Arch Sex Behav (2017) 46:1763–1776

DOI 10.1007/s10508-016-0837-9

ORIGINAL PAPER

Sexually Explicit Media Use by Sexual Identity: A Comparative

Analysis of Gay, Bisexual, and Heterosexual Men in the United States
Martin J. Downing Jr.1 • Eric W. Schrimshaw2 • Roberta Scheinmann1 •
Nadav Antebi-Gruszka2 • Sabina Hirshfield1

Received: 23 March 2016 / Revised: 5 August 2016 / Accepted: 9 August 2016 / Published online: 5 October 2016
 Springer Science+Business Media New York 2016

Abstract Advances in production and distribution of sexually behavior.FindingssuggesttheimportanceofassessingSEMuse

explicit media (SEM) online have resulted in widespread use across media types and contexts and have implications for
among men. Limited research has compared contexts of use and research to address the potential influence of SEM on sexual
behaviors viewed in Internet SEM by sexual identity. The current behavior (e.g., investigate associations between viewing
studyexamineddifferencesinrecentSEMuse(past 6 months)by condomless vaginal sex and engaging in high-risk encounters
sexual identity among an ethnically diverse sample of 821 men with female partners).
who completed an online survey in 2015. Both gay and bisexual
men reported significantly more frequent use of Internet SEM Keywords Sexually explicit media
Pornography

compared to heterosexual men. Although most participants Internet
Sexual identity
Sexual orientation
reported viewing SEM at home (on a computer, tablet, or
smartphone), significantly more gay men reported SEM use
at a sex party or commercial sex venue than either heterosexual Introduction
or bisexual men. Sexual identity predicted viewing of high-risk
and protective behaviors in separate logistic regression models. Advances in the production and distribution of sexually explicit
Specifically, compared to heterosexual men, gay and bisexual media (SEM) have resulted in widespread availability and con-
men had increased odds of viewing condomless anal sex (gay sumption on the Internet (Escoffier, 2009; Rosser et al., 2012;
OR5.20,95 %CI3.35–8.09;bisexualOR3.99,95 %CI2.24–7.10) Weinberg,Williams,Kleiner,& Irizarry,2010).Consistent with
and anal sex with a condom (gay OR 3.93, 95 % CI 2.64–5.83; this greater availability of SEM online, data from the General
bisexual OR 4.59, 95 % CI 2.78–7.57). Compared to gay men, SocialSurvey(GSS)havedocumentedasteadyincreaseinSEM
heterosexual and bisexual men had increased odds of viewing consumption among US men since the 1970s (Wright, 2013).
condomless vaginal sex (heterosexual OR 27.08, 95 % CI According to recent figures from TrafficJunky Media Kit
15.25–48.07; bisexual OR 5.59, 95 % CI 3.81–8.21) and vaginal (2015), a leading advertising network for adult content Web
sex with a condom (heterosexual OR 7.90, 95 % CI 5.19–12.03; sites, their top three SEM Web sites receive a combined average
bisexual OR 4.97, 95 % CI 3.32–7.44). There was also evidence of over 92 million daily visitors; visitors are overwhelmingly
of identity discrepant SEM viewing as 20.7 % of heterosexual- male (75 %), heterosexual (75 %), and age 18–44 (74 %). Not
identified men reported viewing male same-sex behavior and surprisingly, men are more likely than women to access SEM
55.0 % of gay-identified men reported viewing heterosexual (Albright,2008)andtodosomorefrequently(Hald& Štulhofer,
2016; Morgan, 2011; Paul, 2009; Weinberg et al., 2010). Nev-
ertheless, much of what we know about SEM use among men
& Martin J. Downing Jr. pertains to their frequency of use, though emerging research
[email protected]
attention is being paid to the types of media they consume (e.g.,
1
Public Health Solutions, 40 Worth Street, 5th Floor, New York, heterosexual, bisexual, gay, or lesbian), behavioral content and
NY 10013, USA preferences, and contexts of use.
2
Mailman School of Public Health, Columbia University, In a review of Internet-based SEM research conducted over a
New York, NY, USA 10-year period, Short, Black, Smith, Wetterneck, and Wells

123
1764 Arch Sex Behav (2017) 46:1763–1776

(2012) noted that most of the studies published during this It is critical to understand the content of SEM because a
timeframe contained predominantly heterosexual samples significant and growing body of research has suggested that
or did not assess participants’ sexual identity. To our knowl- viewing SEM, and specific behaviors in SEM, may influence
edge, there are no U.S. studies that have compared the behav- real-life behavior among heterosexual and non-heterosexual
ioral content of Internet SEM viewed by sexual identity or viewers. Several studies have revealed potential benefits of view-
sexual orientation. Do men tend to view SEM that corresponds ingSEM.Specifically,someresearchhasshownthatviewingSEM
to their sexual identity or are they more diverse in what they isassociatedwithhavingamorepositiveattitudeaboutsex(Hald&
watch? This is an important distinction to make as researchers Malamuth, 2008; Hald, Smolenski, & Rosser, 2013). Viewing of
consider the role of SEM in sexual behavior, particularly SEM also increases the appeal of certain behaviors (e.g., anal sex),
behaviors that increase one’s risk for HIV and other sexually regardless of gender or sexual orientation (Weinberg et al., 2010).
transmitted infections (STIs), and effective ways to reach SEM Amongheterosexuals,investigatorshavefoundthat they useSEM
audiences with risk reduction messaging (e.g., Rosser et al., to increase arousal before or during sex, to learn new sexual posi-
2012). Peter and Valkenburg (2012) assessed the frequency of tions and activities, to relieve stress and sexual frustration, and to
viewing Internet SEM featuring penetrative sex, group sex, sex fantasize about having sex with the performers (Albright, 2008;
with only men, and sex with only women among exclusively Boies, 2002; Hare, Gahagan, Jackson, & Steenbeek, 2015; Paul &
and nonexclusively heterosexual men in the Netherlands. Shim, 2008; Sun et al., 2016; Traeen & Daneback, 2013). Per-
However, it is not clear from their analyses whether the authors ceived positive effects of SEM use by men who report same-sex
examined differences in content viewed by sexual identity. encounters include an interest in trying new sexual behaviors or
Hald and Štulhofer (2016) recently conducted exploratory fac- positions, enjoyment of sex, and understanding of one’s sexual
tor analyses of pornography types by sexual orientation in a orientation (Hald etal.,2013; Nelson,Leickly, Yang, Pereira, &
Croatian sample, noting similarities among heterosexual and Simoni, 2014a).
non-heterosexual men in viewing group sex behaviors (e.g., Despite the potential benefits of SEM use, other research
bukkake, gang bang scenes featuring one woman and at least has raised concerns that viewing specific behavioral content
three men). may contribute to negative sexual health outcomes. The prolif-
Given the importance of understanding cultural contrib- eration of SEM online has coincided not only with a rise in use,
utors to men’s sexual aggression against women, much of the but also with reported decreases in safer sex behaviors among
extant research on SEM content has focused on heterosexual some viewers (Peter & Valkenburg, 2011a; Rosser et al., 2012;
media and the portrayal of violence toward women (Salmon Wright, Tokunaga, & Kraus, 2016). Moreover, technological
& Diamond, 2012; Sun, Bridges, Johnason, & Ezzell, 2016; innovations have since led to an increase in production and
Wright & Tokunaga, 2016). Additional areas deserving attention distribution of amateur sexual content (Cronin & Davenport,
are non-heterosexual media, comparisons of sexual content 2001; Downing, Schrimshaw, Antebi, & Siegel, 2014b; Green,
across media types, and specific sexual risk behaviors per- 2004). Amateur content, which is not subject to SEM industry
formed by SEM actors (i.e., condomless vaginal and anal sex), HIV/STI testing or condom use policies (Griffith et al., 2012),
especially in light of recent cross-sectional evidence that view- has altered‘‘the narrative and sexual expectations presupposed
ing condomless sexual encounters in Internet SEM is predictive by most porn movies’’ (Escoffier, 2009, p. 347). While some
of engaging in high-risk sex (as discussed below). What remains have speculated that increased competition from a booming
unclear is whether men are differentially exposed to high-risk Internet-based amateur industry prompted gay male SEM stu-
SEM content (i.e., attributed to sexual identity, context of SEM dios to abandon their condom use policies (Hurley, 2009), pro-
use,orviewingpreferences),whichhasimplicationsfortargeting ducersofcommercialbareback(i.e.,condomlessanalsex)videos
SEM-based HIV/STI prevention strategies. Nevertheless, some argue that advances in HIV treatment and prevention (e.g.,
comparisons of sexual behaviors in gay and heterosexual DVD- antiretroviral therapy, pre-exposure prophylaxis) have signif-
based SEM have found that heterosexual videos were less likely icantlyloweredtheriskofHIVtransmission,makingitpossible
todepictanalsex(Salmon&Diamond,2012),condomuseduring to meet consumer demands (Kinser, 2014; Nichols, 2011).
anal sex (Grudzen et al., 2009), and external ejaculation (Salmon Indeed, there has been a substantial rise in sales of SEM fea-
& Diamond, 2012). By the early 1990s, producers of gay male turing condomless sex (Escoffier, 2009). Furthermore, a recent
SEM committed to showing condoms during anal sex scenes for content analysis of 302 gay male SEM videos on the Internet
the occupational safety of performers and to model safer sex for found depictions of condomless anal sex in 34 % of the sample
viewers (Bishop, 2015). Although the heterosexual SEM industry (Downing et al., 2014b); this was considerably higher than that
never fully adopted this practice, it did implement a testing policy previously observed in DVD-based SEM (18 %) by Grudzen
aimedatpreventingacquisitionandtransmissionofHIVandother et al. (2009).
STIs (Escoffier, 2009; Goldstein, Steinberg, Aynalem, & Kerndt, The expanded access to SEM online has perhaps had the
2011). greatest impact on gay, bisexual, and other men who have sex

123
Arch Sex Behav (2017) 46:1763–1776 1765

with men (GBMSM). Limited research suggests that a greater in heterosexual and bisexual media. We found a single study that
proportion of GBMSM view Internet SEM than heterosexual assessed attitudes toward condom use in SEM among male col-
men (Traeen, Nilsen, & Stigum, 2006). Several studies with lege students (Kraus & Rosenberg,2016).The authors reported
online samples of GBMSM found a high prevalence of Internet thatgaymenheldmoresupportiveattitudestowardcondomuse
SEM viewership in the past 3 months ([95 %) (Rosser et al., in SEM than heterosexual men. Yet, there were several limi-
2013; Stein, Silvera, Hagerty, & Marmor, 2012). GBMSM are tations of that study including: the assessment of condom use
also more frequent consumers of Internet SEM than hetero- attitudes without reference to specific sexual behaviors such as
sexual men (Duggan & McCreary, 2004; Peter & Valkenburg, vaginal or anal sex; no assessment of the sexual behaviors (with
2011b). Although SEM may have positive effects on the sexual and withoutcondoms)thatparticipants viewed inSEM; and the
lives of GBMSM (Hald et al., 2013), widespread use may not be inclusion ofsmall subsamplesofgay(n = 32)andbisexualmen
without adverse consequences. Videos that depict condomless (n = 17)comparedtoheterosexualmen(n = 155).Thus,research
anal sex are potentially problematic, as they may influence the is needed to address not only whether men are differentially
intrapersonal sexual scripts of viewers (Štulhofer, Buško, & exposed to high-risk SEM content (as a potential factor of how
Landripet, 2010) and lead some men to engage in similar activ they sexually identify), but also to examine preferences for con-
ities(Jonas,Hawk,Vastenburg,&deGroot,2014;Wilkersonetal., dom use during vaginal and anal sex scenes among heterosexual
2012). Multiple studies to date have documented that GBMSM and bisexual viewers. Answers to these questions are needed
who viewed a greater proportion of SEM depicting bareback sex before researchers can begin to address the potential behavioral
had increased odds of reporting recent anal sex without a condom impact of condomless sex in heterosexual and bisexual SEM.
(Nelson et al., 2014b; Rosser et al., 2013; Schrimshaw, Antebi- Researchers have also called for more attention to venues
Gruszka, & Downing, 2016a; Stein et al., 2012). Furthermore, ofSEMuse(Rosseretal.,2012).Whileitisreasonabletobelieve
consumption of SEM featuring condomless anal sex may be thatmostmenaccessSEMathome,scantattentionhasbeenpaid
most problematic for men who use it as a source of sexual to otherenvironmentswhere SEM maybe availableoraccessed.
information and therefore may normalize such behaviors (Ku- Beyond commercial sex venues where sexually explicit videos
bicek, Beyer, Weiss, Iverson, & Kipke, 2010; Nelson et al., may be streaming for patrons (Holmes, O’Byrne, & Gastaldo,
2014a). 2007; Rosser et al., 2012), there is reason to suspect that men are
Within the literature on SEM use among GBMSM, however, increasingly accessing Internet-based SEM at work (Albright,
researchershaveprimarilyexaminedoutcomesofinterestwithin 2008; Kuchment & Springen, 2008; Perrin et al., 2008). More-
the full sample rather than examining potential differences by over, widespread access to the Internet via smartphones and
sexual identity (i.e., gay vs. bisexual). Most of the published tablets has likely contributed to greater use of SEM at home and
studies that found associations between viewing condomless in other venues. What remains unclear is whether or not viewing
anal sex in SEM and engaging in recent anal sex without a con- context differs by sexual identity.
dom included substantial proportions of gay-identified men The current paper reports on findings from an online sur-
([80 %; Nelson et al., 2014b; Schrimshaw et al., 2016a; Stein vey of self-identified heterosexual, gay, and bisexual men
et al., 2012), which likely accounts for the limited attention to who viewed Internet-based SEM in the past 6 months. This
potential sexual identity differences. Stein et al. (2012) repor- study aims to extend prior research by comparing frequency
ted that participants who did not provide data on their SEM of SEM use, viewing contexts (where and how men access
viewinghistoryweremorelikelytoidentifyasbisexual.Bisexual SEM; use of substances while viewing SEM; masturbation
men remain underrepresented, and perhaps unaccounted for, in during SEM use), content of SEM viewed (vaginal and anal sex,
studies of SEM use and subsequent behavioral impact. Similar with and without condoms), and viewer preferences for condom
concerns have been raised regarding the lack of differentiation use during vaginal and anal sex scenes by sexual identity among
between gay and bisexual men in other sexual health research anethnicallydiversesample.Bytakingamoreinclusiveapproach
(Schnarrs et al., 2012) as well as the relative absence of bisexual to examining SEM use, researchers will be better prepared to
men in studies of disclosure and openness about one’s sexual address concerns about the influence of SEM on sexual behavior
orientationandidentity(Schrimshaw,Downing,&Cohn,2016b) across media types (e.g., SEM that targets heterosexual, bisexual,
and those that examine associations between sexual orientation and/or gay audiences).
and mental health (Dodge et al., 2012; Schrimshaw, Siegel,
Downing, & Parsons, 2013b). There is clearly a need for parity
in sexual health research with heterosexual and non-hetero- Method
sexual populations, particularly studies that investigate viewing
ofdiverse SEM content and the potential influence ofviewingon Participants
sexual behavior, including male and female partners.
Despite the increased attention to bareback sex in gay male Interested individuals were asked to complete an online survey
SEM,thereisalackofequivalentresearchontheuseofcondoms about their experiences with and preferences regarding SEM.

123
1766 Arch Sex Behav (2017) 46:1763–1776

Eligible participants had to: (a) be 18 years of age or older; (b) be providing disqualifying or duplicate data. The overall analytic
able to read and respond in English; and (c) reside within the sample included 1088 surveys. Of those, 821 male-identified
U.S. or its territories. We excluded anyone who selected‘‘Prefer participants indicated that they viewed Internet SEM in the past
not to answer’’ when asked about their recent sexual partners. 6 months and reported their sexual identity as gay/homosexual
Although the study was open to individuals of all gender types, (65.0 %), bisexual (18.6 %), or straight/heterosexual (16.3 %).
the current report is based on data from 821 participants who More than half (65.4 %) of the 821 men were recruited from
identified their current gender as male, reported viewing Internet sexual networking Web sites and GPS-based smartphone appli-
SEM in the past 6 months, and self-identified as straight/hetero- cations, 19.6 % through research-oriented Web sitesora research
sexual, gay/homosexual, or bisexual. We focus on men as they participantregistry,8.2 %fromAmazonMechanicalTurk,4.1 %
are the primary users of SEM. Table 1 describes participant from Craigslist, 1.6 % from social networking Web sites, and
characteristics. 1.1 % through study palm cards.
Recruitment activities occurred from January 23, 2015–
November 1, 2015. To reach a diverse audience of SEM view- Procedure
ers, consistent with our eligibility criteria, study advertise-
ments were posted online to social networking Web sites (e.g., The institutional review board at Public Health Solutions
Twitter), a sexual networking Web site (BGCLive.com), a GPS- approved all study procedures. A waiver of documentation of
based smartphone application for sexual partnering (Scruff), written consent was obtained, given the Internet-based research
Craigslist (Volunteers), research-oriented Web sites (e.g., Social approach. Potential participants accessed the study landing page
Psychology Network), and Amazon Mechanical Turk. Palm by clicking on an online study banner advertisement or study
cards featuring the study logo and detailed description were invitation link. Palm cards directed anyone interested in par-
distributed via LGBT (Lesbian, Gay, Bisexual, and Transgen- ticipating to scan a QR code or visit the study URL. Participants
der) organizations and professional meetings. Study invitations provided informed consent online by reading the consent form
were also sent to members of a national research participant reg-
istry affiliated with the lead author’s institution. The study Footnote 1 continued
employed both general (e.g., online survey of viewer attitudes surveys (M = 34.48, SD = 11.24), F(1, 1467) = 32.31, p\.001. A
greater proportion of men who identified their race as White (71.8 %)
about sexual behaviors shown in sexually explicit videos on the completed the survey compared to men who identified as Black (54.2 %,
Internet;‘‘…study seeks to better understand attitudes of indi- p\.001) or Asian, Pacific Islander, Native American or Alaska Native,
viduals who watch pornographic videos on the Internet’’;‘‘Re- Native Hawaiian or Other Pacific Islander, or Other (44.8 %, p\.001).
searchers seek porn watchers’’) and specific recruitment lan- Further, significantly fewer men reporting up to a high school degree or
GED completed the survey (46.1 %) compared to men with some col-
guage (e.g., ‘‘…study of viewer attitudes about PrEP in the lege, Associate’s degree or Technical degree (59.3 %, p\.001), a 4-year
Adult Film Industry’’; ‘‘Do you like to watch bareback sex?’’ college degree (67.8 %, p\.001), or a professional or graduate degree
‘‘Are you concerned about HIV testing standards in the pornog- (68.6 %, p\.001). Significantly fewer men reporting an annual income
raphy industry?’’). Study advertisements posted to social net- of less than $10,000 completed the survey (50.0 %) compared to those
who earned $40,000–$79,999 (63.1 %, p = .002), $80,000–$119,999
working Web sites incorporated relevant hashtags (#porn, (73.2 %, p\.001), or $120,000 or more (68.7 %, p = .003). Similarly,
#PrEP) to engage a broader audience. Online recruitment noti- significantly fewer men who preferred not to answer the item about
ces provided or embedded a link to access the survey. Where annual income completed the survey (51.7 %) than men who reported
possible, recruitment notices indicated that participants should be earning $80,000–$119,999 per year (73.2 %, p = .001). A significantly
greater proportion of men who reported a zip code corresponding to a US
18 years of age or older, that the survey would take approximately state in the West region completed the survey (73.3 %) compared to men
15–20 min to complete, and that there would be an opportunity in the Northeast (58.8 %, p = .001), Southeast (55.2 %, p\.001), and
to enter a random drawing for a $20 Amazon.com electronic Midwest (60.4 %, p = .001) regions. There were no differences between
gift card upon completing the survey. complete and partial cases in relationship status, HIV status, current
residence, or use of Internet SEM in the past 6 months (yes, no).
There were a total of 2529 visits to the study landing page.
There were significantly more completed surveys among men who
Of those, 186 (7.4 %) broke off immediately, and 32 (1.3 %)
viewed SEM at home on a computer (67.6 vs. 40.3 % of men who did not
chose not to consent to participate. Among the 2311 individ- view SEM at home on a computer, p\.001), at work on a computer (73.8
uals who consented to participate, 1070 (46.3 %) provided only vs. 59.9 % of men who did not view SEM at work on a computer, p\.05),
partial data.1 We excluded 153 (12.3 %) completed surveys for or while attending a commercial sex venue (72.4 vs. 58.5 % of men who
did not view SEM while attending a commercial sex venue, p\.001).
Further, there were significantly more completed surveys among men
1
We examined differences in participant characteristics and Internet who viewed vaginal sex with a condom (87.9 vs. 54.4 % of men who did
SEM use between complete and partial cases. Several differences were not view this behavior), vaginal sex without a condom (89.1 vs. 49.6 % of
noted after applying a Bonferroni correction. A greater proportion of men who did not view this behavior), anal sex with a condom (87.9 vs.
bisexual-identified men had partial surveys (45.6 %) compared to 39.3 % of men who did not view this behavior), and anal sex without a
heterosexual-identified men (32.4 %) (p = .003). No other sexual condom (89.3 vs. 26.2 % of men who did not view this behavior), p values
identity differences were noted. Men with complete surveys were \.001. Lastly, we observed no differences between complete and partial
significantly older (M = 38.08, SD = 12.16) than those with partial cases in the types of Internet SEM viewed in the past 6 months.

123
Arch Sex Behav (2017) 46:1763–1776 1767

Table 1 Comparisons of participant characteristics and recruitment source by sexual identity

Heterosexual Gay Bisexual v2 uc Post hoc
(a) (b) (c)
n = 134 n = 534 n = 153

Race and ethnicity (N = 804) 106.03*** 0.26

White or Caucasian 98 (73.1) 216 (41.5) 32 (21.3) a[b, c; b[c
Black or African-American 19 (14.2) 240 (46.2) 109 (72.7) a\b, c; b\c
Hispanic or Latino 11 (8.2) 47 (9.0) 3 (2.0) b[c
Asian, Pacific Islander, Native American or Alaska Native, 6 (4.5) 17 (3.3) 6 (4.0)
Native Hawaiian, or Other
Relationship status (N = 818) 68.99*** 0.20
Single and not currently in a relationship 55 (41.0) 388 (72.9) 112 (73.7) a\b, c
In a steady relationship 25 (18.7) 79 (14.8) 18 (11.8)
Married or domestic partnership 54 (40.3) 65 (12.2) 22 (14.5) a[b, c
Education (N = 818) 8.38 0.07
BHigh school graduate/GED 15 (11.3) 71 (13.3) 21 (13.7)
Some college, Associate’s degree/Technical degree 50 (37.6) 211 (39.7) 72 (47.1)
College graduate (4 years) 42 (31.6) 144 (27.1) 43 (28.1)
Professional or graduate degree 26 (19.5) 106 (19.9) 17 (11.1)
Annual income (N = 776) 4.19 0.05
\$10,000 15 (11.7) 70 (13.9) 20 (13.8)
$10,000–$39,999 46 (35.9) 192 (38.2) 63 (43.4)
$40,000–$79,999 36 (28.1) 144 (28.6) 36 (24.8)
$80,000–$119,999 19 (14.8) 65 (12.9) 16 (11.0)
$120,000 or more 12 (9.4) 32 (6.4) 10 (6.9)
Current living situation (N = 806) 12.81* 0.09
A house 84 (62.7) 239 (45.6) 74 (50.0) a[b
An apartment 42 (31.3) 233 (44.5) 59 (39.9) a\b
Other 8 (6.0) 52 (9.9) 15 (10.1)
HIV status (N = 769) 223.75*** 0.38
Negative 74 (56.1) 272 (54.8) 113 (80.1) a, b\c
Positivea 0 188 (37.9) 16 (11.3) b[c
Indeterminate or never got test results 0 12 (2.4) 2 (1.4) NA
Never tested 58 (43.9) 24 (4.8) 10 (7.1) a[b, c
Geographic region (N = 813) 16.94* 0.10
Northeast 23 (17.4) 69 (13.0) 28 (18.7)
Southeast 37 (28.0) 171 (32.2) 55 (36.7)
Midwest 33 (25.0) 106 (20.0) 34 (22.7)
Southwest 16 (12.1) 51 (9.6) 14 (9.3)
West 23 (17.4) 134 (25.2) 19 (12.7) b[c
Recruitment source (N = 821) 434.13*** 0.51
Sexual networking 13 (9.7) 396 (74.2) 128 (83.7) a\b, c
Social networking 4 (3.0) 9 (1.7) 0 NA
Craigslist 23 (17.2) 6 (1.1) 5 (3.3) a[b, c
Research (i.e., Web sites, participant registry) 33 (24.6) 112 (21.0) 16 (10.5) a, b[c
Amazon Mechanical Turk 61 (45.5) 3 (0.6) 3 (2.0) NA
Palm cards 0 8 (1.5) 1 (0.7) NA
n (%) presented. Other living situation includes a room in someone else’s house or apartment, drug treatment, detox, shelter or drop in center for
homeless people, or some other place
*** p\.001; * p\.05. Bonferroni corrections applied to post hoc comparisons (p\.017). NA (not analyzed due to low cell counts)
a
Bivariate analysis restricted to gay and bisexual men. v2. uc (Cramer’s V)

123
1768 Arch Sex Behav (2017) 46:1763–1776

on the study landing page and clicking their agreement to par- andonemanengaginginsexualactswitheachother).Participants
ticipate. were asked to report how often in the past 6 months they watched
The survey included a ReCaptcha function to validate human pornographic videos on the Internet (on a computer, tablet, or
responses (i.e.,‘‘Please enter the code shown below in order to smartphone) (adapted from Nelson et al., 2014b). Response
proceed.’’) and offer protection from bots. This feature followed options included less than once a month, once a month, two to
theconsent page. Internet provider(IP)addresseswerecollected three times a month, at least once a week, every day, two to three
for each survey entry to further reduce the likelihood of partic- times a day, once an hour, and more than once an hour. Partic-
ipant fraud. Survey entries with matching IP addresses (match- ipants were also asked to report the number of hours spent
ing on all four quadrants) were considered to be duplicate cases. viewing pornographic videos on the Internet in a typical week
For duplicate cases: (1) if both entries had complete data, the (Downing, Antebi, & Schrimshaw, 2014a). A single follow-up
initial entry was kept for analysis; (2) if the initial entry was itemassessedthelength,inminutes,ofatypicalviewingsession
incomplete, the complete entry was kept for analysis. (adapted from Nelson et al., 2014b). The survey included items
Participants were asked to complete an online survey that to better understand how participants access pornographic
includeditemstoassessdemographicinformation,recentsexual videosontheInternet(i.e.,throughWebsiteswithfreeviewing,
behavior, HIV/STI testing history, history of viewing sexually paid subscription) and viewing context (i.e., at home on a
explicit videosonline, andthecontentofsexuallyexplicitvideos computer, at home on a tablet or smartphone, at home on a
viewedduringthepast6 months.Thesurveytookapproximately television, at work on a computer, at work on a tablet or
15 min to complete. smartphone, at a sex party, in a bar, club or at an adult video
Due to concerns regarding survey length, participants were store). Further, participants were asked to report the percent-
offered the opportunity to be entered into a random drawing for age of time they masturbate while watching pornographic
one of five $20 Amazon.com electronic gift cards. Those who videos on the Internet (none of the time = 0 % to every
wereinterestedinbeingconsideredfortherandomdrawingwere time = 100 %). The survey also included a single item to assess
asked to provide an e-mail address for entry. E-mail addresses the frequency of substance use in the context of watching
were used to distribute e-gift cards to the winners of the drawing. Internet SEM (Downing et al., 2014a). Response options
Participants who completed the survey through Amazon Mechan- included none of the time, some of the time, about half of the
ical Turk received US $0.40, but were not eligible to enter the time, more than half of the time, and every time.
random drawing as this service does not allow for the collection
of personal information such as an e-mail address.
Behaviors viewed in Internet SEM

Measures Participants were asked about the sexual behaviors that actors
were performing in pornographic videos they watched on the
Participant characteristics Internet in the past 6 months. The list of sexual behaviors, along
with instructions to check all that apply, included: solo acts of
The online survey included a set of demographic questions to masturbation, mutual masturbation, oral sex, vaginal sex with a
assess age, race and ethnicity, gender identity, sexual identity, condom, vaginal sex without a condom, anal sex with a condom,
relationship status, education, annual income, housing status, analsexwithoutacondom(barebacking, raw,breeding,seeding),
HIV testing history and status, and zip code (to assess partici- rimming, fisting (vaginal or anal), bondage, sadomasochism
pants’ geographic region in the U.S.). For sexual identity, (S&M), cock and ball torture (CBT), sounding (urethral),
participants were asked the following question:‘‘Do you think watersports (golden showers, pissing/urinating), and felching
of yourself as…?’’ Response options included lesbian, gay or (sucking or eating cum out of someone’s anus). Further, par-
homosexual, straight or heterosexual, bisexual, or something ticipants were asked to report whether or not any of the
else. pornographic videos they had viewed on the Internet in the past
6 months featured group sex scenes with more than two actors
(Nelson et al., 2014b). Response options included: group sex
Internet SEM use
with only men, group sex with only women, and group sex with
men and women.
Survey participants were asked to report (yes or no) if they had
viewed pornographic videos on the Internet (on a computer,
tablet, or smartphone) in the past 6 months. The survey also Condom use preferences in SEM
includedseveralitemstoassessthetypesofSEMthatparticipants
watchedontheInternet(i.e.,videosfeaturingamanandawoman, Participants were asked about their preference for viewing
only men, only women, at least two men and one woman condom use during vaginal and anal sex in SEM (Rosser et al.,
engaging in sexual acts with each other, and at least two women 2013). Specifically, those who indicated that they viewed any

123
Arch Sex Behav (2017) 46:1763–1776 1769

vaginal sex in the past 6 months were asked whether they prefer to Results
watch actors perform vaginal sex with condoms or without. Sim-
ilarly, those who indicated that they viewed any anal sex in the past Comparisons of Participant Characteristics by Sexual
6 months were asked whether they prefer to watch actors perform Identity (N 5 821)
analsexwithcondomsorwithout.Responseoptionsforbothitems
included: without condoms (-1), with condoms (1), and I do not Mean age for the sample was 37.98 years (SD = 12.02). There
care either way (0). were no significant age differences by sexual identity [hetero-
sexual (M = 36.22, SD = 11.53), gay (M = 38.66, SD = 12.13),
bisexual (M = 37.15, SD = 11.94)], F(2, 821) = 2.67, p = .07.
Data Analysis Most of the participants identified their race as Black (45.8 %) or
White (42.1 %). Heterosexual-identified men were more likely
Data cleaning and analyses were performed with IBM SPSS than gay- and bisexual-identified men to be White and less likely
version 22 (IBM, 2013). We reviewed all surveys reporting to be Black. Similarly, gay men were more likely than bisexual
any use of Internet SEM during the past 6 months to ensure men to be White or Hispanic/Latino and less likely to be Black.
that participants also selected at least one type of SEM (i.e., Approximately two-thirds (67.8 %) of the sample reported that
featuring a man and a woman, only men, only women, at least they were single and not currently in a relationship. However,
two men and one woman engaging in sexual acts with each heterosexual men weremore likely than gay andbisexual men to
other, and at least two women and one man engaging in sexual be married or in a domestic partnership and less likely to be
acts with each other) or sexual behavior that they viewed. single. Nearly half of participants had at least a 4-year college
Two participants in the overall analytic sample indicated that degree (46.0 %) and earned less than $40,000 per year (49.5 %).
they viewed at least one type of SEM during the past 6 months, There were no differences in education or income by sexual
butdidnotreportviewinganyspecificsexualbehaviors(perhaps identity. Most participants indicated that their current residence
to indicate that they did not watch any of the behaviors included was a house (48.4 %) or apartment (40.7 %). Heterosexual men
with the item). Because both participants did report at least one were more likely than gay men to report that they currently lived
type of SEM viewed, we retained these surveys for analysis. in a house.Among those who hadeverbeen tested forHIV,more
Descriptive statistics are reported for all participants unless than half reported that their most recent HIV test was negative
otherwise indicated. Comparisons between dichotomous or cat- (55.9 %). Gay men were more likely than bisexual men to report
egorical variables were conducted using chi-square analysis. their HIV serostatus as positive; because no participant identi-
Effect sizes for chi-square tests are reported using uc (Cramer’s V) fyingasheterosexualreportedanHIV-positivestatus,werestricted
(Kline,2013).Comparisonsofcategoricalvariablesoncontinuous this analysis to gay and bisexual men.
variables were conducted using one-way analysis of variance As shown in Table 1, gay and bisexual men were recruited
(ANOVA) for normally distributed variables and Kruskal– primarily through sexual networking and research-oriented
Wallis ANOVA for non-normally distributed variables (with Web sites. Heterosexual men were recruited primarily through
post hoc comparisons using Mann–Whitney U tests). Statisti- Amazon Mechanical Turk and Craigslist. The most represented
callysignificantassociationsarereportedatp\.05.Bonferroni geographic region was the Southeast (32.0 %) followed by the
corrections were applied to the alpha values obtained from post West (21.4 %), Midwest (21.1 %), Northeast (14.6 %), and South-
hoc comparisons. We modeled viewing of high-risk (condom- west (9.9 %). A significantly greater proportion of gay men
less anal and vaginal sex) and protective behaviors (anal and reported a zip code corresponding to a US state in the West
vaginal sex with a condom) using logistic regression (with sexual region compared to bisexual men.
identity as the predictor variable) controlling for participant char-
acteristicssignificantlyassociated(atp\.001)withsexualidentity SEM Viewing Context and Frequency of Use
in bivariate analysis and interactions between sexual identity
and HIV status. Odds ratios and 95 % confidence intervals (CI) MostparticipantsreportedviewingSEMathome:onacomputer
are reported for logistic regression models. We also modeled (85.6 %); on a tablet or smartphone (71.9 %); on a television
condom use preferences in SEM using linear regression, with through an Internet connection (24.2 %); or on a television from
viewing of anal and vaginal sex (with and without a condom) as a DVD or On-Demand service (23.9 %). Although less com-
predictor variables and controlling for participant character- mon, men also reported viewing SEM at work on a tablet or
istics (i.e., sexual identity, race and ethnicity, relationship sta- smartphone(16.1 %)oracomputer(5.4 %).Venue-basedSEM
tus, and HIV status). Unstandardized betas and standard errors viewing was also reported: in a bar, club, or at an adult video
are reported for linear regression models. store (17.9 %); at a sex party (12.4 %). Nearly all men in the

123
1770 Arch Sex Behav (2017) 46:1763–1776

study reported accessing Internet SEM from Web sites with featured only men. Heterosexual men were less likely than gay
free viewing (95.7 %). Only a small percentage of men indi- mentoreportviewingSEMthatfeaturedwatersportsandfelching.
cated having a paid subscription for Internet SEM (10.5 %). They were more likely than gay and bisexual men to report that
Heterosexual men were more likely than gay men to report they viewed SEM featuring group sex with only women, group
viewing SEM at work on a computer, but less likely than gay sex with men and women, and vaginal sex without a condom.
men to view it at home on a television (e.g., DVD, On-De- Heterosexual men were also more likely than gay men to report
mand),ata sex party,orinacommercialsex venue suchasa bar, viewing SEM that featured vaginal sex with a condom.
club, or adult video store (Table 2). Gay men were more likely Gay men were significantly more likely than bisexual men
than bisexual men to report viewing SEM while at a sex party or to report that they viewed SEM featuring group sex with only
acommercialsexvenue.Whenaskedaboutsubstanceuseinthe men, bondage, sadomasochism, cock and ball torture, sounding,
context of viewing Internet SEM, more than two-thirds (69.2 %) fisting,watersports,andfelching.However,theywerelesslikely
of heterosexual men indicated that they did this none of the time. thanbisexualmentoreportviewingSEMthatfeaturedgroupsex
Although there were no differences between gay and bisexual with only women, group sex with men and women, solo acts of
men, gay men were more likely than heterosexual men to report masturbation, and vaginal sex with or without a condom.
that they used substances while viewing Internet SEM. Hetero- Sexual identity significantly predicted viewing of anal sex
sexual men were less likely than bisexual men to indicate doing with a condom (referent: heterosexual; gay OR 3.93, 95 % CI
this some of the time. 2.64–5.83; bisexual OR 4.59, 95 % CI 2.78–7.57), anal sex
As shown in Table 2, heterosexual men were more likely to without a condom (referent: heterosexual; gay OR 5.20, 95 %
view Internet SEM once a week or less compared to gay and CI 3.35–8.09; bisexual OR 3.99, 95 % CI 2.24–7.10), vaginal
bisexual men who were more likely to view Internet SEM at sex with a condom (referent: gay; heterosexual OR 7.90, 95 %
least once a day. Gay and bisexual men viewed more hours of CI 5.19–12.03; bisexual OR 4.97, 95 % CI 3.32–7.44), and
Internet SEM in a typical week (Mdn = 3 h, IQR = 1–5 for vaginal sex without a condom (referent: gay; heterosexual
both groups) compared to heterosexual men (Mdn = 2 h, OR 27.08, 95 % CI 15.25–48.07; bisexual OR 5.59, 95 % CI
IQR = 1–3.75), v2(2, N = 819) = 17.19, p\.001. Further, a 3.81–8.21) in separate logistic regression models. In multi-
significantly greater proportion of heterosexual men reported variable analyses controlling for race and ethnicity, relationship
SEM viewing sessions lasting 10 min or less compared to gay status, and HIV status, sexual identity remained a significant
and bisexual men. Differences by sexual identity in the per- predictorforviewingriskandprotectivebehaviors.Specifically,
centage of time that participants masturbate while watching sexual identity significantly predicted viewing of anal sex with a
Internet SEM approached significance (gay [Mdn = 94.00, condom (gay AOR 4.53, 95 % CI 2.75–7.47; bisexual AOR
IQR = 64.00–100.00], bisexual [Mdn = 90.00, IQR = 50.00– 3.53, 95 % CI 1.96–6.35), anal sex without a condom (gay AOR
100.00], heterosexual [Mdn = 85.50, IQR = 54.25–100.00]), 4.94, 95 % CI 2.75–8.87; bisexual AOR 4.03, 95 % CI
v2(2, N = 812) = 5.20, p = .07. 2.00–8.11), vaginal sex witha condom (heterosexual AOR 9.47,
95 % CI 5.44–16.46; bisexual AOR 4.69, 95 % CI 2.98–7.39),
and vaginal sex without a condom (heterosexual AOR 22.94,
Behaviors Viewed in Internet SEM and Condom Use 95 % CI 11.83–44.49; bisexual AOR 5.76, 95 % CI 3.73–8.89).
Viewing Preferences Men who self-identified as Black or African American, com-
pared to White or Caucasian, had significantly increased odds of
The most common sexual behaviors men reported viewing in reporting that they viewed anal sex with a condom (AOR 1.52,
SEMwereoralsex(86.2 %),analsexwithoutacondom(84.0 %), 95 %CI1.06–2.20).MenwithanHIV-positivestatus,compared
anal sex with a condom (65.3 %), rimming (64.8 %), solo acts of to HIV-negative, had significantly decreased odds of reporting
masturbation (53.1 %), vaginal sex without a condom (40.8 %), that they viewed anal sex with a condom (AOR 0.49, 95 % CI
andmutual masturbation(39.2 %). Menalsoreported viewing,to 0.33–0.72). Participant characteristics were not significantly
a lesser extent, felching (28.5 %), vaginal sex with a condom associated with viewing of anal sex without a condom, vaginal
(26.4 %), bondage/sadomasochism/cock and ball torture/sound- sex with a condom, or vaginal sex without a condom. Adding the
ing (24.7 %), watersports (23.9 %), and fisting (18.1 %). Group interaction between HIV status and sexual identity did not result
sex scenes featuring only men were highly reported (80.1 %) in a significant change to any of the multivariable models.
followed by scenes featuring men and women (46.5 %), and Of those participants who viewed Internet SEM featuring
only women (12.8 %). analsexinthepast6 months,nearlytwo-thirds(61.3 %)reported
Heterosexual men were significantly less likely than gay a preference for viewing condomless anal sex and only 6.8 %
and bisexual men to report that they viewed SEM featuring only reported a preference for viewing anal sex with condoms. Gay
men, group sex with only men, mutual masturbation, anal sex men were more likely than bisexual men to indicate a preference
withorwithoutacondom,andrimming(Table 2).However,one for viewing condomless anal sex. Among those who viewed
in five heterosexual-identified men reported viewing SEM that Internet SEM featuring vaginal sex in the past 6 months, more

123
Arch Sex Behav (2017) 46:1763–1776 1771

Table 2 Comparisons of sexually explicit media use, behaviors viewed, and preferences for condom use by sexual identity
Heterosexual Gay Bisexual v2 uc Post hoc
(a) (b) (c)

Viewed SEM in the past 6 months… (N = 821)

At home on a computer 89.6 85.2 83.7 2.23 0.05
At home on a tablet or smartphone 65.7 74.0 69.9 3.99 0.07
At work on a computer 9.7 4.3 5.2 6.15* 0.09 a[b
At work on a tablet or smartphone 12.7 17.2 15.0 1.79 0.05
At home on a television (e.g., DVD, On-Demand) 14.9 27.5 19.0 11.86** 0.12 a\b
At home on a television via Internet connection 20.1 26.0 21.6 2.75 0.06
In a bar, club, or at an adult video store 5.2 23.2 10.5 30.70*** 0.19 a\b; b[c
At a sex party 4.5 15.9 7.2 17.62*** 0.15 a\b; b[c
Internet SEM viewing frequency (N = 816) 31.34*** 0.14
Once a week or less 37.9 18.4 17.8 a[b, c
Two to three times a week 32.6 30.3 29.6
Once a day or more 29.5 51.3 52.6 a\b, c
Length of typical SEM viewing session (N = 818) 10.60* 0.08
B10 min 41.7 28.7 25.5 a[b, c
30–45 min 38.6 45.2 47.7
1 h or more 19.7 26.1 26.8
Substance use while viewing Internet SEM (N = 814) 11.68* 0.08
None of the time 69.2 54.4 55.6 a[b
Some of the time 18.8 29.0 32.0 a\c
Half or more of the time 12.0 16.7 12.4
Types of Internet SEM vieweda
SEM featuring only men 20.7 98.3 96.0 513.16*** 0.80 a\b, c
SEM featuring a man and a woman 98.5 55.0 88.3 124.26*** 0.40 a[b, c; b\c
SEM featuring only women 83.6 3.3 41.1 387.55*** 0.72 a[b, c; b\c
SEM featuring at least two men and one woman 76.9 52.9 80.4 50.42*** 0.26 a[b; b\c
SEM featuring at least two women and one man 88.5 22.5 61.8 215.32*** 0.53 a[b, c; b\c
Internet SEM behaviors (N = 821)
Solo acts of masturbation 58.2 49.6 60.8 7.62* 0.10 b\c
Mutual masturbation 24.6 42.3 41.2 14.37** 0.13 a\b, c
Oral sex 89.6 85.2 86.9 1.78 0.05
Vaginal sex without a condom 88.8 22.7 62.1 229.32*** 0.53 a[b, c; b\c
Vaginal sex with a condom 56.0 13.9 44.4 129.07*** 0.40 a[b; b\c
Anal sex without a condom 61.2 89.1 86.3 63.07*** 0.28 a\b, c
Anal sex with a condom 37.3 70.0 73.2 55.81*** 0.26 a\b, c
Rimming 32.1 72.7 66.0 77.42*** 0.31 a\b, c
Fisting (vaginal or anal) 14.9 21.0 11.1 8.91* 0.10 b[c
Bondage, sadomasochism, cock and ball torture, soundingb 24.6 27.9 13.7 12.84** 0.12 b[c
c
Watersports 13.4 28.7 16.3 19.52*** 0.15 a\b; b[c
Felchingd 12.7 33.9 23.5 25.93*** 0.18 a\b; b[c
Group sex with only men 8.2 95.9 88.2 525.17*** 0.80 a\b, c; b[c
Group sex with only women 44.8 2.8 19.6 176.99*** 0.46 a[b, c; b\c
Group sex with men and women 85.8 32.6 60.8 137.38*** 0.41 a[b, c; b\c
Preferences for vaginal sex condom use in SEM (N = 413) 10.19* 0.11
No preference 31.7 45.1 41.5
Without condoms 65.0 50.5 50.0 a[b

123
1772 Arch Sex Behav (2017) 46:1763–1776

Table 2 continued
Heterosexual Gay Bisexual v2 uc Post hoc
(a) (b) (c)

With condoms 3.3 4.3 8.5

Preferences for anal sex condom use in SEM (N = 754) 10.66* 0.08
No preference 33.0 29.2 41.3 b\c
Without condoms 61.7 64.4 49.7 b[c
With condoms 5.3 6.4 9.1
Percentages presented. SEM (sexually explicit media)
*** p\.001; ** p\.01, *p\.05. Bonferroni corrections applied to post hoc comparisons (p\.017)
a
Missing data resulted in analytic samples ranging from 743 to 802. b Urethral sounding. c Golden showers, pissing/urinating. d Sucking or eating cum
out of someone’s anus. v2. uc (Cramer’s V)

than half (54.7 %) reported a preference for viewing condom- viewed SEM featuring only men, mutual masturbation, and
less vaginal sex and 5.1 % reported a preference for viewing group sex with only men during the past 6 months.
vaginal sex with a condom. Compared to gay men, heterosexual Nevertheless, the findings also indicated that many men
men were more likely to indicate a preference for viewing con- viewed SEM content inconsistent with their stated sexual iden-
domless vaginal sex. Bisexual men did not differ from either gay tity. It was not uncommon for heterosexual-identified men to
or heterosexual men in their preference for viewing condomless report viewingSEMcontainingmalesame-sexbehavior(20.7 %)
vaginal sex. and for gay-identified men to report viewing heterosexual behav-
We modeled condom use preferences for vaginal sex in SEM ior in SEM (55.0 %). It was also not uncommon for gay men to
using linear regression, with viewing of vaginal sex with a con- report that they viewed vaginal sex with (13.9 %) and without a
dom and without a condom as predictor variables. Unstandard- condom (22.7 %) during the past 6 months. Stein et al. (2012)
ized betas and standard errors are reported in Table 3. The model, found that nearly half of their sample of MSM had ever viewed
controlling for participant characteristics (i.e., sexual identity, heterosexual SEM. Interestingly, though not clear why, bisex-
race and ethnicity, relationship status, and HIV status), explained ual men were more likely than gay men, but not heterosexual
23.9 % of the variance in condom use preferences for vaginal sex men, to report viewing solo acts of masturbation in SEM. Hetero-
in SEM. Viewing vaginal sex with a condom and without a sexualmendidnotdifferfromgaymenintheirviewingofSEMthat
condom was both significantly associated (in expected direc- featured fisting, bondage, sadomasochism, cock and ball tor-
tions) with condom use preferences for vaginal sex. Similarly, ture, and sounding. However, gay men were more likely than
we modeled condom use preferences for anal sex in SEM, with heterosexual and bisexual men to report viewing SEM that
viewing of anal sex with a condom and without a condom as featured watersports and felching, behaviors that have been
predictor variables. The model, controlling for participant found in certain genres (e.g., kink, fetish) of gay male SEM
characteristics, explained 21.7 % of the variance in condom (Downing et al., 2014b). Likewise, gay men were more likely
use preferences for anal sex in SEM. Viewing anal sex with a than bisexual men to report viewing of SEM that featured
condom and without a condom was both significantly asso- fisting, bondage, sadomasochism, cock and ball torture, and
ciated (in expected directions) with condom use preferences sounding in the past 6 months.
for anal sex. These data further suggest a need for clarification of what
is considered bisexual SEM. Hald and Štulhofer (2016) assessed
variations of bisexual behavior in SEM, including several (e.g.,
Discussion threesomes, orgy) that loaded onto a bisexual viewing factor for
heterosexual men. Given the presence of sexual identity dis-
This study is one of the first to compare the behavioral content crepancy in SEM viewing (as reported in the current study and
of Internet SEM viewed by sexual identity, building on prior by Hald & Štulhofer, 2016), there are important questions that
work by Peter and Valkenburg (2012). Findings suggest that warrant subsequent inquiry. More specifically, are bisexual sex
the behaviors men view in SEM tended to reflect their sexual scenes mostly embedded within heterosexual, lesbian, and/or
identity. Specifically, heterosexual men were more likely than gay SEM or is there a distinct presence of this media type
gay and bisexual men to report that they viewed SEM featuring beyond the gender and quantity of actors in a particular scene?
women, vaginal sex, group sex with only women, and group sex Catalog analysis of leading adult industry studios coupled with
withmenandwomenduringthepast6 months.Gayandbisexual a behavioral content analysis of select videos will likely pro-
men were more likely than heterosexual men to report that they vide some answers to these questions. Nevertheless, viewers

123
Arch Sex Behav (2017) 46:1763–1776 1773

Table 3 Multiple linear regressions predicting condom use preferences bisexual-identified men who engage in same-sex encounters,
in SEM avoid these types of venues for fear of discovery (Schrimshaw,
Condom Use
Downing, & Siegel, 2013a). Since multiple researchers have
Preferences in established cross-sectional associations between viewing con-
SEM domless anal sex in SEM and engaging in condomless anal sex
Anal among GBMSM (Nelson et al., 2014b; Schrimshaw et al.,
2016a; Stein et al., 2012), this finding does raise concerns about
b SE
the potential role of SEM in facilitating high-risk encounters in
Viewing of anal sex without a condom -0.59*** 0.08 sexually charged environments. Knowing that some men con-
Viewing of anal sex with a condom 0.29*** 0.05 sume SEM in these contexts may be critical to tailoring delivery
Gay or homosexual -0.07 0.07 strategies of risk reduction messages. Additionally, study find-
Bisexual -0.03 0.08 ings provide further evidence that the workplace is a common
Black or African-American 0.25*** 0.05 site for accessing SEM (Albright, 2008; Perrin et al., 2008),
Hispanic or Latino 0.24** 0.08 particularlyonatabletorsmartphone.Althoughthefrequencyof
Asian, Pacific Islander, Native American or Alaska 0.13 0.12 accessing SEM at work on a computer was low, heterosexual
Native, Native Hawaiian, or Other men were significantly more likely than gay men to have done so
In a steady relationship -0.002 0.06 during the past 6 months. Further research in this area might
Married or domestic partnership -0.05 0.06 consider whether men who view SEM at work are more sexually
HIV-positive -0.13** 0.05 compulsive and whether viewing in this context has a negative
Never tested for HIV 0.01 0.08 impactonproductivityandinterpersonalworkplacerelationships.
As other researchers have reported (Duggan & McCreary,
Vaginal
2004; Peter & Valkenburg, 2011b), the current study found that
b SE use of Internet SEM varied by sexual identity with heterosexual
Viewing of vaginal sex without a condom -0.53*** 0.07 men accessing it less frequently than gay and bisexual men.
Viewing of vaginal sex with a condom 0.37*** 0.06
Heterosexual men were also more likely to report viewing
Straight or heterosexual -0.14  0.08
sessions of 10 min or less. Study findings suggest that hetero-
sexual men are less likely to smoke, consume alcohol or other
Bisexual -0.02 0.07
drugswhileviewingInternetSEMcomparedtogayandbisexual
Black or African-American 0.13  0.07
men.Indeed,nearlyhalfofgay(45.7 %)andbisexualmen(44.4 %)
Hispanic or Latino 0.25* 0.11
indicated using substances at least some of the time in this
Asian, Pacific Islander, Native American -0.13 0.13
or Alaska Native, Native Hawaiian, or Other context. Moreover, bisexual men were significantly more likely
In a steady relationship -0.01 0.08 than heterosexual men to report that they did this some of the
Married or domestic partnership 0.02 0.07 time. Additional research could assess the substances men are
HIV-positive -0.18* 0.08
using while viewing SEM and what motivates their substance
Never tested for HIV 0.07 0.08
use in this context. Furthermore, there was a high rate of mas-
turbation while viewing Internet SEM though no differences by
Unstandardized coefficients (b) and standard error (SE) are reported sexualidentitywerefound.Thisfindingisnotsurprisingasother
*** p\.001; ** p\.01; * p\.05;  p\.10 studies have reported similar rates of masturbation among men
whileconsumingSEM(Kraus& Rosenberg,2014; Nelson et al.,
2014b).
may have a different perspective based on their SEM search Despite growing attention to the potential negative implica-
patterns and preferences that deserves further consideration. tionsofbarebacksexingaymaleSEM(e.g.,Nelsonetal.,2014a;
This study provides important insights into the contexts of Rosser et al., 2012; Schrimshaw et al., 2016a), there has been a
SEM use. Not surprisingly, participants overwhelmingly repor- lack of equivalent research pertaining to the use of condoms for
ted viewing SEM at home either on a computer, on a tablet or vaginal and anal sex in heterosexual and bisexual media. The
smartphone,oronatelevision.Nevertheless,amodestproportion currentstudysoughttobridgethisgapintheliteraturebyassessing
of men reported viewing SEM while attending sex parties or thetypesofsexualbehaviorsthatmen(ofdiversesexualidentities)
commercial sex venues. Gay men were significantly more likely viewed in Internet SEM during the past 6 months, including con
than heterosexual and bisexual mentohaveviewed SEMat a sex domless vaginal and anal sex. Consistent with findings from
party or commercial sex venue. This is perhaps due to group dif- recent content analyses of gay male and heterosexual SEM
ferences in frequency of attendance or the type of venue atten- (Downing et al., 2014; Grudzen et al., 2009; Salmon &
ded. Some research has also suggested that men who are less Diamond,2012),participantsoverwhelminglyreportedviewing
open about their sexuality, particularly heterosexual- and Internet SEM that featured at least one of these risk behaviors.

123
1774 Arch Sex Behav (2017) 46:1763–1776

Moreover, this study builds on research by Kraus and Rosenberg Limitations

(2016) who assessed attitudes toward condom use in SEM
among heterosexual and non-heterosexual men. The authors did Severalstudylimitationsshouldbenoted.First,thedataarecross-
not specifically inquire about the types of behaviors that partic- sectional in nature precluding any causal conclusions about
ipants viewed, thus precluding any analyses into whether their condom use preferences in SEM and behavioral content viewed
attitudespredictedviewing.Thecurrentstudy,however,didassess (i.e., does exposure influence viewing preferences, do prefer-
behavioral content viewed and observed significant associations ences influence what viewers search for and watch, or is it
with preferences for condom use in SEM (also serving as a validity reciprocal and mutually reinforcing). Further, the study relied
check). on self-report retrospective data collected from an online non-
Previous research with GBMSM has shown significant asso- probability sample. Thus, there are potential concerns about
ciations between viewing condomless anal sex in gay male SEM recall bias as well as participant selection and external validity.
and engaging in more condomless anal sex encounters (Nelson To minimize issues with recall, the survey utilized a 6-month
et al., 2014b; Rosser et al., 2013; Schrimshaw et al., 2016a; Stein timeframe to assess recent SEM use. Additionally, the survey
et al., 2012). The current study found that a substantial number of did not include items to assess the types of behavioral content
heterosexual and bisexual men reported viewing condomless participants searched for in SEM, but rather what they actually
vaginalandanalsexinSEM.Thus,itispossiblethatviewingsuch viewed. As such, it remains unclear whether men intended to
high-riskbehaviorsinSEMmaybeassociatednotonlywithmore watch those behaviors that they reported viewing. Nevertheless,
condomless anal sex with male partners (among bisexual men), there was evidence of validity within the findings as preferences
but also with more condomless encounters with female partners for viewing condom use in SEM (a likely marker for content
amongheterosexualandbisexualmen.Furtherresearchisneeded searching)weresignificantlyassociatedwithbehavioralcontent
to examine these potential associations, particularly with regard viewed—specifically condom use or nonuse during scenes fea-
to female partners, as this is an understudied area. Likewise, we turing vaginal and anal sex—during the past 6 months. Lastly,
alsofoundthatgaymenreportviewingcondomlessvaginalsexin we noted several differences between participants who com-
SEM. Additional research should examine whether this viewing pleted the survey and those who only provided a partial survey.
contributes at all to risky sexual encounters with male partners. If Thus, the findings may not necessarily generalize to those
any of these associations are confirmed in subsequent studies, it participants who did not complete the survey.
would suggest that SEM-based risk reduction messaging (as an
HIV/STI prevention intervention) must not only target gay male
Conclusions
media but heterosexual and bisexual media as well.
Further research is also needed to better understand the
The results from this study have extended prior SEM research
mechanisms underlying diverse SEM viewing patterns. For
bycomparingthefrequencyofuse,viewingcontexts,behavioral
example, some heterosexual-identified men may experience
content viewed, and viewer preferences for condom use during
sexual arousal from the homosociality or patterns of male
vaginal and anal sex scenes by sexual identity among an ethni-
bonding (including BDSM) inherent to gay male SEM. For
cally diverse sample of men in the U.S. Gay and bisexual men
gay-identified men who watch bisexual sex scenes, perhaps
reported significantly more frequent use of Internet SEM com-
their interest (gaze) lies not with the female actor(s), but with
pared to heterosexual men. Although most participants reported
the one or more male actors engaging in sexual activity.
viewing SEM at home on a computer, tablet, or smartphone,
Moreover, some men are perhaps aroused by the domination
significantly more gay men reported SEM use at a sex party or
(implied or otherwise) incorporated into bisexual and other
commercial sex venue. Sexual identity predicted viewing of
group sex scenes and may fantasize that they are the person
high-risk (condomless anal and vaginal sex) and protective
being dominated. Of course we recognize that viewing pat-
behaviors (anal and vaginal sex with a condom). Nevertheless,
terns may change over time and are likely associated with
there was evidence of identity discrepant SEM viewing as
more than just sexual identity, including attractions to one or
heterosexual-identified men reported viewing male same-sex
more gender types, SEM content-related sexual arousal, and
behavior and gay-identified men reported viewing heterosexual
prior sexual experiences. Also embedded in this discussion is
behavior. Significant associations were also observed between
the idea that sexual identity is not always reflective of one’s
behavioral content viewed and preferences for condom use in
attractions or experiences (Baldwin et al., 2015). Indeed, the
SEM. Findings suggest the importance of assessing SEM use
current study found evidence of sexual identity discrepancy
across media types (e.g., SEM that targets heterosexual, bisex-
in behavioral content viewed. This suggests, as we have noted
ual, and gay audiences) and contexts and have implications for
above, that SEM-based risk reduction messages have the
future SEM research and prevention strategies to address con-
potential to reach a broad audience.
cerns about the potential influence of SEM on sexual behavior.

123
Arch Sex Behav (2017) 46:1763–1776 1775

Compliance with Ethical Standards Grudzen, C. R., Elliott, M. N., Kerndt, P. R., Schuster, M. A., Brook, R.
H., & Gelberg, L. (2009). Condom use and high-risk sex acts in
Conflict of interest This research was supported by a grant from the adult films: A comparison of heterosexual and homosexual films.
Foundation for the Scientific Study of Sexuality to Martin J. Downing, American Journal of Public Health, 99, S152–S156.
Jr., Ph.D. (no award number provided). The authors declare that they Hald, G. M., & Malamuth, N. M. (2008). Self-perceived effects of
have no other conflicts of interest. pornography consumption. Archives of Sexual Behavior, 37, 614–625.
Hald, G. M., Smolenski, D., & Rosser, B. R. S. (2013). Perceived effects
Human Participants All procedures performed in studies involving of sexually explicit media among men who have sex with men and
human participants were in accordance with the ethical standards of the psychometric properties of the pornography consumption effects
institutional and/or national research committee and with the 1964 scale (PCES). Journal of Sexual Medicine, 10, 757–767.
Helsinki Declaration and its later amendments or comparable ethical Hald, G. M., & Štulhofer, A. (2016). What types of pornography do
standards. people use and do they cluster? Assessing types and categories of
pornography consumption in a large-scale online sample. Journal
Informed Consent Informed consent was obtained from all individual of Sex Research, 53(7), 849–859.
participants included in the study. Hare, K. A., Gahagan, J., Jackson, L., & Steenbeek, A. (2015).
Revisualising ‘porn’: How young adults’ consumption of sexually
explicit internet movies can inform approaches to Canadian sexual
health promotion. Culture, Health and Sexuality, 17, 269–283.
Holmes, D., O’Byrne, P., & Gastaldo, D. (2007). Setting the space for
References sex: Architecture, desire and health issues in gay bathhouses. Inter-
national Journal of Nursing Studies, 44, 273–284.
Albright, J. M. (2008). Sex in America online: An exploration of sex, Hurley, R. (2009). How gay porn undermines safe sex campaigns. British
marital status, and sexual identity in internet sex seeking and its Medical Journal, 338, 775.
impacts. Journal of Sex Research, 45(2), 175–186. Jonas, K. J., Hawk, S. T., Vastenburg, D., & de Groot, P. (2014).
Baldwin, A., Dodge, B., Schick, V., Hubach, R. D., Bowling, J., Malebranche, ‘‘Bareback’’ pornography consumption and safe-sex intentions of
D., … Fortenberry, J. D. (2015). Sexual self-identification among men having sex with men. Archives of Sexual Behavior, 43, 745–
behaviorally bisexual men in the Midwestern United States. Archives of 753.
Sexual Behavior, 44, 2015–2026. Kinser, J. (2014). Michael Lucas explains his controversial decision to
Bishop, C. J. (2015). ‘Cocked, locked and ready to fuck?’: A synthesis make condom-free adult films. Retrieved June 8, 2014 from http://
and review of the gay male pornography literature. Psychology and www.queerty.com/michael-lucas-explains-his-controversial-
Sexuality, 6(1), 5–27. decision-to-make-condom-free-adult-films-20140331.
Boies,S.C.(2002).Universitystudents’usesofandreactionstoonlinesexual Kline, R. B. (2013). Beyond significance testing. Statistics reform in the
information and entertainment: Links to online and offline sexual behavioral sciences (2nd ed.). Washington, DC: American Psycho-
behaviour. Canadian Journal of Human Sexuality, 11, 77–89. logical Association.
Cronin, B., & Davenport, E. (2001). E-rogenous zones: Positioning pornog- Kraus, S., & Rosenberg, H. (2014). The Pornography Craving Questionnaire:
raphy in the digital economy. The Information Society, 17, 33–48. Psychometric properties. Archives of Sexual Behavior, 43, 451–462.
Dodge, B., Schnarrs, P. W., Reece, M., Martinez, O., Goncalves, G., Kraus, S. W., & Rosenberg, H. (2016). Lights, camera, condoms! Assessing
Malebranche, D., … Fortenberry, J. D. (2012). Individual and social college men’s attitudes toward condom use in pornography. Journal of
factors related to mental health concerns among bisexual men in the American College Health, 64(2), 139–146.
Midwestern United States. Journal of Bisexuality, 12, 223–245. Kubicek, K., Beyer, W. J., Weiss, G., Iverson, E., & Kipke, M. D. (2010).
Downing, M. J., Jr., Antebi, N., & Schrimshaw, E. W. (2014a). In the dark: Young men’s stories of sexual initiation in the absence
Compulsive use of internet-based sexually explicit media: Adap- of relevant sexual health information. Health Education and Behav-
tation and validation of the Compulsive Internet Use Scale (CIUS). ior, 37, 243–263.
Addictive Behaviors, 39, 1126–1130. Kuchment, A., & Springen, K. (2008). The tangled web of porn in the
Downing, M. J., Jr., Schrimshaw, E. W., Antebi, N., & Siegel, K. office. Newsweek, 152(23), 14.
(2014b). Sexually explicit media on the internet: A content analysis Morgan, E. M. (2011). Associations between young adults’ use of sexually
of sexual behaviors, risk, and media characteristics in gay male explicit materials and their sexual preferences, behaviors, and satisfac-
adult videos. Archives of Sexual Behavior, 43, 811–821. tion. Journal of Sex Research, 48(6), 520–530.
Duggan, S. J., & McCreary, D. R. (2004). Body image, eating disorders, Nelson, K. M., Leickly, E., Yang, J. P., Pereira, A., & Simoni, J. M.
and the drive for muscularity in gay and heterosexual men: The (2014a). The influence of sexually explicit online media on sex: Do
influence of media images. Journal of Homosexuality, 47(3–4), men who have sex with men believe they‘‘do what they see’’? AIDS
45–58. Care, 26(7), 931–934.
Escoffier, J. (2009). Bigger than life: The history of gay porn cinema Nelson, K. M., Simoni, J. M., Morrison, D. M., George, W. H., Leickly,
from beefcake to hardcore. Philadephia: Running Press Book E., Lengua, L. J., & Hawes, S. E. (2014b). Sexually explicit online
Publishers. media and sexual risk among men who have sex with men in the
Goldstein, B. Y., Steinberg, J. K., Aynalem, G., & Kerndt, P. R. (2011). High United States. Archives of Sexual Behavior, 43, 833–843.
chlamydia and gonorrhea incidence and reinfection among performers Nichols, J. S. (2011). Let’s talk about sex blog. Retrieved June 8, 2014 from
in the adult film industry. Sexually Transmitted Diseases, 38(7), 644– http://ltasex.info/home/2011/4/25/paul-morris-treasure-island-medias-
648. maverick-sets-the-record.html.
Green, S. T. (2004). HIV and AIDS, the internet pornography industry Paul, B. (2009). Predicting internet pornography use and arousal: The
and safer sex. International Journal of STD and AIDS, 15, 206–208. role of individual difference variables. Journal of Sex Research,
Griffith, J. D., Adams, L. T., Hart, C. L., Mitchell, S., Kruger, A., Phares, 46(4), 344–357.
B., … Finkenbinder, A. (2012). Pornography actors: A qualitative Paul, B., & Shim, J. W. (2008). Gender, sexual affect, and motivations for
analysis of motivations and dislikes. North American Journal of internet pornography use. International Journal of Sexual Health,
Psychology, 14(2), 245–256. 20, 187–199.

123
1776 Arch Sex Behav (2017) 46:1763–1776

Perrin, P. C., Madanat, H. N., Barnes, M. D., Carolan, A., Clark, R. B., Schrimshaw, E. W., Siegel, K., Downing, M. J., Jr., & Parsons, J. T.
Ivins, N., … Williams, P. N. (2008). Health education’s role in fram- (2013b). Disclosure and concealment of sexual orientation and the
ing pornography as a public health issue: Local and national mental health of non-gay-identified, behaviorally bisexual men.
strategies with international implications. Promotion and Education, Journal of Consulting and Clinical Psychology, 81, 141–153.
15(1), 11–18. Short, M. B., Black, L., Smith, A. H., Wetterneck, C. T., & Wells, D. E.
Peter, J., & Valkenburg, P. M. (2011a). The influence of sexually explicit (2012). A review of internet pornography use research: Method-
material on sexual risk behavior: A comparison of adolescents and ology and content from the past 10 years. Cyberpsychology, Behav-
adults. Journal of Health Communication, 16, 750–765. ior, and Social Networking, 15, 13–23.
Peter, J., & Valkenburg, P. M. (2011b). The use of sexually explicit Stein, D., Silvera, R., Hagerty, R., & Marmor, M. (2012). Viewing pornog-
internet material and its antecedents: A longitudinal comparison of raphy depicting unprotected anal intercourse: Are there implications for
adolescents and adults. Archives of Sexual Behavior, 40, 1015– HIVpreventionamongmenwhohavesexwithmen?ArchivesofSexual
1025. Behavior, 41, 411–419.
Peter, J., & Valkenburg, P. M. (2012). Do questions about watching Štulhofer, A., Buško, V., & Landripet, I. (2010). Pornography, sexual social-
internet pornography make people watch internet pornography? A ization,andsatisfactionamongyoungmen.ArchivesofSexualBehavior,
comparison between adolescents and adults. International Journal 39, 168–178.
of Public Opinion Research, 24(3), 400–410. Sun, C., Bridges, A., Johnason, J., & Ezzell, M. (2016). Pornography and
Rosser, B. R. S., Grey, J. A., Wilkerson, J. M., Iantaffi, A., Brady, S. S., the male sexual script: An analysis of consumption and sexual
Smolenski, D. J., & Horvath, K. J. (2012). A commentary on the relations. Archives of Sexual Behavior, 45(4), 983–994.
role of sexually explicit media (SEM) in the transmission and Traeen, B., & Daneback, K. (2013). The use of pornography and sexual
prevention of HIV among men who have sex with men (MSM). behaviour among Norwegian men and women of differing sexual
AIDS and Behavior, 16(6), 1373–1381. orientation. Sexologies, 22, e41–e48.
Rosser, B. S., Smolenski, D. J., Erickson, D., Iantaffi, A., Brady, S. S., Traeen, B., Nilsen, T. S., & Stigum, H. (2006). Use of pornography in
Grey, J. A., … Wilkerson, J. M. (2013). The effects of gay sexually traditional media and on the internet in Norway. Journal of Sex
explicit media on the HIV risk behavior of men who have sex with Research, 43, 245–254.
men. AIDS and Behavior, 17, 1488–1498. TrafficJunky Media Kit. (2015). Retrieved May 10, 2016 from http://
Salmon, C., & Diamond, A. (2012). Evolutionary perspectives on the www.trafficjunky.net.
content analysis of heterosexual and homosexual pornography. Weinberg, M. S., Williams, C. J., Kleiner, S., & Irizarry, Y. (2010). Pornog-
Journal of Social, Evolutionary, and Cultural Psychology, 6(2), raphy, normalization, and empowerment. Archives of Sexual Behavior,
193–202. 39, 1389–1401.
Schnarrs, P. W., Dodge, B., Reece, M., Goncalves, G., Martinez, O., Pol, Wilkerson, J. M., Iantaffi, A., Smolenski, D. J., Brady, S. S., Horvath, K.
B. V., … Fortenberry, J. D. (2012). Subjective sexual experiences J., Grey, J. A., & Rosser, B. R. S. (2012). The SEM risk behavior
of behaviorally bisexual men in the midwestern United States: (SRB) model: A new conceptual model of how pornography
Sexual attraction, sexual behaviors and condom use. Journal of influences the sexual intentions and HIV risk behavior of MSM.
Bisexuality, 12, 246–282. Sexual and Relationship Therapy, 27(3), 217–230.
Schrimshaw, E. W., Antebi-Gruszka, N., & Downing, M. J., Jr. (2016a). Wright, P. J. (2013). U.S. males and pornography, 1973–2010: Con-
Viewing of internet-based sexually explicit media as a risk factor sumption, predictors, correlates. Journal of Sex Research, 50, 60–
for condomless anal sex among men who have sex with men in four 71.
U.S. cities. PLoS One, 11, e0154439. Wright, P. J., & Tokunaga, R. S. (2016). Men’s objectifying media
Schrimshaw, E. W., Downing, M. J., Jr., & Cohn, D. J. (2016b). Reasons consumption, objectification of women, and attitudes supportive of
for non-disclosure of sexual orientation among behaviorally violence against women. Archives of Sexual Behavior, 45, 955–964.
bisexual men: Non-disclosure as stigma management. Archives Wright, P. J., Tokunaga, R. S., & Kraus, A. (2016). Consumption of
of Sexual Behavior. doi:10.1007/s10508-016-0762-y. pornography, perceived peer norms, and condomless sex. Health
Schrimshaw, E. W., Downing, M. J., Jr., & Siegel, K. (2013a). Sexual venue Communication, 31(8), 954–963.
selection and strategies for concealment of same-sex behavior among
non-disclosing men who have sex with men and women. Journal of
Homosexuality, 60, 120–145.

123