Predicting Attitudes Towards Transactional Sex: The Interactive Relationship Between Gender and Attitudes on Sexual Behaviour

This article explores explanations for attitudes towards the acceptability of transactional sex. The sparse research investigating attitudes towards transactional sex uncovers a link between gender equality, or feminism, and a lack of support for the trade in sex. However, there are no research agendas that attempt to explain variance in attitudes towards transactional sex where support for gender equality is widespread throughout a population. We estimate regression models utilizing the 2017 Danish Values Survey (Den Danske Værdiundersøgelse) in order to predict views on the acceptability of transactional sex. While the trade in sex is legal in Denmark, we find that a majority of respondents hold negative attitudes towards transactional sex, which conveys a lack of congruence between public opinion and policy. Further, gender is a powerful predictor of attitudes towards transactional sex, with women finding it less acceptable. In addition, the analysis uncovers that general attitudes towards sexual behaviour are the largest predictor of views on the acceptability of transactional sex. That being said, we find that men’s views on the acceptability of transactional sex are more of a function of their general attitudes towards sexual behaviour than they are for women. The findings indicate that, unlike men, women appear to differentiate between their attitudes towards general sexual behaviour and their views on sexual behaviour that they may associate with negative societal implications.


Introduction
The trade in sexual services is gaining ground as a budding area of study as scholars from various fields are paying increasing attention to sex trade-related issues (e.g. Crowhurst 2019; Duff et al. 2015;Febres-Cordero et al. 2018;Garofalo Geymonat 2019;Majic 2014;Tambiah 2005;Della Giusta, Di Tommaso and Strøm 2009;Weitzer 2005). Although interest in this area is growing and important research advancements have been made, relatively little is known about individual level attitudes towards the trade in sex and the predictors of those attitudes. Only a small number of studies have set out to investigate this topic (Abrams and Della Fave 1976;Basow and Campanile 1990;Cosby et al. 1996;Jakobsson and Kotsadam 2011;Libo Yan and Zhou 2018;Lo and Wei 2005;May 1999;Räsaänen and Wilska 2007;Taylor 2005; Valor-Segura, Expósito and Moya 2011).
A better understanding of people's attitudes towards the trade in sex is needed since the results of previous studies are limited. Moreover, research indicates that negativity towards the sex trade may have adverse effects on sex workers' lives, including their access to health, justice, safety and social support (Duff et al. 2015;Febres-Cordero et al. 2018;Kuosmanen and de Cabo 2018;Kurtz et al. 2005;Lazarus et al. 2012;McClelland and Newell 2008;Platt et al. 2018;Scambler and Paoli 2008;Vuolajärvi 2019aVuolajärvi , 2019bWeitzer 2017). Such findings are cause for concern since individuals selling sexual services face excessive risks of suffering abuse, exploitation, poor health, stigma and violence-especially under repressive policies (Levy and Jakobsson 2014;Östergren 2020;Platt et al. 2018;Sloss and Harper 2004). Among the different policy options that have been debated, decriminalization has emerged as a prominent approach for improving the situation of sex workers, with countries, states and organizations all around the world reshaping their policies accordingly (Abel 2014;Albright and D'Adamo 2017;Amnesty International 2016;Armstrong 2016;2015;Miranda et al. 2019;Kilvington, Day and Hard 2001;Open Society Foundations 2015;World Health Organization et al. 2013).
At the same time, repressive prostitution policies continue to prevail in most countries and a persistent focus on criminalization is observable, based on the idea that trade in sex is inherently harmful (Altemimei 2013;Heber 2018;Kuosmanen and de Cabo 2018;Östergren 2020;Strøm 2009;Vuolajärvi 2019aVuolajärvi , 2019b. While the type of policy in place plays a critical role, policies alone are not enough to get to the bottom of the various hardships that many sex workers face. The fact that sex workers continue to suffer abuse even where the sex trade has been decriminalized has led scholars to conclude that a fundamental change in how society views women and sexuality is what is ultimately needed (Armstrong 2015). Such important observations about the short-comings of policies point to the pressing need to better understand people's attitudes towards the sex trade.
To further our understanding of sex trade-related issues, in this study, we ask which factors influence individual level attitudes towards the acceptability of trading sexual services. A nationally representative survey from Denmark inquiring about views on prostitution provides us with the unique opportunity to explore attitudes in a country where it is legal to sell and pay for sexual services, as long as the sex worker has reached the age of 18, registers and pays taxes (Kofod et al. 2011). While the Danish survey used "prostitution", scholars in the field have abandoned "prostitution" due to its negative connotations (McMillan, Worth and Rawstorne 2018). Therefore, we deploy the term "transactional sex" as an alternative to refer to the selling and paying for sexual services when not directly discussing the survey question. Although we recognize the worth of using "sex work" instead, we believe that "transactional sex" is a better substitute when discussing public opinion in Denmark. The reason being that sex work is not recognized as a legal profession, privy to unemployment benefits (Kofod et al. 2011). Given the ideological homogeneity of the Danish population in regards to gender equality and liberalism (Holtug 2012), the Danish survey also allows us to move beyond the explanations provided by previous studies when investigating divergent attitudes towards the sex trade. Our analysis reveals that gender and general attitudes towards sexual behaviour are the largest predictors of attitudes towards prostitution in Denmark, with women finding prostitution less acceptable than men. Further, the findings indicate that the relationship between attitudes towards sexual behaviour more generally and prostitution in particular is characterized by large gender differences. On average, men who view non-committal sex as completely acceptable are twice as likely as their female counterparts to view prostitution as completely acceptable. In contrast to men, women appear more likely to separate their views on prostitution from their attitudes towards general sexual behaviour. The reason may be that women are more prone to associate prostitution with negative societal implications, which are more likely to affect women than men. The occurrence of prostitution has been said to have a negative impact on people's attitudes towards women and women's position in society by perpetuating traditional gender roles and unequal power relations between women and men (Basow and Campanile 1990;Bernardo Ródenas 2001;Valor-Segura et al. 2011).

What Predicts Attitudes Towards Transactional Sex Generally?
Since few studies have looked at individual level attitudes towards the acceptability of transactional sex, the existing literature is currently comprised by more debates than trends. However, some reoccurring themes are discernible. One set of explanations ascribes importance to feminism, gender equality and attitudes towards women in general. A link between feminist orientations and negative attitudes towards prostitution was observed in an earlier study of United States (U.S.) undergraduate students (Basow and Campanile 1990), suggesting that people's views of women may have an impact on their attitudes towards transactional sex. More specifically, the study found that students who supported feminism were unlikely to believe that women who sell sex do so out of economic necessity. Instead, they related prostitution to the exploitation and subordination of women. Such students disapproved of the idea of decriminalizing or legalizing the sex trade, probably as a result of their negative associations.
Using a larger and more representative sample than previous studies, Jakobsson and Kotsadam (2011) found that support for gender equality was linked to negative attitudes towards transactional sex in Norway and Sweden. Support for gender equality also correlated with support for punitive measures against sex workers and clients. That being said, the authors found that Swedes were generally more negative towards transactional sex than their Norwegian neighbours. They explain this dissimilarity as resulting from different feminist traditions and ideas of gender equality. Differences in how political parties on the left have treated transactional sex in the two countries might have also played a role, with the Norwegian left-leaning parties coming to support client criminalization much later than their Swedish counterparts. Jakobsson and Kotsadam (2011) speculate about whether the variance between Norway and Sweden may be related to the countries' different histories of client criminalization, which came about in Sweden in 1999 and a decade later in Norway, in 2009. In 2009, the Nordic neighbouring country of Iceland also introduced client criminalization. The different ways prostitution has been framed in the public debate in Sweden and Norway could also explain the divergence between the two countries. Whereas Sweden has tended more towards a radical feminist frame that treats prostitution as men's violence against women, Norway has leaned more towards liberal feminism and the idea of prostitution as an issue relating to economic inequality or the individual (Jakobsson and Kotsadam 2011). As observed by previous scholars, liberal attitudes are likely to result in more positive views of prostitution (Peracca, Knodel and Saengtienchai 1998).
In contrast to other scholars, Valor-Segura et al. (2011) found a link between hostile attitudes towards women and support for prohibiting prostitution in Spain. Hostile sexism was also associated with negative beliefs about men who pay for sex. In addition, the authors found that people who harboured hostile attitudes towards women and supported prohibitive measures against the sex trade were more inclined to victim-blame sex workers subjected to physical or sexual assault. People who supported the legalization of the sex trade did not engage in victim-blaming to the same extent. People supporting legalization were inclined to perceive of individuals selling sexual services as sex workers. The contrasting findings of previous studies regarding the relationship between attitudes towards women and transactional sex could be a product of different measures. The variable used by Valor-Segura et al. (2011) to represent gender equality was a latent measure composed of several different survey questions, while Jakobsson and Kotsadam's (2011) variable for gender equality was a singular measure of commitment to gender equality. The contrasting findings of previous studies could also be explained by the conflicting position the issue of transactional sex holds in the women's rights movement, with different feminist movements taking diverging positions. There are feminists who view the abolition of prostitution as the only way to achieve gender equality, whereas others view women's rights and sex worker's rights as intrinsically interlinked (Ward and Wylie 2017).
The variation in findings concerning the relationship between individual level attitudes towards feminism, gender equality, women and transactional sex denotes the need for further research. The same can be said about two possible explanations for attitudes towards transactional sex, which previous studies have barely addressed, namely attitudes towards immigration and personal privacy. Attitudes towards immigration may interlink with attitudes towards transactional sex since migrant sex workers constitute a notable share of the sex worker population in several countries (Febres-Cordero et al. 2018;Gülçür and İlkkaracan 2002;Mai 2013;Plambech 2016;Vuolajärvi 2019a;2019b). Despite the fact that numerous scholars have emphasized migrants' long-standing involvement in the sex trade around the world, only one study appears to have looked at the relationship between individual level attitudes towards immigration and transactional sex (Jakobsson and Kotsadam 2011). In particular, Jakobsson and Kotsadam (2011) expected anti-immigrant sentiment to correlate with a more positive evaluation of paying for sex. While the explanation put forward is somewhat unclear, the authors found support for their hypothesis in Norway where they also found a link between negativity towards immigrants and increased support for criminalizing sex workers. The same was not true for Sweden where anti-immigrant sentiment correlated with support for punitive measures against clients. Additional inquiry into the relationship between individual level attitudes towards immigration and transactional sex is needed since scholars have only just begun to touch upon this topic. Moreover, studies indicate that migrants face increased hardships when selling sex in countries where immigration policies play an important role in the regulation of the sex trade, as migrants can be targets of punitive measures relating to both immigration and prostitution policies (Gülçür and İlkkaracan 2002;Hubbard, Matthews and Scoular 2008;Lemke 2017;Levy and Jakobsson 2014;Vuolajärvi 2019a;2019b).
The relationship between individual level attitudes towards personal privacy and transactional sex would also benefit from further investigation. Surprisingly, no study appears to have looked directly at this correlation. One earlier focus group study on the marriage prospects of female sex workers in Thailand addressed the topic only tangentially (Peracca, Knodel and Saengtienchai 1998). While the researchers found that marriage to a sex worker was seen as relatively unproblematic and a matter of personal choice, the study participants largely agreed that it was better to keep experiences of transactional sex concealed from the community. The absence of further inquiry into the relationship between attitudes towards transactional sex and personal privacy constitutes a notable gap in the literature. Not only is prostitution often treated as a public nuisance issue with solicitation constituting part of the offense in many places, the private aspects of engaging in sex for payment are generally also included in repressive prostitution legislation (Altemimei 2013;Perry 1980;Sanders 2005;Stenvoll 2002). The constitutionality of repressive prostitution laws has been challenged in Canada and the U.S. by drawing on the constitutional right of privacy, albeit with differing success (Altemimei 2013;Perry 1980). Moreover, research into other forms of sexual behaviour emphasizes differing notions of public versus private as fundamental to how societies respond to specific sexual practices as well as to the formation of people's attitudes (Ball 2008;Herald 2004;Green 2001;Berlant and Warner 1998;Johnson 1989).
Another set of explanations for attitudes towards transactional sex from the literature attributes importance to traditional attitudes, as studies have associated conservatism and religiosity with negativity towards prostitution (Abrams and Della Fave 1976;Basow and Campanile 1990;Cosby et al. 1996;May 1999;Jakobsson and Kotsadam 2011;Valor-Segura et al. 2011). Conservative and religious individuals may similarly disapprove of transactional sex because it violates the traditional norms concerning sexual behaviour they adhere to-especially the idea that copulation should occur within the marriage for the purpose of procreation. From this perspective, transactional sex is seen as immoral because it is both nonreproductive and extramarital. Negativity towards transactional sex has also been linked to conservative ideas about women's role in society (Valor-Segura et al. 2011; see also : Armstrong 2015). The need to move beyond the explanations provided by previous studies is made clear by the fact that scholars have connected negativity towards transactional sex with such widely different ideological stances as conservatism, feminism, gender equality, hostile sexism and religiosity.
One important predictor for people's attitudes towards transactional sex seems to be their attitudes towards sexual behaviour in general. A more liberal stance on general sexual behaviour has been associated with more tolerant attitudes towards prostitution (Jakobsson and Kotsadam 2011). However, the relationship between attitudes towards transactional sex and general sexual behaviour requires further research as studies are few. Moreover, previous studies do not account for the fact that the impact of general sexual behaviour on attitudes towards transactional sex may be different for men and women. Looking at the existing literature, the correlation between gender and attitudes towards transactional sex appears relatively clear-cut, with men being more likely than women to display positive and tolerant attitudes towards transactional sex (Basow and Campanile 1990;Jakobsson and Kotsadam 2011;Kuosmanen 2011;Libo Yan and Zhou 2018;Lo and Wei 2005). Further, research into sexual attitudes and behaviours in other relevant areas like pornography has demonstrated significant gender differences (Lykke and Cohen 2015;Petersen and Shibley Hyde 2011). As such, when it comes to transactional sex, one would expect to find important differences between women and men who hold sexually liberal attitudes.
In fact, gender is the only socio-demographic factor with a clear pattern in the existing literature. The existence of this pattern is partially explained by the fact that gender is always included as a socio-demographic variable in studies, while other socio-demographics such as class, race and sexuality are not consistently investigated. Nevertheless, studies show that women are not only more likely to view transactional sex in a negative light when compared with men, women also tend to support punitive and restrictive measures against the sex trade to a greater extent than their male counterparts (Basow and Campanile 1990;Jakobsson and Kotsadam 2011;Kuosmanen 2011;Libo Yan and Zhou 2018;Lo and Wei 2005;May 1999;Räsaänen and Wilska 2007). Several factors may explain this gender gap in attitudes. Studies have shown that men are far more likely to pay for sex than women (Kotsadam and Jakobsson 2014;May 1999;Cosby et al. 1996) and that previous experiences of paying for sex may result in a more positive assessment of transactional sex (Stenersen et al. 2020). That being said, previous paid sex experiences alone do not account for this gap in attitudes between men and women since the majority of men do not pay for sex, even in countries where transactional sex is legal. For instance, a recent study on the sexual habits of Danes revealed that 22% of men had paid for sex at some point (Frisch et al. 2019). Nevertheless, the fact that the sex trade largely caters to men may be part of the reason why men tend to be more positive towards transactional sex than women. For instance, Libo Yan and Zhou (2018) found that men in Macau were more tolerant of sex tourism and prostitution than their female counterparts-a difference they attributed to the fact that Chinese culture is accepting of men as patrons of sex trade-related businesses (see also : Lo and Wei 2005). Moreover, men's more accepting stance may be explained by the absence of negative associations ascribed to transactional sex vis-à-vis men's position in society, whereas women's negativity may be explained by the presence of such associations. The occurrence of prostitution has been said to have a negative impact on people's attitudes towards women and women's position in society by perpetuating traditional gender roles and unequal power relations between men and women (Basow and Campanile 1990;Bernardo Ródenas 2001;Valor-Segura et al. 2011).

Case Study Denmark
Since Denmark is a relatively ideologically homogenous country in terms of gender equality and liberalism (Holtug 2012), the Danish case allows us to control for a large portion of the explanations presented by previous studies when accounting for individual level attitudes towards transactional sex. As evidenced by the latest Danish Values Survey, a large majority of Danes support equality between men and women (Frederiksen 2017). Therefore, the Danish case also allows us to move beyond explanations for diverging attitudes towards transactional sex related solely to gender equality. Similarly, traditional values are unlikely to be important since the Danish population is fairly liberal and secular (Frederiksen 2017). While some differences in attitudes towards immigration and personal privacy exist in Denmark, we do not expect the variance to be very large. General attitudes towards sexual behaviour are likely to play an important role. As people hold liberal attitudes to other sexual behaviour, they are more likely to find transactional sex acceptable (Jakobsson and Kotsadam 2011). We also expect gender to be an important socio-demographic predictor for differences in attitudes towards transactional sex even in Denmark, with women finding it less acceptable when compared with men. In line with research into sexual behaviours and attitudes in other relevant areas (Lykke and Cohen 2015; Petersen and Shibley Hyde 2011), we anticipate that the relationship between gender, attitudes towards general sexual behaviour and attitudes towards transactional sex will be characterized by differences between men and women. Given the context of Denmark we posit three hypotheses: H 1 : Gender is the most important socio-demographic predictor, with women finding prostitution less acceptable when compared with men.
H 2 : Attitudes towards general sexual conduct are the most important attitudinal predictor, with respondents holding liberal attitudes being more likely to find prostitution acceptable.
H 3 : Liberal attitudes towards other sexual activity will have a smaller impact on women's views on the acceptability of prostitution when compared with men.

Data
We utilize the 2017 Danish Values Survey-"Den Danske Vaerdiundersøgelse"-in order to predict citizen's attitudes towards the acceptability of transactional sex (Frederiksen 2017). The Danish Values Survey is conducted through Aalborg University in conjunction with the European Values Study (EVS). In fact, the Danish dataset is included as part of the joint EVS and World Values Survey (WVS) collaboration for Wave 7 (EVS/WVS 2020). The survey was conducted in 2017 and published online in April of 2019. Overall, the purpose of the survey was to investigate the values of the Danish population on a number of issues, such as politics, family, role of society, immigration, national sentiment, morality and belonging. While the survey was not designed specifically for the exploration we conduct here, it represents one instance in recent years where a national survey asks about attitudes towards prostitution. The survey is a random sample survey collected through mixed methods. Poststratification weights were calculated and provided in order to account for survey sampling bias. The study surveyed 3362 respondents. After accounting for nonresponses on important variables necessary for the multivariate analysis, there are 2948 respondents represented in the models. The survey was translated by us from Danish to English, and we provide variable coding and descriptive statistics in the Appendix for all variables.

Dependent Variable
The main variable of interest assesses respondents' attitudes towards prostitution. In particular, the survey question asks respondents, "to what degree do you think prostitution is acceptable?" The scale that the respondent must place their response on ranges from 1 to 10 where one indicates "not at all" acceptable and ten indicates "to great degree" acceptable. In survey research, question wording plays a potentially crucial role in impacting responses. Recognizing this fact, we acknowledge that a survey question using the term "prostitution" may elicit different responses than a question using the phrase "transactional sex" or "sex work". Therefore, since the survey specifically refers to transactional sex by using the word "prostitution", we use this term when discussing results.

Independent Variables
The empirical analysis includes several socio-demographic variables that are important for predicting attitudes towards prostitution or are common control variables in behavioural research. Previous studies have come to differing conclusions regarding the impact of age and education (Kuosmanen 2011;Libo Yan and Zhou 2018;May 1999;Sawyer et al. 2001). However, we believe it is important to account for them. The inclusion of age is necessary in order to account for any generational gap in attitudes towards prostitution. Education is an important control variable because schools are an influential agent of socialization on people's attitudes. Education is a continuous measure representing the highest level of education completed from one indicating no schooling (grade school) to 12 indicating a Ph.D. degree. Likewise, religiosity is another agent of socialization that impacts attitudes on sexual behaviour. Religiosity is measured as a continuous variable representing the role that God plays in the respondent's life (1 = does not matter at all to 10 = plays a very big role). Income is a control variable that is useful for assessing any social class differences in attitudes that might exist. Finally, gender is one of the main variables of interest in the analysis (0 = man, 1 = woman).
In addition, there are several attitudinal variables in the analysis in order to account for explanations of variance in attitudes towards prostitution. We include a continuous self-placement measure of political ideology from liberal (1) to conservative (10). A measure of political interest is also important to include since interest in politics has been shown to impact a range of attitudes. The respondent's view on whether marriage is an outdated institution is included to capture variance in support for traditional institutions (0 = no, 1 = yes). Similarly, we include a variable that attempts to indirectly get at attitudes towards gender equality. In particular, the respondents' level of agreement with that idea that men are better political leaders is included (0 = completely disagree, 1 = disagree, 2 = agree, 3 = completely agree). Since sexual activity is commonly thought of as private, we include an indirect measure that represents the respondents' level of commitment to privacy (0 = should not have right, 1 = probably should not have right, 2 = should have right, 3 = definitely should have right). The question that we utilize asks respondents whether they think the Danish state should have the right to collect information about everyone living in Denmark without their knowledge. Considering that street-based sex workers are the most visual of the sex worker population, and that immigrants tend to take up this type of sex work position, we include a measure that broadly accounts for attitudes towards immigrants in the analysis. The survey question asks respondents to what extent do immigrants impact their living conditions (0 = not at all, 1 = to a lesser degree, 2 = to some extent, 3 = somewhat, 4 = greatly). Finally, we include a variable that more generally measures attitudes towards sexual behaviour. The variable measures the extent to which the respondent thinks that non-committal sex is acceptable. The measure is continuous ranging from one indicating "not at all" to ten indicating "greatly".

Method
We employ multivariate OLS linear regression in order to predict respondents' attitude towards the acceptability of prostitution. The dependent variable is measured similar to traditional feeling thermometers or other Likert type survey response questions ordinarily recognized as continuous measures. The Likert-scale is 10 points; thus, respondents are provided substantial variation in response offering when indicating whether they find prostitution more or less acceptable. As a robustness check, ordered logit models were also estimated and the results were indistinguishable from the OLS results. Therefore, for ease of interpretation for the reader, we present the OLS results. In addition, we employ post-stratification survey weights in the estimation of statistical models in order to correct for sampling error and potential non-response bias. In sum, we present results from two models: a basic model with all independent variables and an interactive model with an interaction between gender and the respondent's attitude towards non-committal sex. As a robustness check, we explored the interaction between gender and all other variables of interest and found no other interactive effects. Figure 1 displays density plots of the dependent variable for the full sample of respondents and samples split by respondent gender. The full sample density plot indicates that the modal response is that prostitution is "not at all" acceptable, with around 24% of respondents selecting a one on the scale. In fact, a majority of respondents (54%) selected some type of negative sentiment (responses 1-4) towards the acceptability of prostitution. The second most selected response number is a five on the 1-10 scale, which indicates that about 15% of respondents have a neutral feeling towards the acceptability of prostitution. In sum, over two-thirds of the sample expressed either a neutral or negative attitude towards the acceptability of prostitution. The mean attitudinal value is a 4.17, which means that the sample leans slightly negative in their attitudes towards prostitution. The results indicate that public opinion does not appear to completely align with Denmark's policy, according to which transactional sex is legal.

Results
The density plot presenting the distribution of the dependent variable with the samples split by gender points to some evidence of a gender gap in attitudes on the acceptability of prostitution. In fact, a bivariate statistical test determined a statistically significant gender difference. In particular, around twice as many women than men indicated that prostitution is "not at all" acceptable. Overall, around 68% of women expressed some negative sentiment towards the acceptability of prostitution compared with 40% of men. For women, the modal response category was that prostitution is "not at all" acceptable (32%). On the other hand, more men than women chose either a neutral or positive response towards the acceptability of prostitution. The modal category for men was the neutral response (20%). Notably, around twice as many men selected response values 7 through 10 when compared with women. The result provides some support for H 1 and indicates that more inquiry needs to be made into the gendered dynamics at play in determining attitudes towards prostitution. Table 1 displays results from the multivariate OLS linear regression models and conveys which independent variables have a statistically significant relationship with respondents assessment of the acceptability of prostitution. Additionally, Fig. 2 highlights the average marginal effects of several of the statistically significant independent variables on the attitude. The relationship between socio-demographic variables and respondents' assessment of the acceptability of prostitution convey some expected results. Income has no statistically significant relationship with the attitude. However, age has a small, positive relationship with viewing prostitution as acceptable. In particular, an increase in age of 50 years is related to an increase of one point on the 0-10 scale. In comparison, education has a small, negative relationship with viewing prostitution as acceptable. For example, when comparing a respondent at the lowest level of education to a respondent at the highest level of education, an average drop of almost a point is expected (− 0.96). Similarly, religiosity has a small, negative relationship on the attitude. When comparing a respondent that indicates that God plays no role in their life to a respondent that indicates that God plays a very big role in their life, a drop of about half a point is expected in assessing the acceptability of prostitution. Finally, in the basic model, gender has the largest substantive relationship with attitudes towards the acceptability of prostitution of any The results in regards to the relationship that attitudinal variables have with the dependent variable generally align with our expectations, with only a few exceptions. First, political interest is positively related to the acceptability of prostitution. On average, a respondent that is very interested in politics is little over half a point more accepting of prostitution (0.54) than is a respondent with no interest in politics. Surprisingly, the political ideology variable indicates that as a respondent identifies as more right-wing, they are more likely to find prostitution acceptable. After investigating this variable in-depth, it appears as though the variable contains a lack of variance in responses and a large proportion of respondents identified in the middle of the spectrum. Only 6% of respondents identified as a nine or ten on the 1-10 ideological scale, and only 12% of respondents identified as a one or two. Further, 41.6% of respondents identified as values 4-6 on the ideological spectrum and 68.3% of respondent identified as values 3-7 on the ideological spectrum. The result could very well be driven by this lack of variance.
In terms of traditional institutions, respondents who indicate a belief that marriage is an outdated institution are more likely to think that prostitution is acceptable. However, since only 17.05% of respondents indicate a belief that marriage is an outdated institution, the confidence bounds around the effect of the variable are quite large due to a lack of variance. Similarly, a belief that men are better political leaders appears to have a positive statistically significant correlation with the belief that prostitution is acceptable. That being said, in the interactive model, the variable is not statistically significant. The result is expected since the interactive model does a better job of accounting for nuanced gendered dynamics of the main independent variable. In addition, attitudes towards privacy have no statistically significant relationship with a respondent's attitude towards the acceptability of prostitution.
A view that immigrants negatively impact respondents' life has a negative effect on attitudes towards the acceptability of prostitution. On average, the coefficient indicates that a respondent who conveys that immigrants have a great impact on their life evaluates the acceptability of prostitution over a half a point lower (− 0.60) than a respondent who stated that immigrants have no impact on their life. In Fig. 2, the confidence bounds around the marginal effect are moderately large. To investigate the relationship more in depth, in Fig. 3, we present predicted probabilities with other independent variables held at their survey weighted means. While the confidence bounds are relatively sizable at the ends of the x-axis, we can be confident that there is a negative relationship between attitudes towards immigrants impacting a respondent's life and views on the acceptability of prostitution. Since the survey was not specifically designed to test the relationship between these attitudes, we think that future research should investigate the relationship more in-depth.
The main focus of our empirical analysis is on the relationship between general attitudes towards sexual behaviour, gender and acceptability of prostitution. When exploring Table 1, two substantively important observations stand out. First, in the non-interactive model, the variable with the largest statistically significant and substantively important effect on a belief that prostitution is acceptable is the respondent's attitude towards noncommittal sex. On average, a respondent who indicates that non-committal sex is "to great degree" acceptable is 3.1 points more accepting of prostitution when compared with a respondent who stated that non-committal sex is "not at all" acceptable. The difference is very large and accounts for almost a third of the variance of the dependent variable. Clearly, attitudes towards sexual behaviour in a more general sense are strongly related to attitudes towards prostitution, which provides support for H 2 . Second, in the interactive model, the results convey that there is a statistically significant interaction between gender and attitudes towards non-committal sex when predicting a respondent's level of acceptability of prostitution. In Fig. 4, we plot the predicted probabilities for the interaction holding other independent variables at their survey weighted means. The results of the predicted probabilities indicate that the effect of a respondent's attitude on non-committal sex predicting their attitude towards the acceptability of prostitution is substantively different for women and men. In particular, a man who indicates that non-committal sex is "to great degree" acceptable is 3.34 points higher in their assessment of the acceptability of prostitution when compared with a man who said "not at all" acceptable. In comparison, a woman who indicates that non-committal sex is "to great degree" acceptable is about 2.2 points higher in their assessment of the acceptability of prostitution. The gap in change is over a point. In addition, if we compare men and women who indicate that non-committal sex is "not at all" acceptable, we see that men are only 0.77 of a point higher on their view of the acceptability of prostitution. However, when comparing women and men who indicate that noncommittal sex is "to great degree" acceptable men are almost 2 points (1.92) higher in their view on the acceptability of prostitution? Substantively, holding all other variables at their survey weighted mean, when going from the lowest value to the highest value on the non-committal sex variable women go from viewing prostitution negatively in terms of acceptability to slightly less negatively (one point from neutral). In contrast, when going from the lowest value to the highest value on the same variable, men move from viewing prostitution negatively in terms of acceptability to positively. The results indicate that men's attitudes towards transactional sex is more subject to their overall attitudes towards sexual behaviour. On the other hand, women appear better able to compartmentalize their attitudes towards overall sexual behaviour and their attitudes towards transactional sex. The results provide convincing support for H 3 .

Conclusion
In this study, we asked which are the influential factors on individual level attitudes towards the acceptability of transactional sex. At the beginning of this article, we argued that a better understanding of people's attitudes towards the trade in sex is needed because the existing literature is limited. Another reason we pointed to for further research into this topic is that sex workers may face serious hardships as a result of negativity towards transactional sex. Moreover, we argued that a focus on individual level attitudes could assist in broadening the current debate regarding policies and how to best approach the sex trade and its related issues. Questions concerning policy are undoubtedly important ones, but so are questions about what shapes people's attitudes towards transactional sex. By understanding what predicts individual level attitudes towards transactional sex, policymakers and other key actors may be better equipped to address sex trade-related issues. A nationally representative survey from Denmark inquiring about views on prostitution provided us with the possibility to explore attitudes in a country where it is legal to engage in transactional sex. Since the Danish population is quite ideologically homogenous (Holtug 2012), the case of Denmark made it possible to move beyond some of the explanations that scholars have provided when looking at divergent attitudes towards transactional sex, namely gender equality and traditional values. As our analysis demonstrates, there is a lack of variance in Denmark when exploring these explanations for differences in attitudes towards transactional sex. That being said, we were able to build on the findings of previous studies regarding gender (the one socio-demographic factor following a clear pattern in the literature) to look more closely at how gender interacts with attitudes towards general sexual behaviour. The results from Denmark indicate that diverging attitudes towards transactional sex may persist even when an increasing share of the population holds liberal attitudes and commits to gender equality. Other countries witnessing similar ideologically developments may want to consider this observation.
The findings that we present suggest that the majority of Danes (54%) hold negative attitudes towards the trade in sex. There is thus a lack of congruency between public opinion and policy in Denmark where it is legal to sell and pay for sexual services. Since negativity towards the sex trade may have adverse effects on sex workers' lives, this incongruence is food for thought for anyone engaging in sex trade-related policy debates. That being said, in comparison with other Scandinavian countries, a larger proportion of Danes hold tolerant attitudes towards prostitution (EVS/WVS 2020). A useful next step would be to study the topic in a comparative context to see if further incongruences between public opinion and policy exist across Scandinavia.
We find that gender is the most important socio-demographic predictor of attitudes towards prostitution, with women finding it less acceptable than men. The most powerful attitudinal predictor is attitudes towards general sexual conduct. The interaction between gender and attitudes towards sexual behaviour is notable. On average, men who view non-committal sex as completely acceptable are twice as likely as their female counterparts to view Fig. 4 Interaction between attitudes towards sex and gender on acceptability of prostitution-predicted probabilities calculated holding additional independent variables at their survey-weighted means prostitution as completely acceptable. The reason why women tend to hold more negative attitudes towards transactional may be because women are more prone to relate prostitution to negative societal implications, like the perpetuation of traditional gender roles and unequal power relations between men and women (Basow and Campanile 1990;Bernardo Ródenas 2001;Valor-Segura et al. 2011). Such associations may also explain why women are more likely to distinguish their attitudes towards general sexual behaviour from their attitudes towards transactional sex. Along the same line, the reason why men tend to be more positive towards transactional sex than women may be explained by the absence of negative associations ascribed to the trade in sex vis-à-vis men's position in society. The fact that the sex trade largely caters to men may also play a role in accounting for this gender gap in attitudes.
While our findings contribute to gender theorizing, it would be fruitful to pursue further research into the gendered dynamics at play in determining individual level attitudes towards transactional sex in order to grasp their full make-up. One suggestion moving forward would be to pay close attention to how such attitudes interact with gender and attitudes towards other societal issues both nationally and cross-nationally, as attitudes towards transactional sex are likely to differ between countries. A second suggestion for future research into attitudes towards transactional sex is to include more nuanced survey questions in order to capture the complexity of the issue. A third suggestion is for qualitative studies to continue to deepen our understanding of the links between diversity of experiences, opinions and attitudes towards transactional sex, including those of people who do not conform to binary identity categories.
Immigrants crime: continuous measure, 1 = Immigrants do not make the problems of crime worse to 10 = immigrants make problems of crime worse.

Appendix 2. Descriptive Statistics and Methodological Notes
Descriptive statistics-socio-demographics Methodological notes: • We also explored whether attitudes on divorce and gay sexual activity impacted views on the acceptability of prostitution. The variables were excluded from the analysis due to collinearity with more theoretically grounded variables of interest and an extreme lack of variance. For both variables, over 80% of respondents selected the two highest categories. The findings indicate a great deal of homogeneity on the issues. • Selecting measures from the survey to represent variance in ideas of gender equality is difficult, especially since the survey was not designed with this topic necessarily in mind. Therefore, as a robustness check, we estimated models with alternative measures included. In particular, we assessed whether attitudes towards the importance of women having rights for a democracy and adherence to traditional gender roles of women working in the home had an impact on attitudes towards prostitution. In Denmark, there is almost no variance on these issues and the variables did not have an impact on attitudes towards prostitution in statistical models. • As a robustness check, two additional measures were utilized to represent sentiment towards immigrants. We estimated models that assessed whether attitudes towards immigrants burdening the welfare system or making problems of crime worse predicted respondents' belief in acceptability of prostitution. Neither attitude had a statistically significant relationship with attitudes towards prostitution.