Perceptions of Power and Sexual Pleasure Associated with Sexual Behaviour Profiles Among Latino Sexual Minority Men

Abstract

In a body of research typically focused on risk reduction and disease prevention, other factors motivating the sexual behaviours of Latino sexual minority men, such as resource-based power and sexual pleasure, are less understood. To this end, Latino immigrant sexual minority men living in New York City were surveyed about their sexual behaviours, perceived power differentials, and pleasure from insertive and receptive anal intercourse. Power and pleasure were examined as associations with behavioural profiles identified through latent class analysis, adjusting for age and partner type. Four latent classes of Latino sexual minority men were identified based on behaviours reported during the most recent sexual event: behaviourally insertive (14.2%), behaviourally versatile (25.9%), behaviourally receptive (29.2%), and limited penetrative behaviour (30.7%). Participants who derived pleasure from insertive and receptive anal intercourse had higher odds of belonging in the behaviourally insertive and behaviourally receptive class, respectively. Perceptions of resource-based power were not associated with class membership. Findings highlight the importance of sexual pleasure as a driver of sexual behaviour, irrespective of power dynamics. Sexual health curricula and interventions for sexual minority men should consider sexual pleasure and sex-affirmative frameworks when providing sexuality education and promoting sexual wellbeing.

Introduction

Latino sexual minority men face disproportionate risk for HIV relative to their White counterparts (Chen et al. 2012; Crepaz et al. 2019; Harawa et al. 2004). Despite the implementation of behavioural and psychosocial interventions to promote condom use and HIV testing among this population (Rhodes et al. 2015; 2017), the number of new HIV diagnoses between 2012 and 2016 has increased by 12% among Hispanic/Latino gay and bisexual men, while decreasing by 14% among their White counterparts (Centers for Disease Control and Prevention 2019). One potential explanation for this disparity is the emphasis placed on disease prevention and risk reduction when developing interventions for Latino sexual minority men, with less consideration given to maintaining the enjoyment of their sexual experiences. Relative to other cultural groups, Latino sexual minority men have received less attention in the literature and warrant further research given the growing Latinx community in the USA (Barnett et al. 2019; Flores, Lopez and Krogstad 2019). More nuanced understanding of behavioural practices and psychosocial influences of Latino men is needed to develop programmes and interventions that align with their sexual values and preferences, promote sexual health, and preserve positive aspects of sexuality. This study sought to describe predominant behavioural profiles among Latino immigrant sexual minority men and to examine the relevance of resource-based power and sexual pleasure.

Collectively, the literature on sexual positioning and behaviours among sexual minority men highlights the importance of both sexual pleasure and power (Kippax and Smith 2001; Clark et al. 2013; Carballo-Diéguez et al. 2004; Johns et al. 2012; Xu and Zheng 2018; Ricks, Crosby and Mena 2018; Kiguwa 2015; Hoppe 2011). In a recent narrative review of the sexual positioning practices of sexual minority men, Dangerfield and colleagues identified power, broadly defined, to be one of the key contextual factors that influenced sexual positioning, identity, and practices (Dangerfield et al. 2017). Due to the multifaceted nature of power as a multi-level and socially constructed characteristic (Kaufman et al. 2008; Wingood and DiClemente 2002), researchers have operationalised power as economic capital using resource-based models of power (Perry et al. 2016a; 2016b). Such models posit that relational power is organised by bases, processes, and outcomes where bases, or attributes of individual partners, provide capital in a relationship (i.e. income and education), which provides partners with the capability of expressing such capital via power processes (i.e. communication), which then results in power outcomes, or one partner ‘winning’ the power struggle (Blader and Chen 2012; Cromwell and Olson 1975). In a sample of gay male couples in San Francisco, individual income positively predicted decision-making power among dyads (Perry et al. 2016b). Additionally, resource-based power appears to influence sexual role identities and sexual behaviours among sexual minority men, as a previous analysis of older Australian gay men found “versatile” men, or those willing to engage in both insertive and receptive anal intercourse, tended to have a higher income (Lyons, Pitts and Grierson 2013). These studies suggest that men use information about their partner’s resources to inform decisions around sexual role delegation and the behaviours within encounters that have underlying assumptions of power.

In addition to power, the pleasure associated with engaging in various sexual behaviours is likely to play a significant role. In fact, several investigators have made the argument that power and pleasure are not mutually exclusive, but rather control over pleasure is often a form of exerting power (Dangerfield et al. 2017; Hoppe 2011; Kiguwa 2015). Previous qualitative analyses have found that men who identity as bottoms and prefer the receptive position for anal intercourse feel empowered due to their control over their insertive partner’s stimulation and pleasure (Hoppe 2011; Kiguwa 2015). Additionally, a multivariate path model identified two direct predictors of sexual role identity including whether one finds bottoming pleasurable and one’s desire for sexual control, suggesting pleasure and power may play unique roles in the sexual behaviours sexual minority men practice (Moskowitz and Roloff 2017).

Similar findings have been observed among Latino sexual minority men, though the literature has been partial towards exploring the influence of cultural factors such as machismo, or the socially constructed gender roles for Latino men, on sexual positioning and behaviours (Clark et al. 2013; Carballo-Diéguez et al. 2004; Estrada et al. 2011; Nakamura and Zea 2010). The review by Dangerfield and colleagues (2017) also discusses the importance of cultural contexts in shaping sexual behaviours and the included studies on Latino men emphasised masculinity stereotypes. However, this approach has been critiqued for being highly gendered and decontextualised (Carrillo and Fontdevila 2014). Across cultures, there exist similarities in sexual positioning among sexual minority men globally with consistent thematic preferences for insertive, receptive or versatile positioning (Carballo-Diéguez et al. 2004; Dangerfield et al. 2017; Xu and Zheng 2018; Moskowitz and Roloff 2017; Jeffries 2009). As such, resource-based power and pleasure may influence sexual positioning among Latino men similarly to what has been observed with men internationally. Thus, further research is needed to obtain a more specific understanding of the sexual experiences of Latino men and the potential implications of resource-based power and sexual pleasure for sexual behaviours.

In addition to sexual positioning, other sexual practices may reflect power dynamics and preferences for pleasure such as oral sex and fingering (Dangerfield et al. 2017; 2018). Previous research has explored co-occurring sexual behaviours across sexual role identities indicative of sexual scripts among sexual minority men (Moskowitz, Rieger and Roloff 2008). Top identified participants tended to receive oral sex and perform anal stimulation with the opposite true for bottom identified participants. Versatile participants reported fairly similar rates of performing and receiving the various sexual behaviours (Moskowitz, Rieger and Roloff 2008). Resource-based power and pleasure may be uniquely related to versatility given factors like relationship dynamics and sexual reciprocity (Dangerfield et al. 2017).

In this study, we examined associations of resource-based power differentials between partners and pleasure derived from receptive and insertive anal intercourse with event-level sexual behaviour reported by Latino sexual minority men. To address the complexity of sexual behaviour, Social Action Theory (Ewart 1991) informed the original study (Zea et al. 2009) and selection of variables for this paper given sexual behaviour is a social interaction that is influenced by the surrounding context. Aim one was to identify predominant behavioural profiles (classes) based on participants’ reported engagement in sexual activities during their most recent dyadic male-partnered sexual event using latent class analysis. Profile formation was based on the following sexual activities: performing and receiving oral sex, fingering, and rimming as well as engaging in receptive and insertive anal intercourse. Aim two was to explore whether behavioural class membership was associated with perceptions of resource-based power differentials between partners and/or self-reported sexual pleasure.

Materials and Methods

Participants

Participants were cisgender sexual minority men born in Brazil, Colombia, or the Dominican Republic who had migrated to USA and resided in New York City. Immigrants from these three countries were selected because research on Latinos has traditionally focused on Mexican, Puerto Rican and Cuban participants (Durand and Massey 2010) whereas newer immigrants have been underrepresented in research.

Participants were recruited using purposive sampling techniques at gay venues and Latino cultural events throughout the New York City metropolitan area (2005–2006). To be eligible, individuals had to be at least 18 years of age; male-identified; born in Brazil, Colombia or the Dominican Republic; residing in New York City; and had engaged in sexual activity with men in the last six months. A fuller description of the sample, sampling techniques, and eligibility criteria can be found in Zea et al. 2009.

Analyses utilised data from the participants’ most recent sexual encounter with one male partner. From the original 482 participants sampled, nine men were excluded because their most recent sexual encounters were not with a male and 21 other men were excluded because their most recent encounters were with two or more individuals, yielding a final sample of 452 men.

Measures

Surveys were administered using computer-assisted self-interview technology with audio enhancement (A-CASI) in Spanish, Portuguese or English, depending on participants’ preference. The cross-sectional survey included items assessing sexual behaviour, perceived resource-based power, sexual pleasure and background characteristics.

Sexual Behaviour

Participants were provided with a series of statements regarding sexual activities that may have occurred during their most recent sexual encounter with one male partner and were asked to indicate all that transpired. Sexual behaviours included performing oral sex, receiving oral sex, performing fingering (manual-anal stimulation), receiving fingering, performing rimming (oral-anal stimulation), receiving rimming, engaging in receptive anal intercourse and engaging in insertive anal intercourse. Each behaviour was framed as an action performed by or received by the participant. For example, one item stated, “You inserted your finger into your sexual partner’s anus”, and the complementary item stated, “Your sexual partner inserted his finger into your anus.”

Perceptions of Resource-Based Power

Participants were asked about their perception of their sexual partner’s income and education level relative to their own as a means of assessing access to resources to estimate power differentials. For example, participants were asked, “What do you think this person’s income was?” and response options included “Much higher than mine” (−1), “More or less the same as mine” (0), or “Much lower than mine” (+1). A similar question was used to assess education level. These power-relevant characteristics were informed by resource models of power (e.g. income) (Cromwell and Olson 1975; Perry et al. 2016a; 2016b), as well as specific variables that are likely relevant to sexual minority men dyads (e.g. education). These two items were combined into a proxy for resource-based power using a decisional matrix where scores were awarded based on responses to items measuring income and education differentials (See online supplemental material Figure 1). Socioeconomic status (SES) differentials ranged from −2 to 2, which were collapsed into three categories: participant perceived himself to have more power than his partner (SES differential > 0), participant perceived himself to have less power (SES differential < 0), or participant perceived equal power relative to his partner (SES differential = 0).

Sexual Pleasure

Participants indicated the amount of pleasure derived from insertive and receptive anal intercourse with and without a condom. Questions were asked individually (e.g. “How pleasurable do you find anal sex with a condom: As a top (your penis in his anus)?”), with five response options ranging from “Not pleasurable at all” to “Extremely pleasurable.” Pleasure for each of the four behaviours was coded dichotomously, with “A little bit pleasurable” or higher coded as “pleasurable” (1) and “not pleasurable at all” coded as “not pleasurable” (0).

Background Characteristics

Covariates included participant age, country of birth, HIV status (i.e. positive, negative, unknown), and sexual role identity (i.e. active (top), passive (bottom), versatile, and no label). The type of partner with whom the most recent sexual event was shared (i.e. main or not main) was also included as well as the participant’s assumption of their partner’s HIV status.

Statistical Analysis

Latent class analysis, a statistical technique that identifies latent (unmeasured) classes based on responses to categorical items, was utilised as a person-centred approach to identify classes of Latino sexual minority men based on their sexual behaviours. To this end, latent class analysis used latent class and item-response probabilities to place each participant into a mutually exclusive and exhaustive class (Lanza et al. 2007). Data analyses were conducted using a PROC LCA macro in SAS 9.4 from the Methodology Center at Penn State (Lanza et al. 2015). Latent class analysis has successfully been used previously when studying sexual behaviours of sexual minority men (Calabrese et al. 2015; Card et al. 2018; Chan et al. 2015; Rice, Norris Turner and Lanza 2017; Vasilenko et al. 2019) and is a technique that allows analysis of distinct behavioural profiles of Latino men.

Aim one was tested by identifying a baseline model of behavioural profiles that adequately organised the data based on eight sexual behaviours (see online supplemental material Figure 2). First, we assessed various models that emerged from the data, selecting those that were well substantiated (i.e. maximum likelihood solution >60%) (Collins and Lanza 2010). Subsequently, each model was evaluated by its goodness-of-fit statistics including the G-squared fit statistic, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC) to determine the final number of behavioural profiles (classes) (see Collins and Lanza 2010 for a description of these statistics).

Aim two was tested with latent class regression modelling adjusting for relevant covariates, with resource-based power and sexual pleasure considered as correlates of latent class membership (Collins and Lanza 2010). Differences in the likelihood of class membership were evaluated by comparing each class to every other class separately at each level of the power and pleasure correlates.

Results

Descriptive Statistics

Participants (n=452) included Brazilian (n=138), Colombian (n=158), and Dominican (n=156) immigrant cisgender sexual minority men residing in the USA with an average age of 36 years (SD = 9.5) and have lived in the USA for an average of 15 years (SD = 10.5). The majority of the sample (>65%) earned less than $20,000 yearly although nearly 70% of the sample had attended college. Regarding sexual role identities, 23% of the sample self-identified as active or tops, 16% as passive or bottoms, and 48% as versatile; the remaining participants (13%) indicated they did not use labels. The self-reported HIV prevalence was 25.4%. The majority of the sample (54%) perceived less power relative to their partner, followed by 27% of the sample who perceived equal power, and 19% of the sample perceived more power. Participants reported pleasure from receptive anal intercourse with condoms (48%), receptive anal intercourse without condoms (45.6%), insertive anal intercourse with condoms (56.6%), and insertive anal intercourse without condoms (54.2%). The majority of the sexual encounters (66.8%) were not with a main partner (see Table 1).

Table 1.

Sociodemographic and partner relationship characteristics of participants (N=452)

Characteristic	Brazilian (N=138)	Colombian (N=158)	Dominican (N=156)	Total sample (N=452)
Age (Mean (SD))	37.30 (8.93)	37.99 (9.64)	33.69 (9.25)	36.29 (9.47)
Years in US (Mean (SD))	13.86 (11.04)	12.69 (9.46)	17.21 (10.47)	14.60 (10.47)
Income (monthly) (%)
<$400	13.0	17.1	21.8	17.5
$401–800	18.9	25.9	26.9	24.1
$801–1600	23.9	29.8	25.6	26.6
$1601–2400	18.1	18.3	10.9	15.7
$2401–3200	13.8	5.1	9.6	9.3
$3201–4000	6.5	1.9	2.6	3.5
>$4001	5.8	1.9	2.6	3.3
Education (%)
Grammar/Elementary	0.7	7.0	6.4	4.9
Some high school	6.5	10.1	10.3	9.0
Completed high school	17.4	12.7	10.3	13.3
Trade or vocational school	8.0	1.9	2.5	4.0
Some college	24.6	21.5	35.3	27.2
Completed college	31.9	28.5	21.1	27.0
Some graduate school	4.4	8.2	5.1	6.0
Graduate degree	6.5	10.1	9.0	8.6
Sexual Role Identity (%)
Active (Top)	25.4	18.3	26.3	23.2
Passive (Bottom)	15.9	19.0	14.1	16.4
Versatile	44.9	51.3	48.1	48.2
Don’t Use Labels	13.8	11.4	11.5	12.2
Perceived Resource-Based Power Differential (%)
Participant perceives more power	15.2	20.2	20.5	18.8
Participant perceives equal power	32.6	28.5	21.2	27.2
Participant perceives less power	52.2	51.3	58.3	54.0
Sexual Pleasure from Anal Sex (%)
Pleasure from receptive anal intercourse with condom	52.2	44.3	48.1	48.0
Pleasure from receptive anal intercourse without condom	49.3	39.9	48.1	45.6
Pleasure from insertive anal intercourse with condom	63.8	51.3	55.8	56.6
Pleasure from insertive anal intercourse without condom	55.8	52.5	54.5	54.2
HIV Status (%)
Positive	21.0	30.4	24.4	25.4
Negative	66.7	57.0	66.0	63.1
Unknown	12.3	12.6	9.6	11.5
Partner Type (%)
Main	32.6	33.5	33.3	33.2
Not main	67.4	66.5	66.7	66.8

Four-Class Model

A four-class model was identified using established criteria including fit criteria, interpretability and class separation (see Table 2) (Collins and Lanza 2010). The set of parameters displayed in Table 3 indicate the proportion of participants in the whole sample who endorsed each sexual behaviour. Additionally, Table 3 includes the item-response probabilities for each behaviour by latent class, which represents the conditional probability of reporting each behaviour given membership in a latent class. The cut-off of 0.50 was used to determine which sexual behaviours were characteristic of the latent class, with probabilities closer to one indicating a greater likelihood that individuals in that class engaged in the behaviour (Collins and Lanza 2010).

Table 2.

Fit statistics for competing latent class models of sexual behaviour profiles with one to five classes (N=452)

Number of Classes	G2*	df	Log-Likelihood	AIC	BIC	BLTR	Entropy R2	Solution %
1	771.80	247	−2332.80	787.80	820.71	-	1.00	100
2	519.90	238	−2206.85	553.90	623.83	0.01	0.70	57
3	363.23	229	−2128.51	415.23	522.18	0.01	0.71	72
4	281.17	220	−2087.48	351.17	495.15	0.01	0.73	86
5	241.58	211	−2067.69	329.58	510.58	0.01	0.78	17

Note. G²

^*= Likelihood-ratio chi-square statistic; AIC = Akaike’s Information Criterion; BIC = Bayesian Information Criterion; BLTR = Bootstrap Likelihood Ratio Test. Solution % is the percentage of times solution was selected out of 100 random sets of starting values.

Table 3.

Item-response probabilities for four behavioural profiles reported by sexual minority men (N=452) at last sexual event

Behavioural profile (class)
		Class 1 Behaviourally Insertive	Class 2 Behaviourally Versatile	Class 3 Behaviourally Receptive	Class 4 Limited Penetrative Behaviour
Class Prevalence		14.2%	25.9%	29.2%	30.7%
Behaviour	Proportion in the Sample	Item-Response Probabilities
Performed Oral Sex	0.74	0.58	0.91	0.98	0.43
Received Oral Sex	0.71	0.91	0.96	0.51	0.60
Performed Fingering	0.39	0.87	0.66	0.07	0.25
Received Fingering	0.34	0.00	0.55	0.52	0.13
Performed Rimming	0.33	0.77	0.61	0.17	0.05
Received Rimming	0.37	0.09	0.76	0.40	0.14
Insertive Anal Intercourse	0.44	0.93	0.77	0.03	0.33
Receptive Anal Intercourse	0.47	0.00	0.62	0.90	0.15

Note. Proportions are based on N endorsing each behaviour. Item-response probabilities > 0.50 appear in bold to indicate a behaviour endorsed by the majority of the class.

Class 1: Behaviourally Insertive

Latent class one (14.2% of the sample) was characterised by receiving oral sex (0.91) and engaging in insertive anal intercourse (0.93) as well as performing oral (0.58), rimming (0.77), and fingering (0.87). This class was labelled, “behaviourally insertive” to reflect the class’ tendency to follow an active/insertive behavioural script of engaging in insertive anal intercourse, performing fingering and rimming, and indicating greater probability of receiving oral sex than performing it.

Class 2: Behaviourally Versatile

Latent class two (25.9% of the sample) was characterised by a moderate to high probability of engaging in all behaviours (0.55–0.96), particularly performing and receiving oral sex (0.91 and 0.96, respectively). This latent class was labelled, “behaviourally versatile” to reflect their adaptable sexual positioning and versatility in sexual behaviours within the same sexual encounter.

Class 3: Behaviourally Receptive

Latent class three (29.2% of the sample) was characterised by performing oral sex (0.98) and engaging in receptive anal intercourse (0.90), as well as receiving oral sex (0.51) and receiving fingering (0.52). This latent class was labelled, “behaviourally receptive” to reflect their tendency to follow a passive/receptive behavioural script of engaging in receptive anal intercourse, receiving fingering and rimming, and indicating greater probability of performing than receiving oral sex.

Class 4: Limited Penetrative Behaviour

Latent class four (30.7% of the sample) was characterised only by receiving oral sex (0.60) with low probabilities of other behaviours. This latent class was labelled, “limited penetrative behaviour” to reflect the class’ tendency to only receive oral sex and not engage in any other penetrative behaviours.

Class Association with Identified Covariates

Age, sexual role identity, and partner type were associated with class membership (see online supplemental material Table 1). Sexual encounters with a main partner and encounters had by younger participants had higher odds of belonging in the behaviourally versatile class. Sexual role identities (i.e. active/Top, passive/Bottom, and versatile) aligned with their respective classes of behaviourally insertive, behaviourally receptive, and behaviourally versatile. Country of birth, HIV status and participant’s assumption of their partner’s HIV status were not statistically associated with class membership.

Class Association with Perceptions of Resource-Based Power

The three levels of perceived resource-based power (i.e. participant perceived more, participant perceived equal, participant perceived less) were entered into the latent class analysis as correlates in separate regression models following the protocol established by Collins and Lanza (2010) for conducting a latent class analysis with covariates. Participants’ perception of their resource-based power relative to their partner’s was not associated with class membership at any level of the construct. Whether the participants perceived themselves to possess more, equal or less resource-based power than their partner, their odds of belonging to each class was not different (see Table 4). The same pattern was observed when sexual pleasure and relevant covariates (age and partner type) were adjusted for in the models.

Table 4.

Perceptions of resource-based power and sexual pleasure as correlates of class membership (N=452)

	Sexual behaviour profile comparisons
	Behav. Recep. vs. Behav. Insert.		Behav. Recep. vs. Limited Pen. Behav.		Behav. Recep. vs. Behav. Versatile		Behav. Versatile vs. Behav. Insert.		Behav. Versatile vs. Limited Pen. Behav.		Behav. Insert. vs. Limited Pen. Behav.
	Class 3 vs. 1		Class 3 vs. 4		Class 3 vs. 2		Class 2 vs. 1		Class 2 vs. 4		Class 1 vs. 4
	AORa	95% CI	AORa	95% CI	AORa	95% CI	AORa	95% CI	AORa	95% CI	AORa	95% CI
Resource-Based Power
Participant perceived more	0.73	(0.39, 1.34)	1.08	(0.63, 1.87)	1.03	(0.59, 1.79)	0.70	(0.38, 1.31)	1.05	(0.61, 1.83)	1.49	(0.81, 2.77)
Participant perceived equal	0.99	(0.55, 1.78)	1.04	(0.64, 1.67)	0.74	(0.45, 1.19)	1.34	(0.74, 2.41)	1.41	(0.88, 2.25)	1.05	(0.59, 1.88)
Participant perceived less	1.26	(0.75, 2.14)	0.91	(0.60, 1.41)	1.26	(0.81, 1.97)	1.00	(0.59, 1.71)	0.73	(0.47, 1.12)	0.72	(0.43, 1.21)
Pleasure
Receptive anal intercourse with condom	7.79**	(4.00, 15.15)	4.50**	(2.64, 7.67)	3.77**	(2.18, 6.53)	2.06**	(1.02, 4.17)	1.19	(0.69, 2.06)	0.58	(0.29, 1.16)
Receptive anal intercourse no condom	5.25**	(2.77, 9.97)	4.93**	(2.93, 8.32)	2.90**	(1.67, 5.04)	1.81	(0.93, 3.51)	1.70**	(1.01, 2.86)	0.94	(0.49, 1.79)
Insertive anal intercourse with condom	0.56	(0.31, 1.04)	0.82	(0.50, 1.33)	0.87	(0.53, 1.46)	0.65	(0.33, 1.25)	0.93	(0.56, 1.56)	1.44	(0.76, 2.73)
Insertive anal intercourse no condom	0.36*	(0.18, 0.73)	1.33	(0.83, 2.14)	0.60	(0.34, 1.00)	0.61	(0.29, 1.28)	2.22*	(1.32, 3.71)	3.66*	(1.80, 7.42)

^*p < .01

^**p < .001

^aAdjusted odds ratios. Power analyses ran each level of power (i.e. participant perceived more, participant perceived equal, participant perceived less) in separate regression models, adjusting for significant pleasure variables (i.e. receptive anal intercourse with condom, receptive anal intercourse no condom, insertive anal intercourse no condom) and significant background variables (i.e. age, main partner). Pleasure analyses ran each level of pleasure (i.e. receptive anal intercourse with condom, receptive anal intercourse no condom, insertive anal intercourse with condom, insertive anal intercourse no condom) in separate regression models, adjusting for significant background variables (i.e. age, main partner). Analyses were repeated adjusting for all levels of resource-based power and yielded the same pattern of results (not shown).

Class Association with Sexual Pleasure

We also examined whether participants’ degree of pleasure derived from receptive and insertive anal intercourse with and without condoms was associated with class membership, using the same methodology as the perceived resource-based power construct in separate regression models (Collins and Lanza 2010). For specific adjusted odds ratios (AORs), see Table 4.

For the behaviourally receptive class, the odds of endorsing pleasure from receptive anal intercourse with a condom were multiplied between 4 and 8 times and the odds of endorsing pleasure from receptive anal intercourse without a condom were multiplied between 3 and 5 times relative to other classes. Additionally, for the behaviourally versatile class, the odds of endorsing pleasure from receptive anal intercourse with and without a condom were multiplied between 1 and 2 times relative to other classes. The odds of endorsing pleasure from insertive anal intercourse with a condom did not differ between classes.

The odds of endorsing pleasure from insertive anal intercourse without a condom were multiplied 3 times for those in the behaviourally insertive class and 2 times for those in the behaviourally versatile class relative to those in the limited penetrative behaviour class. For the behaviourally insertive class, the odds of endorsing pleasure from insertive anal intercourse without a condom were multiplied nearly 3 times relative to those in the behaviourally receptive class (see Table 4).

These same patterns were observed when resource-based power and relevant covariates (age and partner type) were adjusted for in the models. Additionally, these patterns were maintained when the pleasure covariate was run as a continuous variable, so the results are presented with the pleasure covariate dichotomised to optimise parsimony and interpretability, an approach used by others (Majeed et al. 2015; Dayton and Macready 2002).

Discussion

Collectively, our findings suggest that sexual pleasure is an influential factor with respect to sexual behaviour and anal sex positioning among Latino sexual minority men, irrespective of power dynamics. These findings also demonstrate the utility of latent class analysis in assessing sexual behaviours of sexual minority men, which corroborates previous applications of latent class analysis with sexual behaviour profiles (Calabrese et al. 2015; Card et al. 2018; Chan et al. 2015; Rice, Norris Turner and Lanza 2017; Vasilenko et al. 2019).

Behavioural Scripts for Latent Classes

Among the behaviourally insertive and behaviourally receptive classes, sexual behaviours appeared congruent with previously identified sexual scripts among gay men (Moskowitz, Rieger and Roloff 2008). In line with these sexual scripts, behaviourally insertive men perform rimming and fingering on their partner and behaviourally receptive men perform oral sex on their partner as foreplay to anal sex. The decreased probability of behaviourally insertive men to receive anal stimulation and behaviourally receptive men to perform anal stimulation suggests that these behaviours are informed by sexual scripts that have both practical and erotic purposes. The behaviourally versatile class demonstrated its own unique sexual script with flexibility in sexual positioning and reciprocity in sexual behaviours. The versatility in their behaviours appear to reject the dichotomous top/bottom scripts and display a greater interest in receiving and performing several sexual behaviours. The limited penetrative behaviour class did not appear to follow any of the previously defined sexual scripts and only receptive oral sex was characteristic of the class (Moskowitz, Rieger and Roloff 2008). While these classes are based on the most recent sexual encounter and behaviours may vary across encounters, these scripts align with sexual role identities suggesting they may transcend encounters. These classes may also be representative of the context of the encounter including factors such as location and timing, which may limit the practicality of certain sexual behaviours.

Pleasure as a Correlate of Sexual Behaviours

Our findings highlight the importance of sexual pleasure in determining the type of sexual activity within encounters, beyond that of resource-based power. Sexual pleasure contributes an additional nuance to the understanding of sexual positioning practices among Latino sexual minority men, which have been largely focused on masculinity stereotypes (Dangerfield et al. 2017). In general, sexual positioning for anal sex at the most recent sexual encounters was aligned with perceptions of pleasure derived from receptive and insertive anal intercourse. A hierarchy of pleasure manifested in the data where insertive men had higher odds of endorsing pleasure from insertive anal intercourse, followed by versatile men, and then receptive men, with the opposite ordering true for receptive anal intercourse. This pattern of pleasure was observed regardless of condom use for receptive anal intercourse, but for insertive anal intercourse, pleasure was only associated with class membership in the context of condomless insertive anal intercourse. These findings may suggest that sexual minority men’s perceptions of pleasure for anal sex is partly based on the presence or absence of condoms. Behavioural versatility was associated with younger age and encounters with a main partner. This link suggests greater reciprocity and versatility in sexual behaviours with an individual with whom there is more intimacy as well as within dyads consisting of younger men.

Analyses provided evidence for the congruency between sexual role identity and actual sexual behaviours at the most recent or second to last encounters with one male partner. Sexual role identity (i.e. active/top, passive/bottom, versatile), behaviour, and pleasure were largely congruent such that men in the sample engaged in behaviour that was consistent with the sexual script associated with their role as well as behaviours they deemed pleasurable. However, analyses also suggested that equating identity with behaviour would be erroneous as there were individuals within insertive and receptive classes who demonstrated versatility in anal sex positioning without identifying as versatile, which has been observed previously among Latino sexual minority men (Jeffries 2009). Additionally, nearly one third of the participants were associated with another behavioural class that did not specifically align with any of the three identities, instead characterised primarily by receptive oral sex and the absence of other penetrative behaviours. These findings highlight the danger of making assumptions about behaviour based on self-labels and prompts further inquiry into the relationship between identity and pleasure. Also, the general lack of standardised ways of measuring sexual positioning and other sexual behaviours complicates studying the sexual behaviour of sexual minority men. Consistent operationalisation of sexual positioning questions across studies may help the conceptual clarity of future investigations and capacity to draw comparisons among them.

Limitations and Future Directions

There are limitations to the measurement of power differentials in this study which were based on the participant’s perceptions of their partner’s income and education relative to their own. This estimate is rather general and may not be objectively accurate, particularly for those encounters that were shared with new or less familiar partners. It is possible that other dimensions of power within a partnership are more relevant to sexual behaviours than resource differentials. Future studies may benefit from expanding the construct of ‘power’ to include measures of racial differences between partners, perceived relative attractiveness, erotic capital, and manifestations of control over pleasure to advance our understanding of the innerworkings of dyadic partnerships among Latino sexual minority men (Dworkin et al. 2017).

The encounter-specific data was advantageous in allowing partner and contextual associations with behaviour to be studied within a single sexual encounter rather than relying on aggregate data. The drawback with using data from the most recent sexual encounters is that it does not allow for assumptions about the representativeness of their behaviours across encounters. Additionally, cross-sectional data do not allow for testing directionality. Although perceptions of resource-based power and sexual pleasure could be conceptualised as predictors in the models presented, the current analyses do not include temporal data to confirm the sequence of these events to deduce causality. Future studies may be better able to answer these questions utilising multiple timepoints for data collection. An additional advantage of the encounter-level data was the ability to learn about specific partner types, which is a well-established factor that influences sexual positioning among sexual minority men (Dangerfield et al. 2018; Johns et al. 2012; Lyons et al. 2011). Given the increased odds of behavioural versatility with main partners, additional studies may seek to better understand the differences in relational dynamics (i.e. power and pleasure) between partner types such as main partners and casual partners.

Studying a sample of Latino immigrant sexual minority men from three Latin American countries recruited in New York City between 2005–2006 offers critical insight about the implications of power and pleasure for sexual behaviour. Replication of this study with Latino immigrant sexual minority men in the future could demonstrate the persistence or evolution of these associations in an era in which a wider array of HIV prevention strategies are available, such as pre-exposure prophylaxis (PrEP) and undetectable = untransmittable (U=U), neither of which existed at the time of data collection.

Implications

Understanding the role of pleasure in the sexual practices of Latino sexual minority men has large implications for future research, community-based interventions, and sexual health curricula. Programmes and interventions that do not consider the importance of pleasure are less likely to be successful, particularly when their aims include changing sexual behaviour. A previous analysis of Latino immigrant sexual minority men found condom use to decline as a function of perceived pleasure loss associated with condom use (Calabrese et al. 2012). For men who perceive condom use to diminish sexual pleasure, sexual health interventions incorporating other strategies beyond condoms, like PrEP, post-exposure prophylaxis (PEP), and U=U, are likely to be a more realistic and effective intervention than those focused on condoms alone (Carballo-Diéguez et al. 2011). However, biomedical interventions may be less accessible to marginalised communities and those with low resources, emphasising the importance of policy and funding to ensure equitable opportunities. Public health interventions should be tailored to ensure the appropriate communities are being reached, resources are accessible and allocated effectively, and sexual safety products (e.g. condoms, dental dams, lubricants) are manufactured to maintain pleasure (Vasilenko et al. 2019). In tandem, the importance of the behavioural components associated with biomedical interventions should be emphasised to maximise effectiveness in order to allow pleasure to remain a priority alongside safety.

Lastly, future investigations and interventions should attempt to integrate positive aspects of sexuality within the disease-prevention lens commonly applied to the sexual behaviours of sexual minority men. A disease-prevention perspective of sexual health serves its purpose in reducing transmission of HIV and STIs, but also has the potential to pathologise sexual behaviours deemed high risk. This may be especially true for Latino sexual minority men given the disproportionate number of new infections they face compared with their White counterparts (Centers for Disease Control and Prevention 2019). Until recently, prevention strategies for HIV transmission among sexual minority men were based on increasing condom use and HIV testing, particularly among Black and Latino men (Cohall et al. 2010; Young et al. 2014; Warren et al. 2008). Findings from our analyses suggest pleasure is a decisive factor in navigating sexual encounters for sexual minority men and should continue to be an important consideration for HIV prevention efforts. Other methods of reducing HIV transmission, within the spirit of a sex-affirming framework, could be to suggest that sexual health educators and medical providers inform clients of the full range of safe sex practices available, including PrEP, PEP, and U=U for those living with HIV, which may be more acceptable options to men for whom condoms interfere with pleasure (Keene et al. 2020). Integrating a prevention lens with a sex-affirming framework will maintain the safety of at-risk communities without pathologising their sexuality.