The Journal of Sex Research
ISSN: 0022-4499 (Print) 1559-8519 (Online) Journal homepage: http://www.tandfonline.com/loi/hjsr20
Variations of Sexual Scripts Relating to Concurrency by Race, Class, and Gender in South Africa Chris R. Kenyon, Kara Osbak, Jozefien Buyze, Saul Johnson & Jacques van Lankveld To cite this article: Chris R. Kenyon, Kara Osbak, Jozefien Buyze, Saul Johnson & Jacques van Lankveld (2015) Variations of Sexual Scripts Relating to Concurrency by Race, Class, and Gender in South Africa, The Journal of Sex Research, 52:8, 878-886, DOI: 10.1080/00224499.2014.963832 To link to this article: http://dx.doi.org/10.1080/00224499.2014.963832
Published online: 28 Oct 2014.
Submit your article to this journal
Article views: 173
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=hjsr20 Download by: [University of California, San Diego]
Date: 15 November 2015, At: 20:37
JOURNAL OF SEX RESEARCH, 52(8), 878–886, 2015 Copyright # The Society for the Scientific Study of Sexuality ISSN: 0022-4499 print=1559-8519 online DOI: 10.1080/00224499.2014.963832
Variations of Sexual Scripts Relating to Concurrency by Race, Class, and Gender in South Africa Chris R. Kenyon, Kara Osbak, and Jozefien Buyze HIV=STI Unit, Institute of Tropical Medicine
Saul Johnson Downloaded by [University of California, San Diego] at 20:37 15 November 2015
Health Development Africa
Jacques van Lankveld Open University, Heerlen, The Netherlands It is unclear whether higher rates of sexual partner concurrency in Black South Africans are due to socioeconomic or cultural factors. We used a nationally representative sample of 9,728 individuals aged 16 to 55 from a study conducted in 2009 to examine how the norms pertaining to concurrency and the practice of concurrency vary by race, class, and gender. The percentage of men reporting point concurrency was 14%, 6.5%, and 2.5% in Blacks, coloreds, and Whites, respectively (p < 0.001). These percentages increased to 45.7%, 24.7%, and 11.7%, respectively, for those reporting lifetime concurrency (p < 0.001). In all the racial groups, men exhibited more favorable attitudes toward concurrency than women did. For a range of indicators, White men and women had less favorable attitudes toward concurrency than Black men and women. These differences remained after controlling for a range of confounding variables. In the adjusted logistic regression model, reported concurrency in men was associated with a younger age, Black race, being in the lowest income tertile, not being in a stable relationship, and expressing various positive attitudes toward concurrency.
Human immunodeficiency virus (HIV) prevalence varies considerably between racial groups among 15- to 49year-old South Africans—from 22.7% (95% Confidence Interval [CI] 21.2% to 24.3%) in Blacks to 4.6% (CI 3.3% to 6.4%) in coloreds to 0.6% (CI 0.2% to 1.8%) in Whites (Shisana et al., 2014). The term colored is a common and socially acceptable term in South Africa for individuals of mixed race. Differences in the prevalence of sexual partner concurrency (having more than one partner at a time) has been put forward as a possible explanatory factor (see Figure 1) (Johnson, Dorrington, Bradshaw, Pillay-Van Wyk, & Rehle, 2009; Kenyon, Buyze, & Colebunders, 2013; Kenyon, Dlamini, Boulle, White, & Badri, 2009). Concurrent relationships are theorized to lead to a nonlinear acceleration of HIV transmission due to two main factors. First, concurrency increases the size of the number of persons that are directly or indirectly sexually connected in a network at any one point in The authors would like to thank Brendan Maughan-Brown for useful comments on an earlier draft of this article. Correspondence should be addressed to Chris R. Kenyon, HIV= STI Unit, Institute of Tropical Medicine, Nationalestraat 155, Antwerp 2000, Belgium. E-mail: [email protected] Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/hjsr.
time; see Figure 1 (Morris & Kretzschmar, 1997; Morris, Kurth, Hamilton, Moody, & Wakefield, 2009). Second, the infection-transmission-lowering effect of ‘‘partner sequencing’’ is lost (Kretzschmar & Morris, 1996). Various studies have found a strong correlation between the point prevalence of concurrency and HIV prevalence by racial and ethnic group in South Africa (Kenyon, 2013a, 2013b; Kenyon et al., 2013; Kenyon et al., 2009). It is not clear, however, whether the higher concurrency prevalence in Black South Africans is due to socioeconomic or cultural factors. A number of authors have argued that socioeconomic factors such as high migration rates (Hunter, 2005; Lurie et al., 2003; Marks, 2002), unemployment (Hunter, 2010), and reduced marital rates are responsible (Hunter, 2010). Other, mostly qualitative, work has found that culturally supported norms, such as the tolerance of having main and side partners, play a major role (Jana, Nkambule, & Tumbo, 2008; Kenyon, Boulle, Badri, & Asselman, 2010; Kenyon, Zondo, & Badri, 2010; Leclerc-Madlala, 2009). One way to investigate whether cultural or socioeconomic factors play a more dominant role would be to examine the relationship between concurrency and the norms pertaining to concurrency by race, class, and gender. We are aware of no published study that has
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
VARIATIONS OF SEXUAL SCRIPTS
Figure 1. Conceptual framework illustrating how differences in norms toward concurrency could, via resultant differences in behavior and connectivity of sexual networks, contribute to differences in HIV prevalence by race in South Africa and offer opportunities for HIV intervention strategies. (Red and blue circles=squares ¼ HIV-infected and HIV-uninfected individuals, respectively.)
examined this relationship. In this article we used a nationally representative survey of South Africans to answer the following related questions: Do norms toward concurrency vary by race and gender—before and after controlling for confounders? If they vary, then is it only subgroups within particular races that hold more favorable views, or is there evidence of a more generalized difference between races? Is there a correlation at an individual level between expressing specific favorable norms to concurrency and engaging in concurrency? In addition, we assessed whether any evidence exists for pluralistic ignorance in regard to the norms toward concurrency. Pluralistic ignorance is a situation in which a majority of group members privately rejects a norm but incorrectly assumes that the majority of the group accepts the norm (Prentice & Miller, 1993; Schroeder & Prentice, 1998). As a result the majority goes along with the norm. The theoretical framework we employ in this article is sexual scripting theory. Sexual scripts can be considered as roadmaps for how individuals conduct themselves in specific sexual situations (Simon & Gagnon, 1987). They operate at three levels: the cultural, interpersonal, and the intrapsychic. Cultural scripts involve societal norms for sexual conduct. Via a process of acculturation, individuals acquire norms that are appropriate to their culture. Individuals, however, make adaptations to these cultural scripts to suit their own needs. Their
interpersonal scripts then involve the translation of the cultural scripts into scripts that are suitable for the persons concerned. They include the individuals’ actual sexual practices and experiences. Intrapsychic scripts involve the sexual fantasies and preferences that a person has. The theory of sexual scripting has been usefully applied to understanding various aspects of sexual behavior (Castillo & Geer, 1993; Parsons et al., 2004). In one study that analyzed data from the National Health and Social Life Survey it was found that sexual partnering was highly segregated along racial and other lines in the United States (Mahay, Laumann, & Michaels, 1994). The authors of this study argued that this segregation resulted in different sexual scripts developing within these subpopulations. It is, however, possible that assortativeness in sexual scripts may influence the structure of sexual networks. Sexual partnering has been shown to exhibit even higher degrees of segregation by race in South Africa (Kenyon & Colebunders, 2013). It is thus plausible this has resulted in differences in sexual scripts between different races.
Method The 2009 National Communication Survey (NCS) was a cross-sectional survey that utilized a three-stage, 879
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
KENYON, OSBAK, BUYZE, JOHNSON, AND VAN LANKVELD
stratified sampling approach. Primary sampling units (2001 census small areas) were selected based on probability proportional to size techniques; thereafter households, followed by individuals, were randomly selected within primary sampling units. The survey produced a nationally representative sample of 9,728 individuals aged 16 to 55. The overall response rate was 58%. For further details of the methodology and possible bias introduced by differential nonresponse, see Johnson and colleagues (2010). The questionnaire included sections on sociodemographic characteristics, sexual behaviors, norms relating to concurrent partnering, social norms and attitudes toward condom use, faithfulness and abstinence, and exposure to HIV programs. The questions relating to the norms toward concurrency were derived from a process of qualitative research. The questionnaire was then piloted in four provinces in February 2009. Based on the findings of this phase, the questionnaire was refined and the final survey conducted from June to August 2009 (Johnson et al., 2010).
Cultural Scripts Five indicators of cultural scripts toward concurrency were derived from the following questions: 1. Extra partner sex okay: It’s okay to have sex with others as long as your main partner does not find out. 2. Real men: Men who have sex with a lot of women are Ingakara (‘‘real men’’). 3. You are the only one: People will tell you that you are the only one even when they are having sex with someone else. 4. Can’t trust anyone: I cannot trust anyone because people are cheating everywhere. 5. Open talk: People in my community are openly talking about the risk of HIV from having more than one partner. Interpersonal Scripts Point concurrency was defined dichotomously based on the following question: How many sexual partners do you currently have? All individuals reporting two or more current partners were coded as concurrents. Cumulative concurrency was defined as answering yes to the following question: Were there any other times in your life when you had more than one sexual partner at the same time? It included those who reported currently having two partners. Perceived partner concurrency was defined as the proportion of persons who answered yes to this yes=no question: Do you think that he=she (your most recent partner) has other sex partners? 880
Intrapsychic Scripts Three indicators of intrapsychic attitudes toward concurrent partnering were constructed using the following questions: 1. Resist temptation: I am confident that I can resist the temptation of having sex with anyone else besides my main sex partner. 2. Fill the gap: I need someone else to fill the gap in case I ever break up with my main partner. 3. Sex so good: Now and then, I go to someone else besides my main partner because the sex is so good. It is acknowledged that this three-way categorization of the indicators is subjective. It was based on an assessment by the authors of which category of sexual scripting the indicator best fitted. Some indicators could fit more than one category. For example, the ‘‘sex so good’’ indicator assesses both the respondent’s intrapsychic preferences in regard to sex as well as interpersonal practices of whether they act on these preferences. As such, it contains elements of intrapsychic and interpersonal scripts. Likewise, the ‘‘extra partner sex okay’’ indicator can be construed as evaluating whether concurrency is culturally acceptable in the community and morally acceptable for the individual respondent. For each of the questions pertaining to the cultural and intrapsychic scripts, the answer was coded as one of four options: Strongly agree, Somewhat agree, Somewhat disagree, Strongly disagree. For each variable we dichotomized all responses into all those who agreed or disagreed. Because of its importance, we report the responses to the ‘‘extra partner sex okay’’ question as dichotomized and coded according to the four options. We did not combine the variables for cultural, interpersonal, and intrapsychic scripts into composite variables but analyzed them individually. Pluralistic Ignorance We used two variables to assess for pluralistic ignorance pertaining to concurrency norms. The variable, based on the statement ‘‘I cannot trust anyone because people are cheating everywhere,’’ was used as a measure of perceived public attitude toward and practice of concurrency. The private attitude to concurrency was assessed via the statement ‘‘It’s okay to have sex with others as long as your main partner does not find out.’’ Statistical Analysis Pearson chi square tests were used to compare categorical variables. Logistic regression was used to evaluate the relationship between cultural scripts and race. This was done both without (Model 1) and with (Model 2) controlling for educational attainment (in four categories),
VARIATIONS OF SEXUAL SCRIPTS
income (in tertiles), age, marital status, and religiosity. Religious individuals were defined dichotomously as those who strongly agreed with the statement ‘‘In my day to day life, my religion or my God is very important to me in helping me deal with life problems.’’ These regressions were performed separately for men and women.
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
Correlates of Concurrency Logistic regression was used to assess the relationship between men’s concurrency on the date of the interview and the exposure variables. The final models were constructed by including all variables with p values < 0.1 0.1 on simple logistic regression. Because the reported prevalence of concurrency was so low in the women (1.1%, 0.7%, and 0.0% of Black, colored, and White women, respectively, reported concurrency), only the results of the regressions for men are reported. Race, Gender, and Class All respondents were asked to self-identify with one of four racial categories: Black (African), colored, White, and Indian. The Indians were dropped from the analyses due to their small sample size (n ¼ 146). Many sexual attitudes and practices are gendered. As a result we have stratified our analyses by gender. We used educational attainment and income tertiles as proxies for class. All analyses were made using STATA 12.0 (College Station, TX). Allowance was made for the complex sampling strategy by using the survey (SVY) methodology. The University of the Witwatersrand’s human research ethics committee granted ethical approval for the NCS.
Results The Black, colored, and White groups differed on all demographic and socioeconomic characteristics (Table 1). Compared to the Whites, the Black group was younger, poorer, less likely to have attained higher educational outcomes, and less likely to be married. The cross-tabulations of the prevalence of each of the three levels of sexual norms, practices, and attitudes by race and gender are presented in Table 2. In all cases the percentage of respondents that agreed with each statement is presented. Table 3 presents the logistic regression results of the relationship between the cultural level attitude variables and race, controlling for a range of potential confounders.
of confounding variables (only in the case of men answering the ‘‘real men’’ question was this difference not statistically significant). Compared to Black men and women, colored men and women had a trend of less favorable attitudes to concurrency but only in the ‘‘real men’’ variable was this statistically significant. A higher proportion of colored men and women than any other group answered that they ‘‘can’t trust anyone because people are cheating everywhere.’’ There was no difference in the proportions reporting ‘‘open talk’’ as to the dangers of concurrency in the different racial groups (47% to 55% answered yes to this question). In all the racial groups men exhibited more favorable attitudes to concurrency than women did. Men were twice as likely to say that extrapartner sex was okay. Women were also approximately 50% more likely to say that ‘‘one cannot trust anyone as people are cheating everywhere.’’ The ‘‘extrapartner sex is okay’’ variable is reported in more detail according to exact responses. It was not just the proportion of Black men and women who strongly agreed with the norm that ‘‘extra-partner sex is okay’’ that was higher but also the proportions who somewhat agreed and somewhat disagreed. In all, 54% of Black men versus 70.5% and 82.6% of colored and White men, strongly disagreed with this statement (p < 0.001). There was no difference between the races in the proportion who agreed that ‘‘people in my community are openly talking about the risk of HIV from having more than one partner.’’ Interpersonal Scripts Large and significant differences were evident in the percentage of each group reporting concurrency. The percentage of men reporting point concurrency varied between 14% in Blacks to 2% in Whites (p < 0.001). These percentages increased to 46% and 12%, respectively, for those reporting lifetime concurrency (p < 0.001). Intrapsychic Scripts There were statistically significant differences between the races for each of the three intrapsychic variables for men and women. The only exception to this was the ‘‘resist temptation’’ variable in the women, where there were no differences between the races. In each case Black respondents had the highest percentage reporting proconcurrency attributes, White respondents the lowest, and colored respondents intermediate. Pluralistic Ignorance
Cultural Scripts White men and women had less favorable attitudes toward concurrency than Black men and women did. These differences remained after controlling for a range
We found evidence that may be indicative of pluralistic ignorance in relation to the prevalence of concurrency. The proportion of Black, colored, and White women who reported that they ‘‘cannot trust anyone because people 881
KENYON, OSBAK, BUYZE, JOHNSON, AND VAN LANKVELD
Table 1. Demographic and Socioeconomic Characteristics of Black, Colored, and White Respondents in the National Communication Survey, 2009 Black
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
Characteristics Gender Women Men Income Richest Middle Poorest Education Primary Up to grade 11 Grade 12 or higher Age 15–24 25–34 35–44 45–56 Married Single Noncohabiting relationshipa Married or living together Divorced=widowed
N
Colored
White
%
N
%
N
%
p Value
4,222 3,704
53.3 46.7
707 496
58.8 41.2
258 173
59.9 40.1
ISSN: 0022-4499 (Print) 1559-8519 (Online) Journal homepage: http://www.tandfonline.com/loi/hjsr20
Variations of Sexual Scripts Relating to Concurrency by Race, Class, and Gender in South Africa Chris R. Kenyon, Kara Osbak, Jozefien Buyze, Saul Johnson & Jacques van Lankveld To cite this article: Chris R. Kenyon, Kara Osbak, Jozefien Buyze, Saul Johnson & Jacques van Lankveld (2015) Variations of Sexual Scripts Relating to Concurrency by Race, Class, and Gender in South Africa, The Journal of Sex Research, 52:8, 878-886, DOI: 10.1080/00224499.2014.963832 To link to this article: http://dx.doi.org/10.1080/00224499.2014.963832
Published online: 28 Oct 2014.
Submit your article to this journal
Article views: 173
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=hjsr20 Download by: [University of California, San Diego]
Date: 15 November 2015, At: 20:37
JOURNAL OF SEX RESEARCH, 52(8), 878–886, 2015 Copyright # The Society for the Scientific Study of Sexuality ISSN: 0022-4499 print=1559-8519 online DOI: 10.1080/00224499.2014.963832
Variations of Sexual Scripts Relating to Concurrency by Race, Class, and Gender in South Africa Chris R. Kenyon, Kara Osbak, and Jozefien Buyze HIV=STI Unit, Institute of Tropical Medicine
Saul Johnson Downloaded by [University of California, San Diego] at 20:37 15 November 2015
Health Development Africa
Jacques van Lankveld Open University, Heerlen, The Netherlands It is unclear whether higher rates of sexual partner concurrency in Black South Africans are due to socioeconomic or cultural factors. We used a nationally representative sample of 9,728 individuals aged 16 to 55 from a study conducted in 2009 to examine how the norms pertaining to concurrency and the practice of concurrency vary by race, class, and gender. The percentage of men reporting point concurrency was 14%, 6.5%, and 2.5% in Blacks, coloreds, and Whites, respectively (p < 0.001). These percentages increased to 45.7%, 24.7%, and 11.7%, respectively, for those reporting lifetime concurrency (p < 0.001). In all the racial groups, men exhibited more favorable attitudes toward concurrency than women did. For a range of indicators, White men and women had less favorable attitudes toward concurrency than Black men and women. These differences remained after controlling for a range of confounding variables. In the adjusted logistic regression model, reported concurrency in men was associated with a younger age, Black race, being in the lowest income tertile, not being in a stable relationship, and expressing various positive attitudes toward concurrency.
Human immunodeficiency virus (HIV) prevalence varies considerably between racial groups among 15- to 49year-old South Africans—from 22.7% (95% Confidence Interval [CI] 21.2% to 24.3%) in Blacks to 4.6% (CI 3.3% to 6.4%) in coloreds to 0.6% (CI 0.2% to 1.8%) in Whites (Shisana et al., 2014). The term colored is a common and socially acceptable term in South Africa for individuals of mixed race. Differences in the prevalence of sexual partner concurrency (having more than one partner at a time) has been put forward as a possible explanatory factor (see Figure 1) (Johnson, Dorrington, Bradshaw, Pillay-Van Wyk, & Rehle, 2009; Kenyon, Buyze, & Colebunders, 2013; Kenyon, Dlamini, Boulle, White, & Badri, 2009). Concurrent relationships are theorized to lead to a nonlinear acceleration of HIV transmission due to two main factors. First, concurrency increases the size of the number of persons that are directly or indirectly sexually connected in a network at any one point in The authors would like to thank Brendan Maughan-Brown for useful comments on an earlier draft of this article. Correspondence should be addressed to Chris R. Kenyon, HIV= STI Unit, Institute of Tropical Medicine, Nationalestraat 155, Antwerp 2000, Belgium. E-mail: [email protected] Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/hjsr.
time; see Figure 1 (Morris & Kretzschmar, 1997; Morris, Kurth, Hamilton, Moody, & Wakefield, 2009). Second, the infection-transmission-lowering effect of ‘‘partner sequencing’’ is lost (Kretzschmar & Morris, 1996). Various studies have found a strong correlation between the point prevalence of concurrency and HIV prevalence by racial and ethnic group in South Africa (Kenyon, 2013a, 2013b; Kenyon et al., 2013; Kenyon et al., 2009). It is not clear, however, whether the higher concurrency prevalence in Black South Africans is due to socioeconomic or cultural factors. A number of authors have argued that socioeconomic factors such as high migration rates (Hunter, 2005; Lurie et al., 2003; Marks, 2002), unemployment (Hunter, 2010), and reduced marital rates are responsible (Hunter, 2010). Other, mostly qualitative, work has found that culturally supported norms, such as the tolerance of having main and side partners, play a major role (Jana, Nkambule, & Tumbo, 2008; Kenyon, Boulle, Badri, & Asselman, 2010; Kenyon, Zondo, & Badri, 2010; Leclerc-Madlala, 2009). One way to investigate whether cultural or socioeconomic factors play a more dominant role would be to examine the relationship between concurrency and the norms pertaining to concurrency by race, class, and gender. We are aware of no published study that has
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
VARIATIONS OF SEXUAL SCRIPTS
Figure 1. Conceptual framework illustrating how differences in norms toward concurrency could, via resultant differences in behavior and connectivity of sexual networks, contribute to differences in HIV prevalence by race in South Africa and offer opportunities for HIV intervention strategies. (Red and blue circles=squares ¼ HIV-infected and HIV-uninfected individuals, respectively.)
examined this relationship. In this article we used a nationally representative survey of South Africans to answer the following related questions: Do norms toward concurrency vary by race and gender—before and after controlling for confounders? If they vary, then is it only subgroups within particular races that hold more favorable views, or is there evidence of a more generalized difference between races? Is there a correlation at an individual level between expressing specific favorable norms to concurrency and engaging in concurrency? In addition, we assessed whether any evidence exists for pluralistic ignorance in regard to the norms toward concurrency. Pluralistic ignorance is a situation in which a majority of group members privately rejects a norm but incorrectly assumes that the majority of the group accepts the norm (Prentice & Miller, 1993; Schroeder & Prentice, 1998). As a result the majority goes along with the norm. The theoretical framework we employ in this article is sexual scripting theory. Sexual scripts can be considered as roadmaps for how individuals conduct themselves in specific sexual situations (Simon & Gagnon, 1987). They operate at three levels: the cultural, interpersonal, and the intrapsychic. Cultural scripts involve societal norms for sexual conduct. Via a process of acculturation, individuals acquire norms that are appropriate to their culture. Individuals, however, make adaptations to these cultural scripts to suit their own needs. Their
interpersonal scripts then involve the translation of the cultural scripts into scripts that are suitable for the persons concerned. They include the individuals’ actual sexual practices and experiences. Intrapsychic scripts involve the sexual fantasies and preferences that a person has. The theory of sexual scripting has been usefully applied to understanding various aspects of sexual behavior (Castillo & Geer, 1993; Parsons et al., 2004). In one study that analyzed data from the National Health and Social Life Survey it was found that sexual partnering was highly segregated along racial and other lines in the United States (Mahay, Laumann, & Michaels, 1994). The authors of this study argued that this segregation resulted in different sexual scripts developing within these subpopulations. It is, however, possible that assortativeness in sexual scripts may influence the structure of sexual networks. Sexual partnering has been shown to exhibit even higher degrees of segregation by race in South Africa (Kenyon & Colebunders, 2013). It is thus plausible this has resulted in differences in sexual scripts between different races.
Method The 2009 National Communication Survey (NCS) was a cross-sectional survey that utilized a three-stage, 879
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
KENYON, OSBAK, BUYZE, JOHNSON, AND VAN LANKVELD
stratified sampling approach. Primary sampling units (2001 census small areas) were selected based on probability proportional to size techniques; thereafter households, followed by individuals, were randomly selected within primary sampling units. The survey produced a nationally representative sample of 9,728 individuals aged 16 to 55. The overall response rate was 58%. For further details of the methodology and possible bias introduced by differential nonresponse, see Johnson and colleagues (2010). The questionnaire included sections on sociodemographic characteristics, sexual behaviors, norms relating to concurrent partnering, social norms and attitudes toward condom use, faithfulness and abstinence, and exposure to HIV programs. The questions relating to the norms toward concurrency were derived from a process of qualitative research. The questionnaire was then piloted in four provinces in February 2009. Based on the findings of this phase, the questionnaire was refined and the final survey conducted from June to August 2009 (Johnson et al., 2010).
Cultural Scripts Five indicators of cultural scripts toward concurrency were derived from the following questions: 1. Extra partner sex okay: It’s okay to have sex with others as long as your main partner does not find out. 2. Real men: Men who have sex with a lot of women are Ingakara (‘‘real men’’). 3. You are the only one: People will tell you that you are the only one even when they are having sex with someone else. 4. Can’t trust anyone: I cannot trust anyone because people are cheating everywhere. 5. Open talk: People in my community are openly talking about the risk of HIV from having more than one partner. Interpersonal Scripts Point concurrency was defined dichotomously based on the following question: How many sexual partners do you currently have? All individuals reporting two or more current partners were coded as concurrents. Cumulative concurrency was defined as answering yes to the following question: Were there any other times in your life when you had more than one sexual partner at the same time? It included those who reported currently having two partners. Perceived partner concurrency was defined as the proportion of persons who answered yes to this yes=no question: Do you think that he=she (your most recent partner) has other sex partners? 880
Intrapsychic Scripts Three indicators of intrapsychic attitudes toward concurrent partnering were constructed using the following questions: 1. Resist temptation: I am confident that I can resist the temptation of having sex with anyone else besides my main sex partner. 2. Fill the gap: I need someone else to fill the gap in case I ever break up with my main partner. 3. Sex so good: Now and then, I go to someone else besides my main partner because the sex is so good. It is acknowledged that this three-way categorization of the indicators is subjective. It was based on an assessment by the authors of which category of sexual scripting the indicator best fitted. Some indicators could fit more than one category. For example, the ‘‘sex so good’’ indicator assesses both the respondent’s intrapsychic preferences in regard to sex as well as interpersonal practices of whether they act on these preferences. As such, it contains elements of intrapsychic and interpersonal scripts. Likewise, the ‘‘extra partner sex okay’’ indicator can be construed as evaluating whether concurrency is culturally acceptable in the community and morally acceptable for the individual respondent. For each of the questions pertaining to the cultural and intrapsychic scripts, the answer was coded as one of four options: Strongly agree, Somewhat agree, Somewhat disagree, Strongly disagree. For each variable we dichotomized all responses into all those who agreed or disagreed. Because of its importance, we report the responses to the ‘‘extra partner sex okay’’ question as dichotomized and coded according to the four options. We did not combine the variables for cultural, interpersonal, and intrapsychic scripts into composite variables but analyzed them individually. Pluralistic Ignorance We used two variables to assess for pluralistic ignorance pertaining to concurrency norms. The variable, based on the statement ‘‘I cannot trust anyone because people are cheating everywhere,’’ was used as a measure of perceived public attitude toward and practice of concurrency. The private attitude to concurrency was assessed via the statement ‘‘It’s okay to have sex with others as long as your main partner does not find out.’’ Statistical Analysis Pearson chi square tests were used to compare categorical variables. Logistic regression was used to evaluate the relationship between cultural scripts and race. This was done both without (Model 1) and with (Model 2) controlling for educational attainment (in four categories),
VARIATIONS OF SEXUAL SCRIPTS
income (in tertiles), age, marital status, and religiosity. Religious individuals were defined dichotomously as those who strongly agreed with the statement ‘‘In my day to day life, my religion or my God is very important to me in helping me deal with life problems.’’ These regressions were performed separately for men and women.
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
Correlates of Concurrency Logistic regression was used to assess the relationship between men’s concurrency on the date of the interview and the exposure variables. The final models were constructed by including all variables with p values < 0.1 0.1 on simple logistic regression. Because the reported prevalence of concurrency was so low in the women (1.1%, 0.7%, and 0.0% of Black, colored, and White women, respectively, reported concurrency), only the results of the regressions for men are reported. Race, Gender, and Class All respondents were asked to self-identify with one of four racial categories: Black (African), colored, White, and Indian. The Indians were dropped from the analyses due to their small sample size (n ¼ 146). Many sexual attitudes and practices are gendered. As a result we have stratified our analyses by gender. We used educational attainment and income tertiles as proxies for class. All analyses were made using STATA 12.0 (College Station, TX). Allowance was made for the complex sampling strategy by using the survey (SVY) methodology. The University of the Witwatersrand’s human research ethics committee granted ethical approval for the NCS.
Results The Black, colored, and White groups differed on all demographic and socioeconomic characteristics (Table 1). Compared to the Whites, the Black group was younger, poorer, less likely to have attained higher educational outcomes, and less likely to be married. The cross-tabulations of the prevalence of each of the three levels of sexual norms, practices, and attitudes by race and gender are presented in Table 2. In all cases the percentage of respondents that agreed with each statement is presented. Table 3 presents the logistic regression results of the relationship between the cultural level attitude variables and race, controlling for a range of potential confounders.
of confounding variables (only in the case of men answering the ‘‘real men’’ question was this difference not statistically significant). Compared to Black men and women, colored men and women had a trend of less favorable attitudes to concurrency but only in the ‘‘real men’’ variable was this statistically significant. A higher proportion of colored men and women than any other group answered that they ‘‘can’t trust anyone because people are cheating everywhere.’’ There was no difference in the proportions reporting ‘‘open talk’’ as to the dangers of concurrency in the different racial groups (47% to 55% answered yes to this question). In all the racial groups men exhibited more favorable attitudes to concurrency than women did. Men were twice as likely to say that extrapartner sex was okay. Women were also approximately 50% more likely to say that ‘‘one cannot trust anyone as people are cheating everywhere.’’ The ‘‘extrapartner sex is okay’’ variable is reported in more detail according to exact responses. It was not just the proportion of Black men and women who strongly agreed with the norm that ‘‘extra-partner sex is okay’’ that was higher but also the proportions who somewhat agreed and somewhat disagreed. In all, 54% of Black men versus 70.5% and 82.6% of colored and White men, strongly disagreed with this statement (p < 0.001). There was no difference between the races in the proportion who agreed that ‘‘people in my community are openly talking about the risk of HIV from having more than one partner.’’ Interpersonal Scripts Large and significant differences were evident in the percentage of each group reporting concurrency. The percentage of men reporting point concurrency varied between 14% in Blacks to 2% in Whites (p < 0.001). These percentages increased to 46% and 12%, respectively, for those reporting lifetime concurrency (p < 0.001). Intrapsychic Scripts There were statistically significant differences between the races for each of the three intrapsychic variables for men and women. The only exception to this was the ‘‘resist temptation’’ variable in the women, where there were no differences between the races. In each case Black respondents had the highest percentage reporting proconcurrency attributes, White respondents the lowest, and colored respondents intermediate. Pluralistic Ignorance
Cultural Scripts White men and women had less favorable attitudes toward concurrency than Black men and women did. These differences remained after controlling for a range
We found evidence that may be indicative of pluralistic ignorance in relation to the prevalence of concurrency. The proportion of Black, colored, and White women who reported that they ‘‘cannot trust anyone because people 881
KENYON, OSBAK, BUYZE, JOHNSON, AND VAN LANKVELD
Table 1. Demographic and Socioeconomic Characteristics of Black, Colored, and White Respondents in the National Communication Survey, 2009 Black
Downloaded by [University of California, San Diego] at 20:37 15 November 2015
Characteristics Gender Women Men Income Richest Middle Poorest Education Primary Up to grade 11 Grade 12 or higher Age 15–24 25–34 35–44 45–56 Married Single Noncohabiting relationshipa Married or living together Divorced=widowed
N
Colored
White
%
N
%
N
%
p Value
4,222 3,704
53.3 46.7
707 496
58.8 41.2
258 173
59.9 40.1
