality. This theory explains sexual behavior as an evolved repertoire of adaptations (Buss, 1998). Men and women encounter different evolutionary challenges in short-term as opposed to long-term mating (Buss & Schmitt, 1993). It is unknown if most dating advertisements lead to short- or long-term mating. It is important to note that evolution by natural selection is not the same as survival selection (Buss, 1998). Differential reproductive success is the key to evolution, not differential survival success (Buss, 1998). In other words, my results could be interpreted by evolutionary theory as providing the best reproductive success and not be only a product of socialization. In this case, age differences in dating could be seen as providing differential reproduction strategies that have been transmitted socially to homosexuals, and in fact, is evolutionary in nature and not mere socialization.

Silverthorne and Quinsey (2000) suggested that gay and heterosexual males have similar age preferences for partners. My research suggests that gay males have different age preferences for partners than heterosexual males dependent on whether or not they self-identify as either more butch or more femme. When lumped together as “gay men,” this specificity is hidden, and they may indeed look like heterosexual men, but when they are separated into groups of butch and femme men, their preferences are revealed to be different, based on their gender role preferences.

Patterned age difference preferences exist in both gay and heterosexual populations, and in this article I suggest that these patterns are a result of gender role socialization and memetic reproduction of gendered behaviors and preferences, including preferences about what comprises an ideal mate. Even in populations in which reproductive fitness does not drive mate choices, observed age hypergamy has been attributed to evolutionary causes7 (Sears-Roberts Alterovitz & Mendelsohn, 2009). Memetic replication can be thought of as a form of cultural transmission of behaviors, thoughts, feelings, and ideas (Distin, 2005), and the challenge to a researcher proposing memetic replication as the method of cultural transmission of behaviors across multiple social groups, such as heterosexual men, heterosexual women, and gay men, is to illustrate how this imitation takes place. One possibility is acculturation to heterosexual norms or heteronormativity (Schilt & Westbrook, 2009). As children are raised in the predominantly heterosexual U.S. culture, they are socialized to gender and cultural norms, such as gender roles and their associated mate preferences, including age difference preferences. When young gay men leave the dominant, heterosexual culture and enter the homosexual subculture, social and cultural artifacts accompany them, including gender roles, norms, and expectations. This artifact from the predominant heterosexual culture likely allows age hypergamy to remain in subsets of the gay male population by the introduction of norms from the predominant culture into the different gender constructs that exist in the subculture of gay American men, hence leading to reproduction of behavior that originated among heterosexuals.

Generalizability is a potential concern in this study. Only 5 % of all the advertisements reviewed for this study met the criteria to be included in the study sample. This was a result of the stringent study criteria, all the advertisements had to include the exact age of the advertiser, and a precise age range of the sought partner. In addition, for the gay male advertisers, gender-role language indicating a more masculine or feminine presentation was required, which severely limited the percentage of gay male advertisements that were selected for the final sample. The question remains: How different are advertisers who list precise age (and gender-role information) from those who do not, when it comes to age difference preferences in a sought-after partner? For example, a man who listed his age as “mid-40s” was thrown out of the study, whereas the man who listed his age as 46 was not. Can we assume that these two men have systematic differences in their preferences in the age differences in a desired partner? More research is needed to determine the generalizability of this research to other populations.

There are a number of limitations inherent in the study of anonymous, on-line dating advertisements that must be considered. The population that places on-line dating ads may be more politically and culturally liberal than the mainstream, as placing dating ads may be seen as a deviant activity (Darden & Koski, 1988). While the Internet newspapers and dating advertisements used did not yield demographic data, individuals outside of the middle-and upper-classes may not have easy access to computers and, hence, Internet dating.

Future research should include a personal interview component in order to clarify some of the issues raised by this research. First, adopting a gender role label as either masculine or feminine is a matter of choice in gay communities and is a choice whose label depends greatly on the labeler (Maracek et al., 1982; Rosenzweig & Lebow, 1992; Wegesin & Meyer-Bahlburg, 2000), and greater clarity on what men mean when they self-identify as masculine or feminine could be gained by a personal interview. Another important area in which a personal interview could clarify these results is in the validation of the surrogates used for gender role. It is unclear whether the surrogates selected, specifically “top,” “bottom,” “submissive,” and “daddy,” were appropriate. For some men, this language may only indicate a sexual preference, but, for other men, it may also extend to a gender role preference (Carrier, 1977; Wegesin & Meyer-Bahlburg, 2000; Weinrich et al., 1992). This critical issue needs to be explored further. Even without the aid of personal interviews, research should be extended to break out the advertiser group that identifies as “butch” and “femme” and those that use the “top” and “bottom” language. For example, some men who identify with the masculine gender role also identify with the “bottom” sexual role. The implications of this on choice of partner are unknown. It is possible that the men who state preferences of “top” or “bottom” are seeking substantially different types of relationships in terms of intimacy, sexual content, or permanency than men who state a preference of “butch” or “femme” and would be, along with heterosexual men and women, a poor comparison. This research assumes that a correlation between the two sets of terms (top to butch, and bottom to femme) is generally assumed and accepted by the gay male culture, but additional research is clearly called for to validate this premise.

This research used a very small sample of only 30 individuals in each group. The research would be greatly enhanced if a larger sample size were used. This will be difficult as modern dating ads are mostly on and and these services do not always offer the ability to state the precise age range of the sought partner, but, if possible, this research should be duplicated with a larger sample size.

A final limitation to the research method is whenever one gathers data anonymously, whether online or through dating ads, lying and misrepresentation is possible (Kenrick & Keefe, 1992). Men and women could both lie about their own age, although they less likely to lie about the age of their preferred partner. My data assume that people were honest about their own ages, but only a personal interview component to this research could partially begin to eliminate this limitation of anonymity.

The data for this research were originally collected in 2003, before sites such as and were as common as they are now, and newspaper ads such as the ones cited in this study have now fallen out of favor. I believe that, when it comes to the age differences requested by seekers, the two mechanisms (newspaper ads and computer dating) operate very similarly, and the results would be very similar if the study were to be reproduced today using online advertisements. However, a reproduction of this study using online advertisements is called for to verify this assumption of the applicability of these results to online dating.


Although gay couples may be driven to reproduce and parent together, current technology prohibits the biological reproduction of two members of the same sex; this research only presumes that gay men cannot physically reproduce with each other and makes no comments on the reproductive desires or parenting plans of gay couples. This article only comments on the sex roles of the gay men and how those sex roles may dictate partner selection.


Gay men are not the only population in which the reproductive dynamic is eliminated. A recent study of older adults, including two age cohorts over age 60, found partner-seeking patterns supportive of sociobiological theory, including age hypergamy, even among those persons who were well beyond reproductive age (Sears-Roberts Alterovitz & Mendelsohn, 2009).


One must consider opportunity factors when considering assortative mating. Rich, successful men tend to have more opportunity to choose young, beautiful women, and young beautiful women tend to have more opportunity to choose rich successful men.


By September 2009, this site was no longer active.


Although individuals, both gay and heterosexual, use dating advertisements for establishing long-term relationships as well as short term sexual liaisons, there was no notable difference in stated preference for short term versus long term partnerships among gay men, heterosexual men, and heterosexual women, so this research did not attempt to stratify advertisers by the emotional content or structure of the relationship sought.


There were eight advertisers who self-identified as both butch and bottom, or femme and top. These ads were not included in the final sampled group as their meaning was unclear and could not be clearly coded and classified; however, their presence suggests it may be culturally widely accepted that the use of one is a proxy for the other-hence the need for an advertiser to specify when he self-identifies as both.


This research specifically examined dating ads and therefore reveals age difference preferences only for dating partners. Accomplished age differences between married partners, both gay and straight, were not examined in the current research. However, some research suggests that there are meaningful differences in the partner selection processes for cohabitating and marriage partners (Blackwell & Lichter, 2000).



The author gratefully acknowledges the close readings and helpful conceptual conversations with Sharon Methvin, Lori Holder-Webb, Richard Williams, and Sarah Rosenfield. Additionally, she gratefully acknowledges the statistical support from Deborah Carr.

Copyright information

© Springer Science+Business Media, LLC 2012