Debe cuidarse en la calle: Normative influences on condom use among the steady male partners of female sex workers in the Dominican Republic

Abstract

Encouragement to use condoms reflects the injunctive norm, or idea that you should use condoms. In our previous research with the regular male partners of female sex workers in the Dominican Republic, we found that encouragement to use condoms with female sex workers from individuals in their personal social networks was not directly associated with condom use. In the current study, we used qualitative interviews to further explore the influence of social network norms on men's sexual risk behaviours. We interviewed eleven steady male partners of female sex workers; participants completed two interviews to achieve greater depth. We analysed data using analytic summaries and systematic thematic coding. All men perceived that the prevailing injunctive norm was that they should use condoms with sex workers. Men received encouragement to use condoms but did not articulate a link between this encouragement and condom use. Additionally, men who did not use condoms lied to their friends to avoid social sanction. Findings highlight that the influence of a pro-condom injunctive norm is not always health promoting and can even be negative. HIV prevention efforts seeking to promote condom use should address the alignment between injunctive and descriptive norms to strengthen their collective influence on behaviour.

Introduction

The Dominican Republic (DR) has a long history of innovative HIV prevention efforts in the female sex industry. These interventions, using peer education and outreach work, community mobilisation, and environmental–structural approaches, have achieved increases in condom use and declines in sexually transmitted infections among female sex workers (Kerrigan et al 2001., Kerrigan et al., 2006). Nevertheless, compared to the adult prevalence of approximately 0.8%, female sex workers are still disproportionately affected by HIV in the DR. The most recent estimates of HIV prevalence among female sex workers in four diverse geographic areas across the country, obtained using respondent driven sampling, indicate a wide range from 3.3% to 8.4% (Consejo Presidencial del SIDA (COPRESIDA) 2008).

The paying and non-paying male sex partners of female sex workers have not been included in HIV surveillance efforts despite their potential vulnerability to HIV infection and potential role in ongoing transmission. These men have been referred to as a bridge population in the epidemiological literature due to their role in sexual networks that include a range of concurrent partners with varying levels of condom use (Lowndes et al., 2007). Dominican women working in the sex industry have articulated the need to improve the involvement of men, especially their steady partners, in negotiating HIV protective behaviours within and beyond sex establishments (Kerrigan et al., 2003).

Male patronage of sex work has frequently been described as a social activity. Vanlandingham et al. (1998) qualitatively examined factors influencing use of sex workers among married men in Thailand. In this setting, routine social gatherings and special events created opportunities for peer groups to go to sex work establishments. Yang et al. (2010a) also found that visiting sex work establishments was also a group activity in China, where men indicated that going with a group of 3-to-5 friends made them feel safer and also helped in negotiating prices. In this study, men also indicated that taking friends or colleagues out for the night could lead to greater social and professional bonding.

By framing involvement with sex work as a social activity among men, the question arises: how does the social nature of seeking sex influence sexual risk behaviours, in particular condom use? A few studies have assessed the influence of social norms on condom use among male sex partners of female sex workers, including both descriptive norms (i.e. perceptions of whether peers use condoms) and injunctive norms (i.e. perceptions of whether peers think you should use condoms) (Cialdini & Reno, 1990; Lapinski & Rimal, 2005). In one of the first studies to explore this question, Ford et al. (2002) found that both perceived condom use and advice to use condoms (an indicator of the injunctive norm) by male friends were significantly associated with reported condom use among male clients of female sex workers in Bali. More recently, Yang et al. (2010b) conducted a more nuanced analysis and found that both perceived condom use and encouragement to use condoms were significantly associated with reported condom use among men who visited sex workers with friends, but not those who visited sex workers alone. Both of these studies focused on male friends as the reference group for normative influence. In our previous quantitative research with steady male partners in the DR, we found that while most (70.4%) received encouragement to use condoms with sex workers from all of their personal network contacts, including male friends and family, female family, female partners and other acquaintances, encouragement was not significantly associated with reported condom use (Barrington et al., 2009). Men who received encouragement from their personal networks were, however, significantly more likely to perceive that all of their male social network contacts used condoms consistently, suggesting that encouragement may fuel the perception of friends' behaviours.

With the goal of improving understanding of normative influences on condom use and adding more depth to our quantitative findings, we qualitatively explored how steady male partners of female sex workers in the DR describe normative influences on their sexual behaviour. In this paper, we address the following aims: 1) Describe male partners' perceptions of the prevailing norms with regard to condom use, 2) Examine how the exchange of advice from personal network contacts reinforces norms, and 3) Assess the influence of personal network norms and advice on condom use.

Methods

Study setting

This study was conducted in La Romana, a town with a population of approximately 200,000 people on the Southeastern coast of the DR. La Romana has over 100 establishments where sex work occurs, ranging from small pool halls with only one or two women to large bars with over twenty women. Sex work establishments thrive in La Romana in part because it is one of the most economically productive cities in the DR as host to the largest sugar company in the country, a growing tourist industry, and free trade manufacturing zones. The majority of male paying and non-paying partners of female sex workers in La Romana are Dominican; when this research was conducted, we also observed foreign tourists from cruise ships and nearby resorts coming to La Romana in pursuit of sex.

Sample and recruitment

Prior to recruitment, the first author conducted a census of all sex work establishments, direct observations of the establishments and interviews with managers/owners (Barrington 2007). This preliminary work was used to improve understanding of the context of sex work in La Romana and develop rapport with key stakeholders in the sex industry. Study participants were recruited from seven different establishments, ranging from informal centro cerveceros (beer halls), to traditional cabarets, to high-end brothels called casas de cita.

Men were eligible if they were a steady partner of a sex worker, which was defined as having had sex at least three times in the previous six months, and were at least 18 years of age. We purposively sampled steady partners as we knew from prior research that condom use is lowest in these partnerships and we wanted to improve understanding of how to engage with steady partners about this topic (Kerrigan et al., 2003; Murray et al., 2007). Men who were known to be steady partners of sex workers were approached by a female peer educator and invited to participate in the study. Eligibility was confirmed by the first author and eligible participants were then invited to engage in an informed consent process. The Institutional Review Boards of the Johns Hopkins University, University of North Carolina at Chapel Hill and the Instituto Dermatalógico y Cirugía de Piel Dr. Huberto Bogaert Diaz in Santo Domingo approved this study.

Data collection

Interviews were carried out with consenting individuals at a mutually agreed upon location immediately following recruitment or shortly after. All interviews were conducted in Spanish by the first author, a white, female, native English speaker from the USA, who is fluent in Spanish and had lived in the DR for over four years. Informed by a social constructionist approach, we view knowledge and meaning as “subjective, constructed, and based on the shared insights and symbols which are recognised by members of a culture” (Grbich 2007, p 8). We view the data in this study as the product of the social interaction between the first author, who conducted the interviews, and the participants. Our study was motivated by an interest in how the social networks of male partners of female sex workers could be a tool for HIV prevention efforts.

We started each interview by soliciting a personal network inventory of the people the participant spent the most time with and/or trusted the most, which was used as a point of reference for the rest of the interview. We did not put any limits on who could be included in the personal network; most men listed male friends, family members, female partners and workmates. We used a semi-structured guide including questions and probes on the following topics: 1) Communication and social influence within social networks, 2) Perception of the HIV situation in La Romana, 3) Patterns around visiting sex work establishments, and 4) Personal sexual behaviour and condom use. The order of the specific questions was flexible and responded to the direction of the participant. All interviews were audio-recorded and transcribed verbatim. Two interviews each were conducted with nine study participants; one interview was conducted with two study participants with whom the entire interview guide was covered in the first interview. Therefore, we conducted a total of 20 interviews with 11 study participants. Each interview session lasted from one to two hours.

Data analysis

Data were analysed in Spanish. Immediately following the first interview, the first author listened to the recording, expanded upon field notes, and prepared follow-up questions for the second interview. For example, while we were initially interested in condom use as a key behaviour of interest, we found that men discussed often partner selection and relationship dynamics and, in response, we probed on those behaviours as well. We developed an analytic memo summarising the life history and social network composition for each study participant (Sandelowski 1995). These memos sought to highlight the stories shared by participants in the context of their lives. Transcripts were also analysed using a more structured categorising approach whereby text was extracted and coded using a template in Word containing key analysis questions and codes related to the three research questions addressed in this paper (Maxwell & Miller 2008). Due to our findings from our previous quantitative research in this setting, we focused the current analysis on improving understanding of injunctive norms, or what men perceived their social network contacts thought they should do, and how they may influence, or not, sexual risk behaviours.

Findings

Following a description of the study population, we present our findings in three sections. We first describe how men talked about their perceptions of norms around HIV-related behaviours, including condom use, partner selection, and relationship dynamics. We then examine how men exchanged advice related to these behaviours within their social networks. Finally, we address how men described the influence of these norms on their own behaviours.

Description of participants

Study participants ranged in age from 19 to 45 years. Most had an 8^th grade education, though this varied from just a few years of primary school to a college educated engineer. All but one of the study participants had a main cohabitating sex partner. All study participants had at least one child or, in the case of two participants, were expecting a baby with their main partner. Most men were recruited at centro cerveceros (n=7), the most popular form of establishment in the city followed by cabarets (n=3) and a casa de cita (n=1). Most men described their steady female sex worker partner as their amiga or friend, though some had more intimate relationships and considered the woman to be his novia (girlfriend) or mujer (main partner, spouse). Condom use was common at the beginning of relationships but many did not use condoms consistently beyond the initial few encounters. There were two participants who never used condoms with any partners.

Injunctive norms related to condom use and relationship dynamics

Among all study participants, the common perception was that condoms should be used when having sex with sex workers. Participants used phrases like “one has to have condoms”, “you should use condoms”, “you have to put on a condom” to reflect the pervasive nature of the injunctive norm around condom use. Ramón, a 44-year-old regular client of cabarets, had a positive attitude about condoms but did not always use them with steady female sex work partners. He considered condom use with one's wife to be a private matter within the couple; this contrasted his beliefs about condom use with a female sex worker, which he considered a common behaviour,

Look, when you are talking about your wife, those are private things, between man and wife. I can't say whether a man uses [condoms] or not with his wife. When it is about a woman who works at a sex establishment, a woman from a cabaret, from a centro cervecero, from a cafe, you should use it [condom]. Why should you use a condom? So that you don't infect your family.

Ramón's words highlight the main motivation for condom use: protecting the family, as has been described elsewhere (Bandali 2011). Javier, 33, who regularly visited two centro cerveceros in his neighbourhood and had a regular sex worker partner with whom he has not used condoms since they both got HIV tests, indicated that he would be embarrassed to admit he didn't use them,

In my case…I would be ashamed to say that I don't use them [condoms]. Now I would be embarrassed to say, “I was with a woman last night without a condom”, that would be embarrassing.

These examples highlight the salience of the expectation that men should use condoms with their sex worker partners.

Participant's narratives also revealed normative expectations around condom use to prevention pregnancy with a sex worker. Pedro, who was 19 years old and expecting his first child with his teenage girlfriend who was not a sex worker, indiated his belief that condoms should be used to prevent having a baby “en la calle” [in the street], which he did not consider an acceptable behaviour,

As people say, you have to use a condom even with your trusted partner. So, it doesn't matter if I know her and we are friends, always a condom…Not for diseases but for pregnancy. You're pregnant, you're pregnant with her…being the father [of a baby] of a woman who working in a [sex] establishment….noooooo.

Santo, who was 45 and frequented mostly cabarets, became sexually active before the advent of HIV when condoms were used to prevent having children with a sex worker, which would not have been acceptable among his family and friends.

Another perceived norm among a sub-sample of participants was that men should not have sex more than once with a sex worker, and if they did have sex more than once, they should not fijar or develop an ongoing relationship characterised by confianza or trust. This perception is noteworthy given the fact that all men in this study had sex at least three times with a female sex worker partner. For some, having repeated sexual encounters with the same woman was a sign of intimacy, affection and love. In contrast, participants who claimed to always use condoms emphasised the importance of not getting into a committed, trusted relationship with a sex worker. Luis, 23, was first exposed to such normative beliefs from his uncle,

An uncle of mine told me, ‘women from the street, you only have to be with them just one time’ and I have always been guided by that…when I was 15, he told me ‘Be with this woman, but only one time’, and that is what I did.

Luis's uncle's advice reflects how HIV prevention strategies are integrated into the gender socialisation process of young men. While Luis did end up having one long relationship with one woman who worked at his neighbourhood sex establishment, even living with her for a few months, he considered that relationship an exception because she was exceptionally beautiful. José, a 38-year-old regular client of a casa de cita who always used condoms, explained that while he would have sex more than once with the same woman, he did not get too close,

…I tell [women] not to get attached to me. For what reason? Because I can come back [to the establishment] tomorrow and like one [woman], I can come the next day and like another. I don't have regular partners, I have friends and I always use protection.

Taken together, these perceptions reveal certain norms of conduct guiding what might be considered acceptable behaviour with female sex workers, both in terms of condom use as well as relationship dynamics.

Advice about HIV-related behaviours within social networks

The exchange of advice, including but not limited to condom use, was a common event among men interviewed for this study. Several participants used the phrase vive aconsejando or “spend their life giving advice” to describe this frequent exchange of advice. The recurring use of this hyperbolic phrase reflects the central role of advice exchange in social interactions, including about female sex workers, as explained by Luis in reference to his closest friend,

We talk about a lot of crazy stuff. Sometimes when we had something pending that one of us didn't like we'd say, ‘I didn't like this and that, come, let's talk’. Sometimes we have a disagreement because he pawned this woman's jewelry and she came and made a scandal at [sex establishment] and he told me that he had given [the woman] a television and he was going to take it back [to sell]. I told him that does not look good, after a man gives something to a woman, to go and take it back for a no reason.

Through these exchanges of advice, Luis made clear to his friend what he considered to be acceptable behaviour for a man.

Within the ongoing process of giving and receiving advice, there were examples of two different types of advice: advice that reinforces what men already know and advice that introduces new information. One example of advice that reinforces what men already know is the mantra, debe cuidarse en la calle, which literally translates to “You should be careful in the street”. Male partners receive this advice from social network contacts on a daily basis as a reminder of the dangers associated with la calle, including violence, drugs, and the financial costs of spending time in la calle. However, debe cuidarse en la calle is also a euphemism for “you should use condoms”, and the response, yo me cuido a euphemism for “I always use condoms”. The exchange of advice about being careful was an ongoing process mentioned by all men in this study that served as a constant reminder of the prevailing norm that men should use condoms with female sex workers. These euphemisms allowed people to make reference to an intimate behaviour without having to directly make reference to sex and condoms. Such exchanges occurred between male partners and all different types of social network contacts, including friends, workmates, sex establishment owners and employees, wives, and family. For example, Roberto, 43, who disliked condoms and rarely used them, described such an exchange with the owner of one of the sex establishments that he frequents,

He has told me that I should be careful, but he hasn't told me ‘You have to put on a condom’. He has told me, “Be careful, that there are lots of diseases in the street, since you are a man who supposedly takes care of himself, be careful, take care of yourself.

José, who only went out with women from one casa de cita, described receiving similarly veiled advice from his mother,

[She says to me] that I should be careful. ‘Be careful…be careful in the street because you are alone, when you go out with a woman, protect yourself.’

Edgar, a 28-year-old consistent condom user, explained that his wife, who he has never directly told that he has sex with other women, will simply say to him Protegete, or protect yourself, without having to spell out exactly what that means.

In addition to receiving advice about condoms, study participants also encouraged their friends to use them. Santo, who was a consistent condom user since his youth, asked his friends if they are “preparado” (“prepared”) when they are going to have sex with a female sex worker as a reminder to use condoms. Ramón, who generally stopped using condoms with sex work partners he had sex with regularly, still encouraged his friends by telling them that he himself used them. He justified this since he always uses condoms at first. Ramón's use of his own behaviour as an example to his friends reflects the interconnectedness between descriptive and injunctive norms, whereby telling others that you use condoms can itself be a form of encouraging them to use them. Of note, Juan and Roberto, the two participants who reported never or rarely using condoms, indicated that they did not tend to give advice to their social network contacts about using condoms. Roberto explained that while he occasionally encouraged his friends to use condoms, he felt that it was a waste of time because he could never really know if anyone used them,

No, no, no…the thing is, I supposedly tell him to be careful, all the time I tell him to be careful…[But] whatever I may say, to put on a condom when you get to the room, you know what you are going to do. I am not seeing what you are doing. So, when two people do this, shut themselves in a hotel room, I don't see what they are doing! What do I gain by telling them to use condoms or not to use condoms.

Roberto's words highlight a challenge to using normative influence to promote private behaviours such as condom use, since there is no way to know what goes on during private encounters.

In addition to the reinforcement of condom use via advice, study participants also described exchanging advice and/or opinions about the character and reputation of a sex worker with their social network contacts. Such advice fits into the process utilised by most participants to “estudiar una mujer” (“study a woman”) to decide if they are going to have sex with her. All participants mentioned this process of gathering information as well as using their own instincts to decide if they should have sex with a woman. Implicit in this process of studying a woman was an assessment of the man's perceived risk of getting HIV from the woman. For some men this process was first guided by specific norms about acceptable sex partners. For example, Santo did not think that he should have sex with sex workers who have tattoos or wear wigs. José considered only the women from the high-end casas de cita as “appropriate” sex partners because they wear “thousands of pesos (worth) of clothing” and are the most beautiful. In contrast to these assessments of the exterior qualities of a woman, participants with a propensity for developing long-term relationships with female sex workers described focusing their assessments on the interior nature and character of the woman with whom they wanted to establish confianza (trust). Among men who rarely used condoms, this assessment of character played a central role in their decision about whether or not to have sex with a woman as they had to decide if they could risk having unprotected sex.

Participants also described giving advice or sharing their opinions about a woman with their social network contacts to help them decide whether or not to have sex. For example, Luis explained that he had warned a friend who never used condoms about not having sex with a woman who was rumoured to be infected with HIV, emphasising the impact it could have on his life in the following manner,

… he was with a woman who had been rumored that she was sick…with HIV. I told him, ‘That woman is sick. Be careful…I told him this because I know he doesn't like to use condoms.

While Roberto was reluctant to advise his social network contacts about using condoms because there was no way to confirm whether or not they actually used them, he was more open to advising his contacts to avoid women whom he considered “too active”, which referred to the perception that women who went out with more clients could pose a risk for HIV,

Now, I advise my friend, when I see that the woman is too active, if she is too active, I draw his attention to this, if he is my friend. But if I tell him and he calls the woman over and tells her, never again in my life [will I give advice], never again in my life. Because you need to know how to take care of yourself.

As indicated by these quotes, participants gave advice about women to their friends when they felt that they had information that could help protect them from exposure to HIV. In contrast to the advice about condom use, which reinforced the well-established norm that condoms should be used with female sex workers, advice about the character or reputation of a woman provided new information related to HIV risk that may have been previously unavailable to the male partner.

How do norms and advice influence HIV-related behaviours?

The narratives of male partners highlight their perception that condoms should be used with sex workers, as well as the frequent exchange of advice about condoms as a reinforcement of the norm. The lingering question, however, is whether these norms and advice impact HIV-related behaviours among male partners and if so, how? While participants articulated a direct link between receiving advice about the character or reputation of a sex worker and their decision to have sex with her, the link between receiving advice about using condoms and actual condom use was less direct. For example, Pedro's friend “studied” a woman Pedro was interested in having sex with and warned him against it because she did not inspire confianza and Pedro decided not to have sex with her,

It is better to just keep talking with her, stay friends, friends, but not pass beyond that. I mean, but if you go by it, go by….'that I love her, and I want it… it is better to prevent than to lament, as they say.

José explained that if a friend advised him against having sex with a woman, he may just give her a tip and look for an excuse to leave the sex establishment,

Because sometimes you sit with a woman [in the establishment] and someone tells you, “She is not good for you”. More or less, you think and analyse and you say to the woman, “Look, take this tip because I am going home' or ‘they are calling me, I have to go’.

These two experiences highlight how friends' advice may influence the decision of whether or not to have sex with a specific woman.

With regard to condom use, most participants considered condom use with a sex worker to be the correct and expected behaviour of what they “should” do. Not all of these participants always used condoms with sex workers, however, either due to dislike of condoms in general or because they had developed confianza or a relational dynamic with their partner that made them not want to use condoms. Participants who reported consistently using condoms with sex workers identified specific motivators including having previously been infected with an STI and wanting to avoid repeat infection, and the desire to prevent having a baby with a sex worker. Santo shared a story of how he first started using condoms with sex workers at a young age after being inspired by a friend who was prepared to not have sex because he did not have a condom,

So, one night we had a couple [women].We were drinking and he came and knocked on the door. ‘What happened?’ ‘Nothing, there are no condoms here’, I said, ‘What? But, just do it [without one]’. ‘If I don't get a condom, I am leaving’.

Being the younger of the two friends Santo had to go out in the middle of the night looking for condoms, which were not easy to find as this event occurred in the early 1980s before the rise of HIV. He referred to this experience as a type of consejo or piece of advice, which reflects the role of example and modeling by peers, or the descriptive norm, as an influence on behaviour.

Twenty-four year old Juan, who had not used condoms in over a year, reflected on how he believed that he would start to use condoms once he himself was ready to poner de su parte” (give your part), which reflects the idea of internalising the norm – or acting on it due to a sense of personal meaning and consequence (Kerr et al 1997),

But then I think, ok, no, I always like to follow advice, but I think that in life there are stages and sometimes you are changing, you, something enters your thinking and you put in your part.

Related to this idea of internalising norms, Santo, who had been inspired by his friend at a young age to use condoms with female sex workers to prevent pregnancy, reflected on the challenges to achieving impact with advice about behaviours that are good for you,

Any positive influence is hard for people to accept, the negative [influence] anyone can accept. And more so when you are drinking, understand? The positive [influence], you have to have lots of awareness [to follow], especially when you are drinking, but the negative, no, the negative….

This quote again reflects how in order for norms to truly influence behaviour they have to be internalised by the individual.

A final example of the impact of norms and advice was brought up by participants in reference to how they disclose their own behaviour to their social network contacts. Juan and Roberto, the two participants who never used condoms, shared a common strategy regarding disclosure of their lack of condom use: when asked, they tell people that they use condoms. This strategy was used by both men to avoid judgment and critique from their social network contacts, and reflects their perception that others think they should use condoms. For example, Juan explained below that he would tell his friends that he used condoms to avoid getting lectured at,

Without telling you a lie, sometimes when I with person X and sometimes they ask me, ‘Are you with that person?’ and I say, ‘Yes’. ‘But are you protecting yourself? ‘Of course, papa’. But it is a lie…I say this because, if I say that I didn't use one, they are going to start talking, and they are going to start looking for obstacles and say to me, “Think about this…’ So I sometimes say, ‘of course’, but it's a lie.

Juan and Roberto's tendency to lie about condom use reflects that while not inspiring them to use condoms, both men feel pressure to make others think that they do use them. Edgar, who had always told people that he used condoms even when he didn't, doubted that people would ever really say they do not use condoms,

Look, people never say they are not using. You know why? An example, you would ask me, ‘Do you use condoms?’ If I were to tell no, you are not going to trust me. What are you going to say about me? For example, ‘Ah no! He goes to bed with the first woman he finds, I don't know what type of disease he can have.’ Or whatever, a guy who is not trustworthy doesn't use condoms. So, the majority of people have this system. As soon as you ask them, ‘do you use condoms’ the first thing they say, ‘Yes’, even if they don't use them.

Edgar makes a connection between trustworthiness and condom use as further evidence of the salience of the pro-condom norm in this setting. In this example, the impact of these norms, however, is to lie rather than to use condoms.

Discussion

We explored men's perceptions of norms related to condom use and how these norms influence behaviours. Based on our previous research, we were particularly interested in improving understanding of what injunctive norms around condom use, as communicated through encouragement to use condoms by personal social network contacts, mean to men and through what mechanisms they may influence behaviour. In contrast to our previous study, in which we used a deductive approach and quantitative methodology, in this study, we elicited men's narratives about this phenomenon to improve our understanding of processes of normative influence on condom use.

All study participants, regardless of their own condom use behaviour, shared the perception that the prevailing injunctive norm was that men should use condoms with sex workers. Given the long history of HIV prevention work in the female sex industry of the DR, the existence of this strong norm is not surprising (Kerrigan et al., 2006). Even within this normative environment that dictates condom use as the acceptable behaviour, however, participants described unique and nuanced approaches to managing their own behaviours. They did not exist in what Brennan (2004) has referred to as a “coherent or homogenous” culture where all men act in the same manner. Therefore, while norm-based social network interventions may communicate and reinforce pro-condom norms, there is also a clear need for integrated efforts at the individual, dyadic, and socio-structural levels, to address the distinct influence of social and network-level norms, that are maintained through social surveillance, and internalised or “self-administered” norms (Kerr et al 1997). Additionally, our findings suggest that peer-based efforts to promote consistent condom use need to extend beyond the normative discourse of “be careful in the street” to promote more meaningful communication between men and promote experience-based discussion about condom use in the context of steady partnerships (Hirsch et al 2009).

We also found that encouragement to use condoms is frequent and is provided from friends and family. While this advice serves as a reminder of condom use norms, men did not articulate a direct link between advice and condom use behaviour, consistent with our previous quantitative findings (Barrington et al., 2009). The lack of a clear articulation of this link, or what Lincoln and Guba (1982) (p 242) describe as “plausible semantic attributional linkages”, suggest that the pathways of influence from perceived norms to behaviour in the context of steady partnerships is nuanced and complex; men may draw on what others say they should do, but it is not the only source of influence. One explanation is that injunctive norms are not as influential on a private behaviour like condom use since the threat of social sanction, which is theorised as the cause of influence, is not salient with a behaviour that is not usually observed. This interpretation is challenged, however, by findings from other studies with male clients of female sex workers, in which both descriptive and injunctive norms have been significantly associated with condom use (Ford et al., 2002; Yang et al., 2010b; Anastario et al., 2013). One challenge to comparing these studies, however, is that each used a unique definition of the social reference group and a distinct approach to eliciting information about these groups, which could itself influence the nature of the findings.

There is an emerging experimental literature comparing the influence of descriptive and injunctive norms and assessing the interplay between norms (Smith et al 2012). Findings from these experimental studies are varied. In one study of motivations to intervene when people have been drinking, injunctive norms had a positive and longer lasting effect on intervening behaviours compared to descriptive norms (Mollen et al., 2013). In contrast, in a study of fruit consumption among adolescents a health promoting descriptive norm positively affected fruit intake among adolescents while a health promoting injunctive norm did not; the injunctive norm actually caused a decrease in fruit intake intentions (Stok et al., 2013). Smith et al (2012) explicitly studied the interplay between descriptive and injunctive normative influences on pro-environmental behaviours. These authors argue that a supportive injunctive norm would only motivate behaviour if supported by the descriptive norm; that is, alignment between norms is critical to their influence on behaviour. In their experimental study they found that when individuals were exposed to incongruent descriptive and injunctive norms, behavioural intentions were reduced to the same level as exposure to negative injunctive and descriptive norms. The power of each norm was dependent on the extent to which the other norm was in alignment.

Our findings provide additional support for the potential interplay between norms and need for alignment. While most men in our sample described using condoms at some point when they had sex with sex workers, there were two unique characteristics of the men who never used condoms: first, they lied to others about their condom use; and second, they did not give advice about condoms to others. We consider the fact that men lied about their condom use as a reflection that pro-condom injunctive norms are indeed influential, but just not always in a positive way (Stok et al., 2013). These men felt pressure but rather than using condoms, the influence of these norms was to create a negative reactive response whereby they lied in order to avoid lectures and judgment from their friends. Other studies in Latin America have also described of men lying to members of their social networks about their sex lives (Fleming et al., 2013). While lying is not a direct HIV risk behaviour, such insincerity can be a barrier to intimacy or trust between friends (Nardi 1992). Another study with male partners of female sex workers in the DR found that many men lack trusted friends and that even with those they trust, conversations about sexuality rarely extend beyond the same “you should be careful” discourse (Fleming et al 2012). Hirsch et al (2009) found that among Mexican migrant men in the USA, those who had larger social networks had more sex partners than their counterparts with smaller networks, but those who had more frequent contact with their networks had lower numbers of partners. While the outcome is different than our study, these findings still suggest that the nature of social ties may also shape the nature of social influence. These authors emphasise the need understand both what men say and what they do to understand the context of risk (Hirsch et al 2009).

Additionally, when men who do not use condoms lie to their friends and say they do, they are falsely contributing to the descriptive norm; that is, they are making it appear that they adhere to the norm that “men who have sex with sex workers use condoms”. In contrast, by not giving advice about condom use, they are not contributing to the injunctive norm that “men who have sex with sex workers should use condoms”. In our previous research, we found that while injunctive norms were not associated with condom use, they were associated with descriptive norms: men were significantly more likely to think their friends used condoms if these friends had provided encouragement towards condom use. Our current qualitative finding that men who do not use condoms also don't tend to give advice provides additional support for Smith's et al (2012) argument that aligned norms are potentially more influential. While encouragement may not have a direct influence on condom use, it does appear that it may suggest that the provider of the encouragement is a condom user. One important grey area is the men who use condoms but not consistently and still encourage their friends to do so. There is a need to identify innovative ways to promote more sincere communication among men about their experiences with condoms to extend communication beyond the superficial “you should be careful” discourse (Fleming et al 2012). Future research on condom use and other sexual behaviours should continue to explore the interplay between injunctive and descriptive normative influences to develop innovative approaches to norms-based interventions.

While the focus of this study was condom use, participants mentioned other sexual behaviours that were prone to normative influence through advice from male friends. In contrast to advice about condoms, which was either passively received or provided to passive listeners, study participants described actively seeking out the advice of their male friends regarding the character and reputation of a woman they were interested in having sex with. The receptivity to advice about potential partners among both users and non-users of condoms highlights the fact that male partners used more than just condoms to reduce their risk of HIV infection. One concern with these homegrown strategies is that male partners' use of advice about potential sex partners can increase their HIV vulnerability as this advice may make men who do not like using condoms feel that they can assess a woman's possibility of HIV infection. It is still important to understand men's prevention strategies in order to address them in future interventions. Additionally, there is an urgent need to assess normative influence on other HIV-related behaviours among men, in particular HIV testing and care seeking behaviours, which are critical components of HIV prevention that have not been explored in depth as sources of normative influence.

Limitations

Our study had several limitations. Our sample of eleven was fairly small, but we conducted two interviews with most participants, therefore achieving greater depth. We monitored key themes related to social networks, norms and condom use and determined that we had reached a point of saturation of key themes related to our study aims. While our use of purposive sampling to recruit participants does not allow us to generalise beyond the context of La Romana, we do believe our thick description of the setting allows our findings to be transferred to other contexts (Lincoln and Guba 1982). Based on our study aims, we focused on condom use as the key behaviour of interest but we identified other behaviours related to sex partner selection and relationship dynamics that should be explored in more detail in future research. Finally, the very presence of the strong pro-condom norms could have influenced participant's responses. Due to the fact that a few participants indicated their dislike for condoms and revealed lying about condom use, however, we are fairly confident that this norm did not overly influence our exchanges.

Conclusions

We found a strong pro-condom injunctive norm among male partners of female sex workers in La Romana, Dominican Republic. This injunctive norm was not articulated as being highly influential on condom use behaviour among participants, suggesting the need to complement encouragement to use condoms with other forms of influence through communication. Such efforts should seek to promote alignment or coherence between injunctive and descriptive norms as this may strengthen normative influence. Additionally, men who did not use condoms lied others saying they did, highlighting a potential negative reactive response to norms. Findings support the need for integrated approaches at the individual, dyadic and social network levels to promote and reinforce norms around HIV protective behaviours among male partners of female sex workers. Future research should continue to explore the interplay and alignment between descriptive and injunctive condom norms in the context of both private and public HIV-related behaviours.