Self-Presentation, Desired Partner Characteristics, and Sexual Behavior Preferences in Online Personal Advertisements of Men Seeking Non-Gay-Identified Men

Abstract

Despite attention to the sexual behaviors of non-gay-identified (NGI) men who have same-sex encounters, virtually no research has focused on issues of partner desirability and selection. Limited evidence suggests that a subgroup of men who have sex with men (MSM) advertise online for sexual encounters with NGI men. Exchange theory provided a framework to investigate this seeking of NGI men, based on the content of Internet personal advertisements for same-sex encounters. Researchers analyzed 282 ads posted to an online bulletin board. Ads by men who explicitly desired encounters with NGI men were compared with those by men who did not indicate this preference in potential partners. Multivariate analyses revealed that NGI-seeking men had significantly increased odds of identifying as discreet (Adjusted odds ratio [AOR] = 2.82), seeking a discreet encounter (AOR = 4.68), seeking a masculine partner (AOR = 2.18), being willing to host (AOR = 2.77), as well as seeking oral-receptive sex (AOR = 2.69), unprotected oral sex (AOR = 6.76), and anal-receptive sex (AOR = 2.18). Further, NGI-seeking ads were more likely to not mention condom use or safer sex practices (AOR = 4.13) and were less likely to indicate a desire for oral-insertive sex (AOR = 0.34) and rimming (AOR = 0.21). Findings suggest that some men may deliberately present themselves in ways that they perceive as being attractive to NGI men, and have research implications for NGI MSM, their partners, and the risk outcomes of these online ads.

Despite attention to the sexual behaviors of non-gay-identified (NGI) men who conceal their sex with other men, especially from female sexual partners (Benoit & Koken, 2012; Hightow et al., 2006; Prabhu, Owen, Folger, & McFarland, 2004; Reback & Larkins, 2013; Schrimshaw, Downing, & Siegel, 2013; Zellner et al., 2009), virtually no research has focused on issues of partner desirability and selection. There is reason to suspect (Cheeseman, Goodlin- Fahncke, & Tewksbury, 2012; Ward, 2008) that there exists a subgroup of men who have sex with men (MSM) who actively seek out NGI men for sexual encounters, either because they themselves do not identify as gay (and want a similar partner who will be motivated to be discreet and non-disclosing about their same-sex encounters) or because NGI men are perceived by some gay men to have other desirable qualities (e.g., they are masculine, dominant, “straight - acting”).

Supporting the perceptions that NGI men are more masculine, dominant, and “straight - acting” compared to openly gay men are data on the sexual practices these men report when discussing their same-sex encounters or relationships. Several studies have found a higher prevalence of oral-insertive acts (recent or lifetime) than oral-receptive acts among behaviorally bisexual male samples, including NGI men (Dodge et al., 2013; Reback & Larkins, 2013; Siegel, Schrimshaw, Lekas, & Parsons, 2008; Stokes, Vanable, & McKirnan, 1997). The majority of research further suggests that the prevalence of anal-insertive sex (both in general and unprotected) is higher than that of anal-receptive sex among these men (Dodge et al., 2013; Siegel et al., 2008; Spikes et al., 2009; Stokes et al., 1997; Wheeler et al., 2008; Zellner et al., 2009; Zule et al., 2009), although not universally (Mercer, Hart, Johnson, & Cassell, 2009; Reback & Larkins, 2013). This preference for insertive behaviors may not only contribute to the perception that these men are more masculine and dominant, but also may serve to preserve their own self-image as masculine and heterosexual.

Exchange Theory and the Virtual Sexual Marketplace

The Internet has become a dominant strategy for individuals to meet sexual partners (Chiasson et al., 2006; Downing, 2011; 2012; Grosskopf, Harris, Wallace, & Nanin, 2011; Grov, Parsons, & Bimbi, 2007; Lever, Grov, Royce, & Gillespie, 2008). This is particularly true of MSM, who have been found to use the Internet extensively for sexual partnering, in part because it allows them an easy and efficient method to search for partners with specific qualities they are looking for (e.g., sexual role preferences, physical types) and provides opportunities to negotiate the details of sexual encounters prior to meeting (Brown, Maycock, & Burns, 2005). Indeed, the Internet is widely used by individuals seeking specific sexual subgroups or behaviors (e.g., bareback sex, HIV-positive, younger men; Balán, Carballo-Diéguez, Ventuneac, & Remien, 2009; Klein, 2012; Moskowitz & Roloff, 2007; Sowell & Phillips, 2010) that might be more challenging to find in an offline venue such as a gay or mixed-sexuality bar. One sexual subgroup in particular, NGI MSM, uses the Internet because it affords them the opportunity to search for anonymous sexual encounters while being discreet and potentially avoiding discovery (Ross, Rosser, McCurdy, & Feldman, 2007; Schrimshaw, Downing, et al., 2013). For example, Ward (2008) found evidence that straight-identified MSM constructed advertisements (hereafter referred to as ad or ads) that would potentially appeal to desired partner types. This literature would suggest the utility of an exchange theory framework to explore the ways MSM present themselves to prospective sexual partners through Internet-based personal ads and whether these presentations differ if they are looking for NGI MSM.

Exchange theory postulates that individuals knowingly bring certain assets into an interaction or relationship that they use to bargain for valued commodities (Collins, 1994). Each party to the exchange performs a cost-benefit analysis with interactions likely to continue provided each judges his or her potential gains to outweigh potential losses (Baumeister & Vohs, 2004). An exchange theory framework has particular relevance for a number of sexual issues including partner selection and the negotiation for initiating a sexual relationship (Sprecher, 1998). Indeed, researchers have proposed a sex exchange theory that builds on this framework and argues that for women sexual activity is a resource offered to male partners in exchange for financial security, love, or commitment (Baumeister & Vohs, 2004; Vohs, Catanese, & Baumeister, 2004). Although some have suggested that exchange theory requires individuals to have specific roles as “buyer” and “seller” (e.g., the male provides financial stability in exchange for sex), a more egalitarian view may be offered (especially in contemporary society where women are not financially dependent on their male partner but seek other assets from their partners). Thus, financial exchange is not required and that both parties to the exchange may bring various desirable qualities in addition to financial resources and sexual behavior.

While some researchers have questioned the applicability of exchange theory to same-sex relationships (Baumeister & Vohs, 2004) due to a presumed absence of inherent sexual marketplace-defined roles (i.e., buyer and seller), this contention fails to consider the importance that MSM place on their own as well as their partner’s sexual roles (e.g., top-bottom, dominant- submissive, masculine-feminine) during sexual encounters (Johns, Pingel, Eisenberg, Santana, & Bauermeister, 2012; Wei & Raymond, 2011). Likewise, as we argued earlier, buyer and seller roles are not critical to a sexual exchange framework. Just like heterosexual couples, both individuals of a potential same-sex couple may bring a variety of (not necessarily financial) assets that one prospective partner may find attractive in the other partner. Regardless of gender, the relationship may be viewed as attractive when both individuals see the exchange as beneficial.

Evidence that MSM believe they bring (non-financial) assets to the sexual marketplace in bargaining for potential partners exists in the personal ads men place in gay-oriented publications (Bailey, Kim, Hills, & Linsenmeier, 1997; Gonzales & Meyers, 1993) and telephone personals (Bartholome, Tewksbury, & Bruzzone, 2000), and more recently, in personal ads or profiles on Internet websites designed to meet casual sexual partners (Cheeseman et al., 2012; Ward, 2008). As Bartholome et al. (2000) suggest, “By conducting their search on paper, online, or via telephone sources, advertisers can stress the qualities that are most valued by them. By addressing specific qualities the advertiser is more likely to receive responses that close ly match their interests.” (p. 310).

The growth of Internet websites for sexual partnering has created a virtual sexual marketplace where MSM can advertise the assets they would bring to a sexual encounter in exchange for those offered by a prospective sexual partner (Chiasson et al., 2006; Paul, Ayala, & Choi, 2010; Ross, 2005). Such assets might include youthfulness or maturity (Slevin & Linneman, 2010; Sowell & Phillips, 2010; Steinman, 1990), masculinity (Bailey et al., 1997; Connell, 2005; Drummond, 2005; Filiault & Drummond, 2007), fitness or muscularity (Brennan, Craig, & Thompson, 2012; Drummond, 2010; Martins, Tiggemann, & Churchett, 2008), physical endowment (Chiasson et al., 2006; Drummond & Filiault, 2007; Grov, Parsons, & Bimbi, 2010; Ward, 2008), sexual role preferences (Chiasson et al., 2006; Wei & Raymond, 2011), a specific HIV serostatus (Grov & Parsons, 2006; Moskowitz & Roloff, 2007), as well as preferences for risky or safer sexual encounters (Chiasson et al., 2006; Downing, 2011; Grov, 2010; Klein, 2012). To date, however, limited research has examined (Cheeseman et al., 2012; Ward, 2008) the use of a heterosexual or bisexual identity, heterosexual marriage, or ‘down low’ image as potential assets in the sexual marketplace.

Self-presentation is an integral component to the design of online personal ads. The provision of often extensive self-descriptions suggests that personal ads are not merely a statement of preferred partner preferences. Rather, they imply that online personal ads are created in ways that are believed will attract partners with certain qualities. It has been noted that “connecting successfully with men on the Internet is reliant upon one’s ability to sell oneself as a product—with hook-ups facilitated by an awareness of one’s bartering power in a given sexual marketplace” (Paul et al., 2010, p. 7). For instance, researchers have argued that the use of phrases that evoke a “straight-acting” persona when creating same-sex online profiles or personal ads represent a desire by the poster to be perceived as masculine (Payne, 2007; Phua, 2002; Ward, 2008). The assumptions being made are that men need to give signals or indications that they are “straight-acting” in order to attract a certain type of partner; or that prospective partners prefer someone who acts straight. As such, users will “advertise themselves in ways suitable to market forces and with particular kinds of consumers in mind” (Payne, 2007, p. 531). Indeed, researchers have suggested that some gay men consider masculinity to be a valuable commodity in themselves and in their partners (Bailey et al., 1997; Connell, 2005; Drummond, 2005; Filiault & Drummond, 2007). Moreover, Ward (2008) found evidence that straight-identified men actively claim the ‘down low’ label (i.e., men who have covert sex with other men), and suggested that this is another way they maintain or “affirm” their heterosexual masculinity.

MSM with characteristics that are devalued in the sexual marketplace will perhaps have less bargaining power (or at least perceive themselves as having less power) and therefore may be willing to place themselves at risk for HIV and other sexually transmitted diseases (STD) in order to obtain a sexual partner. Given the value that some gay men place on thinness (Martins et al., 2008) and their desire to lose weight (McArdle & Hill, 2009), a recent study found that MSM with a higher body mass index were less likely to reject a sex partner and reported less consistent use of condoms, though with fewer anal sex partners (Moskowitz & Seal, 2010). Researchers have also reported that MSM will bargain with and for specific racial or ethnic backgrounds when seeking sex partners online (Callander, Holt, & Newman, 2012; Paul et al., 2010; Ward, 2008; Wilson et al., 2009). However, ads containing a preference for—or a rejection of—certain racial and ethnic groups may be perceived by members of those groups as discriminatory, resulting in a devalued sense of self (Wilson et al., 2009). As such, some ethnic minority men may place themselves at risk in order to obtain a sexual partner from the Internet (Han, 2008). Recent qualitative findings with an ethnically diverse sample of MSM suggest that stereotyping of African American men as sexually dominant, and Latino men as low risk for HIV/STDs may contribute to sexual risk behavior (Rhodes et al., 2011). Further, the desire to preserve one’s masculinity or manhood through sex may lead to risk taking behavior (Rhodes et al., 2011). The authors found that some men will use condoms inconsistently and may increase their number of male and female sexual partners in order to be perceived as a “real” man (p. 145).

This study applied an exchange theory framework to investigate how MSM communicate through the Internet to find prospective sexual partners with desired characteristics and how this could lead to risky or safer encounters. Specifically, this report examines the existence of a subgroup of men who advertise for sexual encounters with NGI male partners (i.e., heterosexual or straight, bisexual, heterosexually married, ‘down low’). Content analysis procedures were used to describe the prevailing characteristics of men who placed these ads, to identify their preferred characteristics of male partners (i.e., perceived assets), and to assess the types of sexual behaviors and level of HIV/STD risk that advertisers desired to engage in with prospective partners. Subsequent analyses compared the content of personal ads by NGI-seeking men with that of men who did not indicate a preference for the sexual identity of potential partners.

Method

Sample Selection

To examine the existence of a subgroup of NGI-seeking men in the online sexual marketplace, a total of 282 Internet personal ads were selected and subjected to content analysis (Krippendorff, 2004; Weber, 1990) in March 2009. Personal ads are particularly useful for understanding the self-presentation of individuals and their perceptions of an ideal sexual partner and sexual encounter. Thus, even if the eventual sexual behavior does not exactly reflect the personal ad, it does provide an understanding of the sexual fantasies and the expressed intentions of the ad poster as communicated to potential partners. Likewise, personal ads achieve this without the self-presentation biases that would be introduced by the researcher (Gonzales & Meyers, 1993).

This study utilized ads posted to Craigslist, a publicly accessible and highly-trafficked online bulletin board. Craigslist users can post virtually any type of ad including for real estate, employment, volunteer services, and personals. It was selected as the website for data collection because it is cost-free and facilitates expedited interactions between users by not requiring a membership to view, post, or reply to ads. Craigslist has been widely used to find sexual partners both by gay (Downing, 2011; Grov, 2010) and non-gay-identified MSM or men who have sex with both men and women (MSMW; Schrimshaw, Downing et al., 2013; Ward, 2008). Indeed, the New York City “men seeking men” personals section posts two to four thousand ads each day (Grov, 2010). Furthermore, researchers have successfully used Craigslist to recruit MSM and MSMW participants (Downing, 2012; Grov, 2012; Moskowitz & Seal, 2010; Schrimshaw, Siegel, Downing, & Parsons, 2013).

Traditional Internet dating websites typically require users to respond to a series of predetermined questions (e.g., age, race/ethnicity, height, weight, body type, HIV status) that are subsequently utilized to build an online dating profile. Craigslist does not utilize any predetermined questions. Alternatively, users write an ad that describes, in as few or as many words as they choose to use, what they are looking for in a partner and/or sexual encounter (see Measures for examples of Craigslist personal ads). The requirements to post in the personals category of Craigslist are: (1) acknowledgment of website terms of use (be at least 18 years old); (2) selection of a single category for dating, romance, long term relationship or sex with no strings attached; (3) selection of a subcategory (i.e., a man or men seeking another man or men); (4) an indication of geographic location including one’s neighborhood or borough; and, (5) a title and description of the ad.

Procedure

The research team systematically reviewed and collected ads from the New York City section of Craigslist. We focused on two of the website’s personals categories, Men Seeking Men and Casual Encounters, which are frequently used to solicit and coordinate sexual encounters. All five boroughs of New York City were represented in this section as well as neighboring areas including Long Island, Westchester, New Jersey, and Connecticut). No attempts were made to contact the posters of any personal ads. Further, no private or personally identifiable information (e.g., emails, phone numbers) was obtained, as Craigslist does not permit users to include such details in ads for personals categories. The Institutional Review Board for Columbia University approved the study protocol.

The authors read through 1194 ads posted during 11 different 1-hour time blocks across both Craigslist categories. Data collection times were assigned in 1-hour time blocks, randomly selected each day and varied from early afternoon to late evening and early morning hours (i.e., after midnight) so as to ensure a more representative sampling of ads. Researchers reviewed all ads posted during an assigned time block. Those that upon review appeared to be posted by a man seeking a sexual encounter with a NGI man (i.e., looking for a straight/“str8”, bi/bisexual, married, curious, “DL”/down low man) were printed for later content analysis. In addition, for the purpose of having a comparison group, the first ineligible ad to appear following each NGI-seeking ad was also collected. If two NGI-seeking ads appeared together, the next two ineligible ads were collected. However, in a few instances, an NGI-seeking ad appeared at the end of the 1 - hour block, thus a comparison ad was not available. Researchers maintained a record of the number of ads that were reviewed during each 1-hour time block.

Measures

Following standard content analysis procedures (Krippendorff, 2004; Weber, 1990), the coding system for this study was developed prior to data collection based on a preliminary set of Craigslist ads. During this pilot phase, the authors independently read and classified 169 ads that were posted to the Men Seeking Men category over a 1-hour period. Inter-rater reliability was assessed as part of the pilot phase using Cohen’s kappa (κ) to verify that the researchers were consistent in determining which ads sought NGI men and which did not. The findings revealed modest agreement between the authors on those ads that appeared to be posted by a man seeking a sexual encounter with a NGI man (κ = 0.77). Coding discrepancies were identified and discussions were held to resolve differences and ensure greater consistency in subsequent coding.

The research team developed a coding sheet based on this preliminary review to systematically extract information on the following variables: demographics of poster (age, race), preferred demographics of partner (e.g., older or younger, race), sexual identity of poster and prospective partner, characteristics of poster (e.g., discreet, masculine, body type, cut or uncut penis), exchange of money for sex, location of encounter (poster or partner hosts at his home, hotel room, office, etc.), preferred sexual behavior(s), condom use or non-use, health of poster (HIV status and STD history), preferred health of partner, and substance use (e.g., in the context of a sexual encounter, poster is drug free, partner should be drug free). Sexual behaviors that were coded include: oral-insertive and receptive (e.g., blowjob, suck), anal-insertive and receptive (e.g., top, bottom), rimming, masturbation (e.g., jackoff), massage or body contact, the use of sex toys, spanking or discipline, and watersports (e.g., piss, urinate). Unprotected oral sex was coded whenever an ad made reference to ejaculation in the mouth using colloquial terms such as cum, load, or swallow. Ads that used terms such as bareback and raw were coded as unprotected anal sex. Rimming was coded whenever an ad made reference to oral-anal contact using such terms as rim, eat, or lick.

The following examples of ads demonstrate the context in which study variables were coded. Italicized words or phrases were coded as descriptors of the poster, and boldfaced words or phrases were coded to describe a prospective partner, preferred type of encounter, and sexual behavior preferences.

NGI-seeking ads

(1) Seeking Bi/Str8 Top/Vers Guy. I Host. No Pic Required. Discreet. – 55 (Location). Hi, and thanks for taking the time to read my post. Well-mannered, refined, safe, sane guy seeking partner-in-crime today or preferably for ongoing good times. Am smaller in stature but have a defined, hairy body. D & D free; alcohol and poppers OK, otherwise no drugs. Married is fine, open to all ages and races, and am even open to couples. Am masculine, non-scene, and will protect your privacy with ease; I’ll expect the same in return. Am not interested in strict bottoms. I have no photo and will not require one of you. Safe only. Hope to hear back from you soon. (Posted in Men Seeking Men, 4:42 P.M.)

(2) Amazingly Cocksucker Is Here In (Location). Any top needs a skillful cocksucker for chilling?? I am the one you are looking for…like to suck off, fuck my face, cum on my face or in mouth. Black or Latino and bi or str8 only!!! You be masc, discreet, noooo fem!!!. If you are fem, do not responding!!! My stats: very masc, discreet, handsome, body shape, 6’1”, 170 lb, 8” cut cock, HIV-, STD-, can host on (street). (Posted in Men Seeking Men, 9:17 P.M.)

(3) COME BY AND BE HUNG AND FUCK MY MOUTH HARD! – 22 (Location). I’m 6’1 tall great face, into sports, great body. Very discreet here with a gf!! Into young hot guys ONLY… U have to be str8 or bisex good body, normal SDD FREE NEGATIVE and u r able to come to my room… lights r very low take your dick out… be dom top and fuck the shit out of my mouth! It will be ONLY our secret! U have to be MAX 30 years old Be hung, thick and a good fucker and lets do this! poppers r ok NO DRUGS! Send body and face pics with your info in your first email!!!!!!!! If this is still on means “I’m looking” and still didn’t find the right guy for this :-) So once again u come by no talking drama… take your dick out… u r enjoying fucking guys mouth… we take poppers u fuck hard… and when we r done u leave!2 (Posted in Men Seeking Men, 12:54 A.M.)

(4) Looking? Curious? (Location). Is your wife or girlfriend not satisfying your sexual needs? Or perhaps you have thought about or been curious about what it would be to have a sexual experience with another man? A man knows exactly how to please another man. I think that I can help if you allow me to. My preference is ages 18–35 but not limited to that, uncut a plus and you must be good-looking. I am very discreet and can be DL for you. Must be clean, d/dfree, honest, and very respectful as I am. I am able to host at my place in [Location]. I enjoy giving oral (no reciprocation), HJ, and open to anything that will satisfy your sexual needs and hunger. Endless emails will be ignored. Also do not ask for my pic UNLESS I have yours first. If I appeal to you get back to me with your age and stats and how I can satisfy you. I am 5’7”, 160, 7”C, Hope to hear from you soon as I am available today or tomorrow. (Posted in Men Seeking Men, 11:52 A.M.)

Comparison ads

(1) Uncut looking for now – 30 (Location). 30, uncut, 175 lbs, versatile bottom looking for action tonight. disease free, clean safe action. looking for verbal guys. mild to wild. pic for trade. yours will get mine. (Posted in Men Seeking Men, 9:08 P.M.)

(2) Arse to fuck – m4m – 29 (Location). Guy Fit 29 masculine horny arse love to ride nice big cocks, avail tonight 6 till 11 pm and tomorrow till 3pm prefer hung guys white black or latinos, group also possible. (Posted in Casual Encounters, 4:53 P.M.)

(3) Hung Hairy Athletic Jock for Tight Ass Bottom – m4m – 29 (Location). Hanging at home and horned. Need a cute tight bottom to go to work between my legs, maybe more. Me: 5’9”, 160 lbs, toned athletic top, hairy, a true 7.5c and thick. Masculine and verbal Italian. You: My height or shorter, lean, young, tight bottom who can suck deep and, if it goes there, can get fucked deep as well. Safe only, d&d free, looking for same. Send face pic in first email or no response. (Posted in Casual Encounters, 9:05 P.M.)

(4) Take on my delicious ass – 41 (Location). Looking for a negative dude who wants to eat and fuck ass safely. No BB needed. Tight ass, nicelyu [sic] hung (UC) bottom here. 5’10”, 175-pounds, toned body, 8”uc. Lots of making out and sucking before you flip my legs over your shoulder. A fuck buddy situation would be great. Let me know. I prefer to travel. Send stas [sic] and pic. (Posted in Men Seeking Men, 4:17 P.M.)

Data Analysis

Descriptive statistics were computed for all variables. Bivariate analyses were conducted to assess differences between NGI-seeking and comparison ads: (1) an analysis of variance test to examine differences by poster age; and (2) a chi-square test to examine differences by website category, neighborhood, demographics, online sexual marketplace characteristics, and sexual behavior preferences. Analyses were not conducted when two or more cell counts were less than 10 or one cell count was less than 4. Demographic variables and online sexual marketplace characteristics with a p-value of < .10 at the bivariate level were entered into a logistic regression model to further assess the differences in sexual partner preferences between NGI-seeking and comparison ads. Sexual behavior preferences with a p-value of < .10 at the bivariate level were entered into a second logistic regression model to further assess behavioral differences between NGI-seeking and comparison ads. We conducted a post-hoc power analysis to ensure that the logistic regression models were sufficiently powered to detect differences in the least prevalent (and therefore the least powered) exchange characteristic (i.e., hairy) and sexual behavior preference (i.e., rimming). We found that despite the small differences observed on these variables between NGI-seeking and comparison men (3–4% vs. 10%) and accounting for the model variance already explained by other variables, the analytic sample of 282 still had sufficient power (> .98) to assess the significance of additional variables in both of the logistic regression models.

Results

A total of 1194 ads were read during the study period (Men Seeking Men = 1027; Casual Encounters = 167). Among these, 137 (11%) were ads placed by men seeking NGI sexual partners. Fourteen additional ads that sought NGI and gay-identified men were collected, however, upon further review they were not included in the NGI-seeking sample in order to more clearly examine differences between NGI-seeking and non-NGI-seeking men. A greater proportion of NGI-seeking ads were collected from Casual Encounters (16%) than Men Seeking Men (11%), χ² (1, N = 1194) = 4.21, p < .05. In addition, the research team collected 145 (12%) comparison ads. There was no difference in the proportion of comparison ads collected from Casual Encounters (12%) and Men Seeking Men (12%). Furthermore, there were no geographic differences between NGI-seeking and comparison ads [Manhattan (61% vs 70%); all other NYC locations (39% vs 30%); χ² (1, N = 282) = 2.56, ns].

Comparisons between ad groups

Sample demographics and online sexual marketplace characteristics

Table 1 presents bivariate comparisons of demographic information and online sexual marketplace characteristics between the NGI-seeking and comparison ads. Men who posted an NGI-seeking ad were more likely to self-identify as bisexual, married, and/or discreet (p < .05), and to seek out a discreet encounter (p < .001). Men who posted a comparison ad were less likely to indicate their sexual identity (p < .001). Men who posted ads for NGI male partners were similar in age to those who posted comparison ads (M = 34.29, SD = 8.37 vs M = 33.49, SD = 8.42), F(252) = .57, ns. There were no group differences regarding the race or ethnicity of the poster.

Table 1.

Bivariate Comparisons of NGI-seeking Advertisements (n = 137) and Comparison Advertisements (n = 145) on Demographics and Online Sexual Marketplace Characteristics

		Characteristics of Poster			Characteristics of Desired Partner
Racea		NGI	Comparison	χ2	NGI	Comparison	χ2
White		27 (20)	33 (23)	2.26	13 (10)	9 (6)	NA
Non-White		15 (11)	23 (16)		11 (8)	9 (6)
Seeks White and Non-White		---	---		3 (2)	6 (4)
Did not identify/No preference		95 (69)	89 (61)		110 (80)	121 (83)
Age
Younger		---	---		13 (10)	16 (11)	0.18
Older		---	---		9 (7)	4 (3)	NA
Sexual identityb
Straight/Heterosexual		6 (4)	5 (3)	NA	61 (45)	0	---
Bisexual		27 (20)	6 (4)	16.53***	78 (57)	0	---
Gay/Homosexual		5 (4)	3 (2)	NA	0	1 (1)	---
Curious		2 (2)	1 (1)	NA	15 (11)	0	---
Down low		0	1 (1)	NA	14 (10)	0	---
Did not identify		90 (66)	127 (88)	19.04***	0	144 (99)	---
Married		15 (11)	6 (4)	4.74*	68 (50)	0	---
Discreet		49 (36)	17 (12)	22.71***	37 (27)	10 (7)	20.51***
Masculine		29 (21)	33 (23)	0.10	36 (26)	18 (12)	8.75**
Very good looking		26 (19)	26 (18)	0.05	9 (7)	5 (3)	NA
Body type
Regular		7 (5)	1 (1)	NA	5 (4)	1 (1)	NA
Hairy		6 (4)	14 (10)	2.98†	6 (4)	4 (3)	NA
Athletic, in shape, muscular		37 (27)	38 (26)	0.02	22 (16)	20 (14)	0.29
Can host		57 (42)	33 (23)	11.52**	---	---	---
Cannot host		23 (17)	41 (28)	5.30*	---	---	---
Generous		3 (2)	3 (2)	NA	1 (1)	14 (10)	NA
Penis
Cut		21 (15)	23 (16)	4.67†	1 (1)	2 (1)	NA
Uncut		4 (3)	13 (9)		7 (5)	1 (1)	NA
HIV/STDs
Disease free/Healthy		34 (25)	36 (25)	0.00	22 (16)	20 (14)	0.29
HIV-negative		17 (12)	26 (18)	1.66	20 (15)	17 (12)	0.51
HIV-positive		1 (1)	0	NA	0 (0)	1 (1)	NA
Did not indicate		115 (84)	119 (82)	0.18	113 (83)	128 (88)	1.55
Drug free		23 (17)	27 (19)	0.16	20 (15)	13 (9)	2.16

Notes: n (%) are presented. NGI (ads seeking a non-gay-identified partner);

^aChi-square analysis included White, Non-White, and Did not identify categories; df = 2. Latino, Black, Asian and Other/mixed race were collapsed into Non-White for chi-square analysis due to small numbers of ads that identified the poster’s or potential partner’s racial/ethnic group.

^bFourteen additional ads that sought both NGI and gay-identified men were collected, however, they were excluded from the analyses in order to more clearly examine differences between NGI-seeking and non-NGI-seeking men.

NA = not analyzed due to low cell counts. Dashes (---) = data not collected (e.g., host) or analysis was not appropriate (e.g., groups were not compared on sexual identity of partner sought because the groups were created based on this characteristic).

^***p < .001,

^**p < .01,

^*p < .05,

^†p < .05,

Although there were no group differences on the prevalence of men who identified themselves as masculine, significantly more NGI-seeking men sought encounters with masculine men (p < .01). NGI-seeking ads were more likely than comparison ads to indicate the poster’s willingness to host a sexual encounter (p < .01) and less likely to indicate an inability to host (p < .05). There were no group differences in the proportion of posters who identified as “very good looking,” athletic, in shape, or muscular. Similarly, there were no differences in the percentage of posters who sought sexual encounters with athletic, in shape, or muscular men. The two groups did not differ in terms of drug status (“drug free,” drugs ok) for either the poster or his desired partner. Furthermore, there were no differences between NGI-seeking and comparison ads regarding the indication of disease status for the poster (poster stated his HIV status, poster stated that he is “healthy” or “disease free”) or desired partner (poster stated that partners should be “disease free”).

Those demographics and online sexual marketplace characteristics with a p-value of < .10 at the bivariate level were then included in a logistic regression model. Table 2 presents the role of each independent variable in its ability to classify the type of ad (NGI-seeking vs comparison). The full model containing all predictors was statistically significant, χ² (11, N = 282) = 85.33, p < .001, indicating that the model was able to distinguish between both groups. Relative to comparison men’s ads, men who posted NGI-seeking ads were 2.8 times more likely to describe themselves as discreet, were 4.7 times more likely to indicate that they sought a discreet partner or encounter, and nearly three times more likely to be able to host a sexual encounter. Additionally, these NGI-seeking men were more than twice as likely to seek a masculine partner and less likely to describe their body type as hairy or their penis as uncut relative to comparison men’s ads.

Table 2.

Multivariate Comparisons of NGI-seeking and Comparison Men’s Advertisements on Demographics and Online Sexual Marketplace Characteristics

	AOR	95% CI
Bisexual (P)	2.64	0.68 – 10.27
No sexual identity (P)	0.38	0.13 – 1.11
Married (P)	1.26	0.31 – 5.12
Discreet (P)	2.82**	1.39 – 5.73
Hairy (P)	0.16**	0.05 – 0.57
Can host (P)	2.77**	1.47 – 5.23
Cannot host (P)	0.65	0.31 – 1.37
Uncut penis (P)	0.22*	0.05 – 0.90
Discreet (S)	4.68**	1.96 – 11.21
Masculine (S)	2.18*	1.04 – 4.58

Notes: (P) = Characteristics the poster used to describe themselves, (S) = Characteristics sought in a potential partner. AOR = adjusted odds ratio. Nagelkerke R square value = 0.35.

^**p < .01,

^*p < .05.

Sexual behavior preferences

Table 3 presents bivariate comparisons of sexual behavior preferences between the NGI-seeking and comparison ads. As shown, men who sought sexual encounters with a NGI male partner were more likely than comparison men to desire oral-receptive sex (p < .001), unprotected oral sex (p < .01), and anal-receptive sex (p < .05). However, NGI seeking men were less likely to desire massage or body contact and rimming than comparison men (p < .05). Additionally, NGI-seeking ads were less likely than comparison ads to indicate a preference for safer sex or condom use (p < .01). Moreover, NGI-seeking men were more likely than comparison men to leave out any mention of, or requests for safer sex or condom use (p < .001). Regarding substance use during sexual encounters, a similar proportion of ads in both groups indicated that this was acceptable to, or desired by the poster (7% vs 7%).

Table 3.

Bivariate and Multivariate Comparisons of NGI-seeking and Comparison Men’s Advertisements on Sexual Behavior Preferences

	NGI-seeking (n = 137)	Comparison (n = 145)	χ2	AOR	95% CI
Masturbation	25 (18)	21 (15)	0.73
Massage/Body contact	10 (7)	22 (15)	4.34*	0.61	0.25 – 1.49
Oral-receptive	69 (50)	29 (20)	28.64***	2.69**	1.46 – 4.99
Oral-insertive	9 (7)	18 (12)	2.78†	0.34*	0.13 – 0.87
Anal-receptive	28 (20)	17 (12)	3.99*	2.18*	1.03 – 4.62
Anal-insertive	14 (10)	21 (15)	1.18
Rimming	4 (3)	14 (10)	5.35*	0.21*	0.06 – 0.73
Othera	27 (20)	33 (23)	0.39
No specified sexual behaviors	22 (16)	26 (18)	0.18
Safe sex/condom use	21 (15)	48 (33)	12.04**	1.25	0.45 – 3.51
Did not indicate safe sex/condom use	90 (66)	61 (42)	15.81***	4.13**	1.60 – 10.65
Unprotected oral	19 (14)	7 (5)	6.88**	6.76**	1.66 – 27.46
Unprotected anal	1 (1)	3 (2)	NA

Notes: n (%) are presented. NGI (advertisements seeking a non-gay identified male partner);

^a(watersports, sex toys, spanking or discipline, etc.).

AOR = adjusted odds ratio. Nagelkerke R square value = 0.28. NA = not analyzed due to low cell counts.

^***p < .001,

^**p < .01,

^*p < .05,

^†p < .10.

Those sexual behavior preferences with a p-value of < .10 at the bivariate level were then included in a second logistic regression model (Table 3). The full model containing all predictors was statistically significant, χ² (8, N = 282) = 66.78, p < .001, indicating that the model was able to distinguish between the two groups of ads. Relative to comparison men, men who posted NGI-seeking ads were 2.7 times more likely to seek oral-receptive sex and more than twice as likely to seek anal-receptive sex. However, these men were less likely to seek oral-insertive sex or rimming than comparison men. Furthermore, NGI-seeking men were over four times more likely to not mention safer sex or condom use in their ads and 6.8 times more likely to seek unprotected oral sex relative to comparison men.

Discussion

Despite interest in the sexual lives of NGI men who have same-sex encounters, there has been almost no research focusing on issues of partner desirability and selection. This report sought to identify the existence of a subgroup of men who actively advertise on the Internet for sexual encounters with NGI men. Contrary to suggestions that an exchange theory framework is not suitable to understanding same-sex relationships (Baumeister & Vohs, 2004), our findings strongly support the assumptions of exchange theory. Specifically, our study found evidence that men who desire sexual encounters with NGI partners will market or “sell” themselves in ways that may be perceived as accommodating to the needs of NGI men.

In seeking an NGI male partner, certain characteristics or assets were considered to be particularly important above others. For instance, a central concern to men who conceal their same-sex behavior is their need to effectively search for, and engage in sexual encounters while maintaining anonymity and avoiding discovery (Schrimshaw, Downing et al., 2013). Our NGI-seeking ads were more likely than comparison ads to describe the poster as discreet and to request a discreet encounter. Likewise, these men were more likely to seek a masculine sexual partner. In doing so, these men communicated to their desired partners a commonality or shared understanding that presumably would facilitate the coordination of sexual encounters. Moreover, some research suggests that behaviorally bisexual men are more likely than men who only have sex with other men to have bisexual partners (Hightow et al., 2006), and that disclosure of one’s bisexuality may be easier when the partner is also bisexual in identity or behavior (Dodge, Jeffries, & Sandfort, 2008). Further, NGI-seeking men were also more likely to offer to host a sexual encounter. This can be particularly important for non-disclosing men who live with a wife or girlfriend, or who may be unwilling to risk having a male partner seen at their home or in their neighborhood (Schrimshaw, Downing et al., 2013).

The findings of this study further suggest that NGI-seeking men used their desire for specific sexual behaviors to facilitate negotiations with prospective partners. Indeed, NGI-seeking ads were more likely than comparison ads to include language indicating the poster’s preferences for oral-receptive sex, unprotected oral sex, and anal-receptive sex. In other words, the posters were looking for men who wanted to receive oral sex, who were less concerned about using condoms for oral sex or wanted to ejaculate into their partner’s mouth, and who preferred an insertive role during anal intercourse. Conversely, NGI-seeking men were less likely to seek out oral-insertive sex and rimming. These behaviors often require a more participatory role by the prospective partner, which may be off-putting to some men who are less comfortable about engaging in same-sex encounters. The results suggest an assumption by the poster that NGI MSM prefer to engage in oral-insertive and anal-insertive behaviors, as previous studies have shown that behaviorally bisexual men report more insertive than receptive acts (Dodge et al., 2013; Reback & Larkins, 2013; Siegel et al., 2008; Spikes et al., 2009; Stokes et al., 1997; Wheeler et al., 2008; Zellner et al., 2009; Zule et al., 2009). Furthermore, these data offer indirect support for the idea that some non-disclosing NGI MSMW may want to avoid engaging in sexual behaviors that they consider to be characteristic of gay men or that would conflict with their efforts to maintain a heterosexual, bisexual, or other non-gay identity (Siegel et al., 2008).

As one of the first studies to investigate NGI-seeking men, these findings have unique implications for sexual health research targeting non-disclosing, NGI MSM and their same-sex partners. Further investigation is warranted to determine the extent to which NGI-seeking men are successful in finding partners for same-sex encounters who identify as straight or heterosexual, bisexual, ‘down low’ or married, as well as the types of behaviors these men engage in and the level of sexual risk. Although our analysis revealed that very few ads (by NGI seeking or comparison men) specifically sought unprotected anal sex, NGI-seeking men were more likely to indicate a desire for unprotected oral sex and to not mention safer sex or condom use relative to comparison men’s ads. This suggests that some NGI-seeking men may avoid the topic of condom use in their ads (perhaps in an effort to attract NGI partners without raising concerns about HIV/STD infection), which could lead to potential unprotected encounters. The lower likelihood of mentioning condom use or safe sex found here contradicts an earlier study (Cheeseman et al., 2012) of ads posted by MSM seeking married men. Therefore, future research on NGI-seeking men is needed to better understand their risk behaviors prior to the implementation of intervention efforts. Specifically, it is important to investigate the sexual outcomes of these ads for evidence of risk beyond that of just unprotected anal sex (e.g., ejaculation in the mouth, ejaculation into or on the anus, rimming). Information from such investigation would be useful for the development and targeting of HIV/STD prevention and intervention efforts.

In addition to sexual partnering websites, there has been a growing trend among sexual minorities to utilize Smartphone and mobile-based applications to initiate or search for same-sex encounters (Antebi & Schrimshaw, in press; Landovitz et al., 2013). Using GPS technology, these applications can accelerate the speed with which men can locate and meet sexual partners over that of traditional Internet websites that are used to facilitate casual encounters. Additional research is needed to investigate the use of these applications by NGI men who are concerned about remaining discreet or anonymous. More specifically, researchers can examine if and how these men present themselves in this type of virtual sexual marketplace and the strategies they employ to screen potential partners. Since these applications allow users to meet partners who are in close proximity at a particular time, their use may be associated with a sense of sexual urgency rather than desired partner characteristics that can result in more vulnerability to exposure of their sexual orientation and less negotiation of behaviors (safer or otherwise).

Limitations and Conclusions

These findings and their generalizability should be interpreted within the context of this study’s limitations. First, Craigslist was the only website used for data collection. However, unlike other personal dating websites, this site does not require user accounts in order to post ads for sexual encounters. Thus, it does not systematically exclude MSM who would not have the financial resources to pay a membership fee or who are less open about their sexuality (and would not want to provide name and credit card information to a gay website). Likewise, Craigslist has been previously documented to be particularly popular among MSM (Downing, 2011; Grov, 2010) and particularly NGI MSM (Schrimshaw, Downing et al., 2013). Further, ads were collected at randomly assigned one-hour time blocks, and from more than one Craigslist category in an effort to obtain a more representative sample. Second, there was no contact with posters, and therefore no way of verifying their self-presentations or knowing whether men actually met a partner and the types of sexual behaviors that occurred. Contact with the posters of these ads would have required human subjects protections, including a consent process. We suspect that very few posters would agree to participate and provide us with follow-up information regarding their ads. Further, any replies to these ads with a research solicitation would constitute a violation of Craigslist terms and could result in a negative reaction by the poster who likely anticipates a different type of response to their ad. Indeed, as argued earlier, personal ads offer the researcher information on the self-presentation, desired partner characteristics, and sexual behaviors that are unbiased by the participants knowledge that they are being researched.

Regardless of the study limitations, the study has allowed us to document the existence of a subgroup of men who actively seek out same-sex sexual encounters with men who do not identify as gay. Moreover, the findings suggest that men with a preference for NGI men attempt to alert such prospective partners through a combination of self-described characteristics, desired partner attributes, and behavioral preferences all of which serve to attract more discreet and masculine men. Given the attention to the sexual behaviors of NGI MSM, the results of this study support the need for additional research to investigate behavioral outcomes of NGI-seeking men’s personal ads.