Risky Sex- and Drug-Seeking in a Probability Sample of Men-for-Men Online Bulletin Board Postings

Abstract

There has been limited research on men who have sex with men’s postings in online sex-seeking bulletin boards. This study uses a probability sample of 1,438 advertisements ostensibly posted by men-seeking-men in the New York City section of Craigslist.org. Ad’s qualitative text were coded for various sex- and drug-seeking behaviors. The proportion of ads seeking unprotected/uninhibited sex (3.0%) and party-n-play (i.e., sex/drug) encounters (4.2%) was low. In contrast, 12.7% of postings specified safe sex encounters, and 17.5% of postings specified that they were “drug and disease free.” Prevalence of certain behaviors varied by time that ads were posted. Implications for health/community service providers are discussed.

Introduction

In the US, men who have sex with men (MSM) account for almost half of HIV/AIDS cases, and the number of HIV/AIDS diagnoses among MSM from 2001–2006 increased by 8.6% [1]. Urban centers, such as New York City (NYC), bear disproportionate HIV prevalence among MSM and remain important arenas to target research and prevention [2]. Researchers have highlighted how the Internet has altered patterns of dating and sex-seeking among MSM [3, 4]. A 2006 meta-analysis found that 40% of MSM seek sex partners online [5]. Benotsch et al. [4] reported that MSM who meet partners online have more sex partners, are more likely to use methamphetamine and report higher rates of sexual risk behavior. The increase in online sex-seeking among MSM has inspired research inquiries assessing whether the Internet facilitates HIV-risk behaviors in this population [6].

Much of the research on the Internet’s role in MSM’s sexuality has focused on men’s behavior with online and offline partners [6]. A much smaller body of research has focused on the content of the websites where MSM meet partners (i.e., the electronic environment). Certainly, it is useful to understand men’s behavior with online partners as this has implications for targeting HIV prevention/education; however, it is equally important to understand the ways in which HIV risk behaviors are portrayed in virtual environments.

Although limited, content-based research has included analyses of men’s profiles on membership-based “hookup” websites and analysis of such websites in-and-of themselves. Carballo-Dieguez et al. [7] used cybercartography to identify the six most popular free Internet sites MSM used to meet partners for intentional unprotected sex (i.e., bareback) and concluded that the characteristics of these sites allowed men to search for bareback sex partners—even though the websites themselves were not themed around unprotected sex. This study noted similarities in the types of content on bareback-themed websites and more “general” gay websites (e.g., chatrooms, instant message, e-mail account on site).

Grov and Parsons [8] and also Moskowitz and Rolof [9] analyzed Internet profiles of men seeking to either become infected with or to transmit HIV (i.e., bug chasing and gift giving) on a membership-based barebacking website (n = 1,228 and 300, respectively). These studies found some men’s profiles conveyed explicit intent to transmit HIV. Moskowitz and Rolof noted that bug chasers ranked higher on behavioral and psychological measures of sexual addiction than those just identified as barebackers. In a separate study, Grov [10] presented a content analyses of two membership-based websites devoted to barebacking, highlighting how one of the websites encouraged frequent HIV-testing and limiting number of partners, while the other outright negated the seriousness of HIV and STIs. This study suggested these electronic environments create an atmosphere that can influence its members’ attitudes and behavior.

Notably, the aforementioned studies have focused research questions around barebacking and the intentional spread of HIV. As a result, much of this work has been limited to barebacking-themed membership-based web-sites. There has been less research analyzing the content of mainstream MSM hook-up websites. In addition, to date, there has been no research assessing the prevalence of advertisements for risky sexual behavior on MSM online hook-up bulletin boards (e.g., Craigslist.org), which may represent distinct virtual environments from membership-based sites.

For instance, on a membership-based website such as Manhunt.net or Adam4Adam.com, users register for memberships (sometimes with subscription fees) and create profiles for themselves. Users then log onto the website to engage other members. Profile content is often screened by a central administrator such that certain materials are prohibited (e.g., using one’s “real” email address in profile text, references to drug use, or using photos that depict sexual acts). In addition, correspondence between members is often structured such that users must initiate contact within the website, via chat-room or email messages contained on the site.

In contrast, a bulletin board (such as Craigslist.org) is a medium where individuals post single advertisements for sexual encounters. Bulletin boards typically do not require membership registration and may be open to the public. Rather than being moderated by a central administer, bulletin boards are often moderated by their users. Further, given their nature as online classified ad websites, bulletin boards often do not possess correspondence mechanisms (such as chat rooms); instead they facilitate users contacting each other outside of the website. Given the dearth of research on MSM bulletin boards, this study used a probability sample of 1,438 men-seeking-men bulletin board postings. This analysis reports characteristics of bulletin board postings and the prevalence of ads seeking unprotected and “uninhibited” anal sex and drugs use.

Method

Participants and Procedure

This project utilized a large-scale US-based online bulletin board (Craigslist.org). On this site, individuals post ads in a variety of sections (employment, housing, services, personals, etc.). Craigslist is organized by city such that users search for and post ads in their location. Data were taken from the NYC men-seeking-men section (herein “NYC M4M”), which is extremely active; two to four thousand ads are posted every day. The M4M section was free to use, required no membership, and was not moderated by a central administrator (Note, the site had automated filters preventing duplicate ads or including another website’s URL within one’s ad).

Ads were displayed in reverse chronological order (newest on top). Potential sex partners browse ads and responded via email. Ads could remain online for up to a week. Unlike Internet profiles on membership-based sites, where individuals “checked off” responses that appeared in their profile (e.g., age, race, HIV status), the content of ads on Craigslist were almost entirely qualitative and open-ended. Thus, user’s text could range from minimalist (“Looking for ‘now’ in Brooklyn”) to detailed descriptions including type of sex act, condom use, HIV status, and physical description of self or sex partner sought.

Sample Selection

The research team adapted Time–Space sampling [11, 12] in order to generate a probability sample of ads. First, we systematically selected days (every 2nd day) from a 30-day period using a random start date from the month of February 2009. Next, we divided each day into 30-min increments (1 am, 1:30 am, 2 am, 2:30 am, etc.) and used a random-digit generator to select an increment of time. At that randomly-selected time, a member of the research team visited the NYC M4M section and download the first 100 ads. The team downloaded 1,500 ads over the 30-day period. Consistent with the randomization, 18.8% (n = 270) of the downloaded ads were posted between 8 am and 1:59 pm, 21.4% (n = 308) between 2 pm–7:59 pm, 30.6% (n = 440) between 8 pm–1:59 am, and 29.2% (n = 420) between 2 am–7:59 am.

Coding

Prior to the sampling period, the lead researcher spent 4 months conducting virtual ethnographic field work, reviewing the content of thousands of NYC M4M postings. He then created a codebook consisting of 73 items that coders would assess for in each posting. Most items were coded yes/no (presence/absence). For example, “Oral sex: does advertisement indicate looking to give/receive oral sex? 1 = yes, 0 = no.” Other variables included age, race, time of posting, absence/presence of text that would indicate a desire for safe sex, party-n-play behavior (defined as sexual encounters involving methamphetamine, cocaine, or generic use of the words “party” or “pnp” with overt intimations of drug use), presence of a photo placed within the ad, and codes for the content of any photo (e.g., includes face, genitals, depiction of oral sex). During the ethnographic period, we learned that “uninhibited” was often used to describe unprotected anal sex, and thus a code was also created for text specifying a desire for unprotected or “uninhibited” anal sex. All variables were coded 1 = yes, 0 = no.

Five coders were trained to recognize vernacular for various codes. For example “BJ,” “blow job,” “give head,” “go down,” to indicate oral sex; “BB,” “bareback,” “raw,” and “uninhibited” anal sex to indicate high risk/unprotected sex; and “420,” “Mary J.,” “smoke a blunt” to indicate marijuana. Team members were assigned the same ads to code and reviewed coding with the lead researcher until criterion was achieved (94% inter-coder reliability achieved in the third round). Upon achieving criterion, members coded individual ads using SPSS. The full content of each ad (body text, header text, and the content of any photos) was reviewed In the event that a team member was unsure of whether to assign a code, s/he reviewed that ad with another coder and the lead investigator to establish a consensus. Once all ads were coded, coders cross-checked each other, verifying a random 20% of another’s work. Negligible errors (<1%) were identified in this process. Of the 1,500 ads, n = 62 could not be coded due to (a) having been deleted from the website, but the link was still active when downloaded (n = 55), or (b) being miscategorized (e.g., “apartment for rent,” n = 7). Thus the final sample size was 1,438 online postings. Brooklyn College’s Institutional Review Board approved all procedures.

Analytic Plan

For this brief report, the prevalence of various sex-seeking, drug-seeking, relationship infidelity, and the content of photos were assessed. In addition, bivariate analyses identifying differences in variables by the time of day that ads were posted (dichotomized: daytime 8 am–7:59 pm, nighttime 8 pm–7:59 am) are presented. In so doing, chi-square and odds ratios were utilized to evaluate differences among variables. Given the relative infrequency of some behaviors assessed (see Table 1), splitting time into two categories (versus four or six) was used to maximize statistical power, minimizing degrees of freedom.

Table 1.

Comparing the characteristics of men-seeking-men online bulletin board postings by the time of day the ad was posted, N = 1,438, February–March 2009, New York City

	Overall prevalence N = 1,438 n (%)	Daytime posting 8 am–7:59 pm n = 578 n (%)	Nighttime posting 8 pm–7:59 am n = 860 n (%)	χ2 (1)	P	Odds ratio	95% CI
Sexual behaviors
Looking for oral and/or anal sex	921 (64.0)	365 (63.1)	556 (64.7)	0.34	0.56	–	–
Looking for anal sex	592 (41.2)	223 (38.6)	369 (42.9)	2.67	0.10	–	–
Looking for oral sex	569 (39.6)	229 (39.6)	340 (39.5)	0.001	0.97	–	–
Looking for “safe sex” or mentions wanting to use condoms	182 (12.7)	86 (14.9)	96 (11.2)	4.32	0.04	0.72	0.53–0.98
Looking for an “uninhibited” or “bareback” (unprotected anal sex) encounter	43 (3.0)	22 (3.8)	21 (2.4)	2.22	0.14	–	–
Looking for group sex (3+ partners)	104 (7.2)	51 (8.8)	53 (6.2)	3.65	0.06	–	–
Looking for “kinky” sex (fisting, feet, urine, spanking, or BD/SM)	86 (6.0)	42 (7.3)	44 (5.1)	2.84	0.09	–	–
Looking for rimming (i.e., oral sex with anus)	59 (4.1)	34 (5.9)	25 (2.9)	7.78	< 0.01	0.48	0.28–0.81
Is looking to be paid money (i.e., escorting)	45 (3.1)	33 (5.7)	12 (1.4)	21.22	< 0.001	0.23	0.12–0.46
Looking for a relationship, boyfriend	37 (2.6)	16 (2.8)	21 (2.4)	0.15	0.70	–	–
Drug use and drug-seeking
Wants to “party” (with explicit language intimating drug use), PnP, or use cocaine or methamphetamine in sexual context	60 (4.2)	13 (2.2)	47 (5.5)	8.94	< 0.01	2.51	1.35–4.69
Looking for an encounter with marijuana	56 (3.9)	33 (5.7)	23 (2.7)	8.51	< 0.01	0.45	0.26–0.78
Looking for an encounter with poppers	47 (3.3)	27 (4.7)	20 (2.3)	6.02	0.01	0.49	0.27–0.87
Says is, or looking for, “drug and disease free” or “ddf”	252 (17.5)	105 (18.2)	147 (17.1)	0.28	0.60	–	–
Discretion and relationship infidelity
Was seeking partners outside of his own primary relationship	48 (3.3)	26 (4.5)	22 (2.6)	4.03	0.04	0.56	0.31–0.99
Looking for a partner who is married or in committed relationship	55 (3.8)	28 (4.8)	27 (3.1)	2.73	0.10	–	–
Is on the “DL” or “Down Low”a	30 (2.1)	12 (2.1)	18 (2.1)	< 0.001	0.98	–	–
Mentions discretion or “discreet”	227 (15.8)	109 (18.9)	118 (13.7)	6.86	< 0.01	0.68	0.51–0.91
Has a picture in posting	630 (43.8)	257 (44.5)	373 (43.4)	0.17	0.68	–	–
Includes face (valid n = 630)	117 (18.6)	49 (19.1)	68 (18.2)	0.08	0.78	–	–
Includes torso (valid n = 630)	271 (43.0)	110 (42.8)	161 (43.2)	0.01	0.93	–	–
Includes buttocks (valid n = 630)	168 (26.7)	70 (27.2)	98 (26.3)	0.07	0.79	–	–
Includes genitals (valid n = 630)	274 (43.5)	108 (42.0)	166 (44.5)	0.38	0.54	–	–
Includes oral sex (valid n = 630)	33 (5.2)	13 (5.1)	20 (5.4)	0.03	0.87	–	–
Includes anal/vaginal sex (i.e., penetration) (valid n = 630)	25 (4.0)	8 (3.1)	17 (4.6)	0.83	0.36	–	–
Includes “other” content that could not be coded (photo of mountain, smile face, sex toy, etc.) (valid n = 630)	70 (11.1)	31 (12.2)	39 (10.4)	0.44	0.51	–	–

^aTerm that describes a man having female partners who were unaware of his sexual behavior with other men

Results

The stated mean age of individuals posting ads was 32.5 (SD = 8.54, Range 18–67, valid n = 1,316). Most posters did not indicate a racial identification (64.4%); however, among those who did (n = 512), 55.9% were White (n = 286), 16.0% were Latino (n = 82), 13.9% were Asian/Pacific Islander (n = 71), 9.8% were African American/Black (n = 50), and the remainder were of mixed or other races/ethnicities (n = 23). Of the 130 (9.0%) men who specified their HIV status, 95% (n = 124) said they were HIV negative and 4.6% (n = 6) said HIV positive. Most ads (64.3%) were posted in Manhattan, followed by Brooklyn (11.9%), Queens (8.2%), Long Island (5.8%), upstate New York (2.5%), Connecticut (2.4%), the Bronx (2.0%), New Jersey (1.9%), and Staten Island (1.0%). Given that all ads were taken from a men-seeking-men section, most posters did not specify a sexual orientation (91.0%). If sexual orientation was mentioned (n = 130), most indicated they were bisexual (70.8%, n = 92), followed by heterosexual (15.4%, n = 20), and gay (13.8%, n = 17).

Sexual Behaviors

Table 1 reports the prevalence of select characteristics of online ads. Although some ads were ambiguous about the poster’s desires/intentions, 64.0% of ads indicated an explicit interest in either anal and/or oral sex: 41.2% anal sex, 39.6% oral sex. Among all ads, the prevalence of seeking unprotected/bareback sex or “uninhibited” sex was low (3.0%), whereas 12.7% of ads specified they were looking for “safe sex” (i.e., indicating “safe sex” or condom use). Seven percent of ads were for group sex (i.e., 3+ partners, or specifying a “group” encounter), 6.0% sought specialized sexual behaviors (i.e., fisting, foot fetish, urine fetish, spanking, bondage/domination, sadism/masochism), and 4.1% to engage in rimming (oral-anal sex).

The prevalence of some sexual behaviors differed by the time ads were posted. Compared with daytime ads, ads posted at night were at a 28% (1–0.72) reduced odds of specifying a “safe sex” encounter (P < 0.05). In addition, daytime ads were more likely to specify a desire for rimming (5.9 vs. 2.9%, P < 0.01) and being paid for sex (i.e., escorting; 5.7 vs. 1.4%, P < 0.001). There were no time-of-day differences in the prevalence of other sexual behaviors.

Drug Use and Drug-Seeking

In total, 17.5% of posters indicated that they were “drug and disease free” (commonly abbreviated “ddf”) in their ads. Prevalence of indicating “drug and disease free” did not significantly vary by the time of day that ads were placed. Compared with daytime ads (2.2%), nighttime ads (5.5%) were 2.51 times the odds of indicating a desire for a party-n-play encounter (operationalized as involving either crystal methamphetamine and/or cocaine, or specifying “pnp” or “party” with explicit language intimating drug use, P < 0.01). In contrast, daytime ads were significantly more likely than nighttime ads to indicate a desire for marijuana (5.7 vs. 2.7%) and poppers/nitrate inhalants (4.7 vs. 2.3%, Ps < 0.01).

Discretion and Relationship Infidelity

Mentions of being on the “down low” or “DL” (2.1%) (i.e., vernacular used to describe having female partners who were unaware of one’s sexual behavior with other men) and seeking a partner who is married or otherwise in a committed relationship (3.8%) were infrequent, and did not significantly vary by the time of day. However, daytime ads (4.5%) were significantly more likely than nighttime ads (2.6%) to indicate that the poster was seeking partners outside of his own primary relationship (P = 0.04). Ads indicating that the poster was seeking partners outside of his own primary relationship (compared with ads that were not) were 4.89 times the odds of indicating they were also looking for a “discreet” encounter (45.8 vs. 14.7%, 95% CI = 2.72–8.80), and 11.92 times the odds of indicating that they were looking for someone who was married or otherwise in a committed relationship (27.1 vs. 3.0%, 95% CI = 5.88–24.17).

Posting a Picture

In total, 43.8% of ads included a picture. Among those with pictures (n = 630), 62.4% included one photo, 22.9% included two, 6.2% included three, and 8.6% included four. Most often these pictures included a person’s torso (43.0%), genitals (43.5%), or buttocks (26.7%). Only 18.6% of ads with photos included a face, and fewer included depictions of oral sex (5.2%) or penetrative sex (4.0%). Whether an ad contained a photo and the content of photos did not significantly vary by the time of day ads were posted.

Discussion

In this study, the proportion of advertisements seeking unprotected/uninhibited sex and party-n-play encounters involving methamphetamine, cocaine, or non-specified “party” drugs—behaviors that potentiate the spread of HIV—were low. Although the prevalence of ads seeking to engage in unprotected or “uninhibited” anal sex did not differ by time of day (daytime 8 am–7:59 pm versus nighttime 8 pm–7:59 am), indicating a desire for “safe sex” was significantly less frequent at night. In addition, seeking “party-n-play” encounters were highest among nighttime ads. These findings highlight that Internet-based outreach tailored for men who engage in sex under the influence of drugs might be best targeted at night.

Even though the proportions of advertisements for unprotected/uninhibited anal sex and party-n-play encounters were relatively low, there were thousands of ads posted in the NYC M4M section of Craigslist each day. Given this study’s findings, this would equate to about one-hundred ads per day seeking uninhibited/unprotected anal sex and/ or party-n-play encounters. With the website’s text-based search feature, users could easily identify advertisements for “uninhibited,” “bareback,” or “raw” encounters (for example). Thus, even though not pervasive on this online bulletin board, the website was an environment where users could locate risky sex should they wished to do so. Carballo-Dieguez et al. [7] noted how such features, embedded within mainstream gay websites, made searching for bareback sex partners possible. Nevertheless, and as highlighted by Chiasson et al. [6], this may also be a medium through which community/health service providers could initially locate individuals engaging in HIV risk behaviors.

In contrast to ads for “risky” encounters, approximately one out of every eight postings specified they were looking for a safe sex encounter. In addition, approximately one in six postings specified that they were “drug and disease free.” These findings highlight how postings for safer sex and being drug free were more common than those seeking risky encounters.

This study is subject to several limitations. First, ads were taken from a single city’s M4M section. NYC has one of the highest HIV prevalence rates among MSM in the US [2] and thus a continued need for focused research on factors that influence HIV transmission in the geographic population. Nevertheless, it is uncertain how results might vary for other urban, suburban, or rural areas. Second, ads were taken from a single website. Although Craigslist.org is possibly the most active M4M online bulletin board in the United States, other smaller-scale sites also host bulletin boards and may cater to less mainstream (i.e., more specialized) populations. Third, prevalence data reported in this study relied on analysis and interpretation of qualitative text and self-reported information (e.g., age, physical characteristics) which may not have been entirely factual. Even though qualitative data can provide rich detail regarding the sexual behaviors users may be seeking, some ads lacked adequate text from which to draw definitive conclusions. Erring on the side of caution, ambiguous text was not coded. Given the stigma associated with engaging in high-risk sexual behavior and the illegality of drug use, some men may have been more allusive in their language. Thus, the prevalence of some riskier or more illicit behavior reported in this study may be underestimated. Finally, it is uncertain if users’ postings ultimately resulted in the sexual encounters they were seeking. Similar limitations are common among studies that analyze Internet personals/profiles.

Despite these limitations, this is one of the first studies to have used a probably sample of MSM’s qualitative online bulletin board postings; reporting on various sex-and drug-seeking behaviors. Although Craigslist may not be the most common venue MSM use to meet online sex partners (e.g., Manhunt.net has over 3 million active members), it is one of the largest M4M bulletin boards in the US. Because it is free, un-moderated, and requires no membership, this bulletin board might be an attractive option for MSM seeking immediate sexual encounters on a casual basis; a population that could be distinct from membership-based “hook-up” websites that have been reported on in previous studies (see Liau et al. [5] for a review). In addition, on membership/profile-based sites, men “check off” interests and personal characteristics that appear in their profiles, but this feature is not provided on Craigslist. In this study, 91% of ads made no mention of men’s HIV status, compared with (for example) only 11.5% of the membership-based profile that Grov and Parsons [8] analyzed. Such structural characteristics inherent to the electronic environment could impact how men negotiate HIV-associated risk behavior [10].

In recent years, community and health service providers have recognized the need to increase their online presence in an effort to reach MSM—engaging men with Internet-based outreach and counseling [6]. In this study, we found some behaviors did not differ between daytime and nighttime ads (e.g., seeking oral or anal sex). In contrast, postings for “party-n-play” encounters were highest at night, and some behaviors, such as relationship infidelity and seeking “safe sex,” were most frequent during the day. These findings may be useful for community and health service providers as they highlight how the services provided and target populations sought by online health/community groups may need to be tailored according to the time of day. Specifically, efforts to reach men who engage in “party-n-play” behavior might be best targeted at night. Thus it may be necessary for providers to offer resources in times that are outside of “typical” business hours.