By the ninth grade, more than one-third of adolescents have had sexual intercourse; by grade 12, nearly two-thirds are sexually experienced.1 Sexually active adolescents tend to have multiple partners (consecutively or concurrently) and to be inconsistent in their practice of safer sex.2 The combination of these factors places adolescents at risk for unplanned pregnancies and STDs, including HIV.
In the United States in 2000, more than 800,000 pregnancies occurred among women between the ages of 15 and 19.3 Additionally, adolescents and young adults (aged 20–24) account for between one-fifth and one-third of reported cases of syphilis, gonorrhea and chlamydia nationwide.4 Reducing or eliminating the problems associated with the early initiation of sexual intercourse and adolescents' inconsistent use of condoms and contraceptive methods requires an understanding of the aspects of young people's sexual experiences—including relationship contexts and the types of partners they have had—that may influence their decisions regarding their sexual behavior.
The majority of adult studies that distinguish among types of sexual partners include primary partners (e.g., spouse, main, steady, established, long-term) and one or more secondary relationship partners (e.g., side, casual, nonmain, new, anonymous, one-night stand). Although the manner in which these partner types are defined varies among populations and studies, significant differences emerge in behaviors with different partner types. In sexually active adult samples, the frequency of sexual events, the likelihood of specific sexual behaviors, the use of substances in conjunction with sex and the disclosure of relevant sexual risk information vary by partner type.5 Individuals are more likely to use condoms with casual or new partners than with main partners.6 Since condom use rarely is completely consistent even with casual or new partners, when main relationships are not monogamous, main partners are placed at risk of contracting STDs.
Research examining factors that affect adolescents' sexual decision-making with main and casual partners has identified different influences on behavior with different partner types. For example, the more adolescents value health, the stronger their intention to use condoms with casual partners, but not with main partners.7 Similarly, the importance placed on sex and intimacy operates in determining intentions to delay sexual intercourse with new main partners, but not with new casual partners.8
Differences in adolescents' behavior with main and casual partners may be partly related to the type of sexual partner they are making decisions with, but may also be related to differences between adolescents who choose different types of partners. Research examining patterns of sexual experiences (i.e., relative, or long-term, monogamy; serial monogamy; and nonmonogamy), but not types of partners, has identified significant demographic, psychosocial and behavioral differences among groups.9
In the current study, we sought to identify demographic and psychosocial differences among adolescents who report having had only main partners, those who report having had both main and casual partners, and those who report having had only casual partners. In particular, we examined differences between pairs of these groups and within the group who reported experience with both partner types. We also were interested in identifying differences in sexual behavior and behavioral intention with main partners between those who reported only main partners and those who reported main and casual partners; with casual partners between those who reported only casual partners and those who report both main and casual partners; and with main and casual partners among those who reported experience with both. We predicted that sexually experienced adolescents with different partner-type experiences would have statistically significant differences in demographic and psychosocial characteristics, as well as in behavioral and intention profiles.
Data for the current analyses were collected as part of a larger study examining perceived risk of STDs, perceived risk of pregnancy and sexual decision-making.10 The study was conducted in a San Francisco STD clinic between June 1996 and June 1998, using a protocol approved by the institutional review board at the University of California, San Francisco. Adolescents were recruited in the waiting room prior to clinician visits, and were eligible to participate if they were 14–19 years old, spoke English, had had vaginal or anal intercourse in the preceding three months and lived within the local metropolitan area. Of the 305 adolescents approached to participate, 276 (90%) participated.
After obtaining written informed consent, a research assistant conducted a structured interview with each participant in a private room and recorded the adolescent's answers on a standard form. (Because California law considers adolescents younger than 18 obtaining sex-related health services to be emancipated, informed consent from parents was not required.) Adolescents were offered compensation of $15 to participate in the interview.
Participants indicated their age, gender, racial or ethnic group, and mother's educational attainment (as a proxy for socioeconomic status).
We measured partner-type experiences by asking participants if they had ever had sex with each type of partner. We did not define "had sex" for participants, but defined partner types as follows: A main partner was described as "someone that you have sex with and you consider this person to be the person that you are serious about." A casual partner was described as "anyone that you have sex with but you do not consider this person to be a main partner to you. This person can be someone you've had sex with only once, or a few times, or you have sex with them on an ongoing, casual basis. The important thing, however, is that this person is not a main partner to you." Development of these definitions has been described previously.11 Results of pilot-testing ensured that these definitions were meaningful and distinct from one another.
We measured perceptions of risk of contracting an STD with two five-item scales, one referring to main and one to casual partners. Each scale asked participants how likely it is (with possible responses ranging from 1="not at all" to 5="extremely"), what the chances are (0–10), what the risk is (1="no risk" to 5="extra high risk"), how strongly they agree that they will (1="disagree a lot" to 6="agree a lot") and how strongly they agree that they will not (1="disagree a lot" to 6="agree a lot") get an STD if they have unprotected sex. For each scale, we summed the items to obtain a total perceived risk score (range, 4–32); higher scores indicate greater perceived risk of contracting an STD.
We assessed condom attitudes by creating composite scores from two existing multi-item scales reflecting participants' expectations and values regarding condom use.12 To measure expectations, we asked participants to rate, on a five-point scale (1="never" to 5="always"), the frequency with which consistent condom use would have 18 specific outcomes on their sexual health (i.e., prevent pregnancy and STDs), their relationship (e.g., make their partner angry) and various aspects of their well-being (e.g., decrease their sexual pleasure or make them worry less). To measure values, we asked them to rate, again on a five-point scale (1="very bad" to 5="very good"), how bad or good each of those outcomes would be. Scores from the expectations scale were multiplied by corresponding scores from the values scale to create the composite condoms attitude score. We then used the mean of these products to calculate four multi-item scales: perceived condom efficacy to prevent STDs and pregnancy (two items, alpha=.58), effect of condom use on trust in relationships (four items, alpha=.78), negative aspects of condom use (six items, alpha=.84) and positive aspects of condom use (six items, alpha=.77). Scores for each of the subscales could range from 1 to 25; higher scores indicate stronger condom attitudes.
Partner-specific condom self-efficacy was measured with existing six-item scales13 that ask participants to rate their confidence that they could insist on condom use in each of a variety of challenging situations (e.g., "Can you insist on condom use if a casual/main partner does not want to use one?" and "Can you continue to insist on condom use with a main/casual partner even if she/he gets angry when you suggest it?"). Responses were rated on a five-point scale (1="definitely no" to 5="definitely yes"); higher scores indicate greater self-efficacy to use condoms (alpha=.89 for main, .87 for casual).
Perceived social norms and compliance with those norms were measured through four items. Two asked participants whether most people who are important to them and whether their parents think that they should or should not use condoms in the next six months (possible responses ranged from 1="definitely should not" to 5="definitely should"). Two additional items assessed how often, in general, participants want to do what most people who are important to them and what their parents want them to (1="never" to 5="always"). Scores on the first items were multiplied by corresponding scores on the second items to obtain one score pertaining to most people and one pertaining to parents (range, 1–25 for each); higher scores indicate greater condom-supportive perceived social norms.
The Value on Health Scale14 assesses a number of domains of health that may be valued differently by adolescents. Participants were asked to indicate how important each of five items (e.g., having plenty of energy for everyday activities, feeling physically fit, staying in good health) is to them. Items were rated on a five-point scale (1="not at all" to 5="extremely"); scores were standardized and averaged, and had a possible range of 1–5 (alpha=0.81).
Pregnancy and STD histories were assessed by questions asking participants to indicate if they had ever been (or gotten someone) pregnant and if they had ever had an STD.
The importance of intimacy and sex in relationships was assessed by items asking participants how important (1="not at all" to 5="very") seven intimacy issues (e.g., sharing feelings with their partner) and four sexual relationship issues (e.g., having sex frequently with their partner) are to them. Two subscales emerged: importance of intimacy (alpha=.74) and importance of sex (alpha=.77). Higher scores indicate greater importance.15
Relative power within the relationship was measured through two items assessing which partner usually wins arguments and which has the final word in decision- making, and three items measuring who decides when the partners will see one another, what activities they will engage in and which friends they will see (-2=the respondent usually does to 2=the partner usually does). The five items were reverse-scored and were averaged to create a scale (alpha=.58). Higher scores indicate that participants have more decision-making power than their partners.16
Sexual behavior history was assessed by questions on how many partners of either type participants had had and the length of their most recent relationship. Relationship length was calculated in days from participants' responses.
Participants were asked whether they had used a condom and whether they had used a contraceptive during their last sexual encounter with their most recent main and most recent casual partners.
Participants were asked if they had had a concurrent sexual relationship during their relationship with their most recent main partner.
Past delay and delay intentions were assessed through three items for each partner type: "How long did you wait before having sex for the first time with your most recent main/casual partner?" "How long do you plan to wait before having sex for the first time with your next main/casual partner?" and "How long are you likely to wait before having sex for the first time with your next main/casual partner?" Answers for both past delay and intention to delay were 1="less than one day," 2="one day," 3="a couple of days," 4="one week," 5="two weeks," 6="three weeks," 7="one month," 8="two months," 9="three months" and 10="more than three months." For past delay, we split measures at their medians to determine "long delays" (more than two months for main partners; more than one month for casual partners). For delay intention, we averaged the scores for the second and third items.
Condom intentions were assessed by four items for each partner type. The items asked participants to indicate, on five-point Likert scales, how often they will use condoms (1="never" to 5="every time"), how likely it is that they will use condoms every time (1="not at all" to 5="extremely"), how sure they are that they will use condoms every time (1="very sure I will not" to 5="very sure I will") and how likely it is that they will not use condoms in the next six months (1="not at all" to 5="extremely"; reverse-scored). Higher scores indicate greater intention to use condoms (alpha=.90 for casual partner, .93 for main partner).
Intentions to have a side partner with future main partners were assessed through two items. Participants were asked how likely it is that they will have a side partner (either main or casual) at the same time as they are going out with a main partner in the next six months (responses ranged from 1="not at all likely" to 5="extremely likely"), and how sure they are that this will happen (1="very sure I will not" to 5="very sure I will").* A total score was calculated by averaging the responses to these items. Higher scores indicate greater intentions to have a side partner in the next six months.
We categorized participants into three mutually exclusive groups according to partner-type experiences: main only (52% of the sample), main and casual (32%) and casual only (16%). We then conducted a series of chi-square tests and one-way analyses of variance, as appropriate, to determine psychosocial variables that differentiate these groups. We also conducted post hoc analyses (Tukey's studentized range tests) to determine differences between pairs of groups, and paired t tests and McNemar chi-square analyses, as appropriate, to examine differences between partner-specific variables within the main-and-casual group. All analyses were conducted using SAS (version 8.2) software.
The participants ranged in age from 14 to 19; their average age was 17 (standard deviation, 1.41). Overall, 34% were black, 21% were white, 16% were Hispanic, 16% were Asian, and 14% were of mixed or other race or ethnicity. Whereas 39% of the adolescents' mothers had at most a high school education, 24% had attended some college and 36% were college graduates; 2% of participants were unaware of their mother's educational attainment.
Demographic and Psychosocial Differences by Partner-Type Experience
The gender and racial/ethnic distributions of participants differed significantly by type of partner experience (Table 1). Seventy-three percent of those in the main-only group were female, compared with 60% in the main-and-casual group, and 40% among those who had had only casual partners (Χ2=16.49, df=2, p=.0002). Given these differences, all subsequent analyses controlled for gender. The proportion of adolescents reporting "other" or mixed race or ethnicity was higher in the casual-only group than in either of the other groups (28% vs. 9–13%; Χ2=17.15, df=8, p=.03).
Participants who had had only main sexual partners reported greater perceived risk of contracting STDs from both main and casual partners than those who had had experience with both partner types. Compared with adolescents who had had only casual partners, those reporting only main partners had less favorable attitudes toward condoms' effect on trust in relationships, placed greater importance on intimacy and less on sex in relationships, and reported greater relative power in relationships. Additionally, those who had had only casual sexual partners indicated more favorable attitudes toward condoms' effects on trust in relationships and greater relative power than those who had had main and casual relationships; women in the casual-only group were the least likely to have been pregnant.
Those with sexual experience with both main and casual partners perceived themselves to be at greater risk of contracting an STD from a casual partner than from a main partner (t=–9.04, p<0.0001) and were more confident in their ability to use condoms with casual than with main partners (t=–5.47, p<0.0001).
Behavioral Differences by Partner-Type Experience
As Table 2 indicates, the average lifetime number of sexual partners reported differed significantly among groups (F=42.42, df=5, 270, p<.0001). There were no differences in condom or contraceptive use with main or casual partners. There were also no significant differences among groups in intentions to use condoms with future main and casual partners
A smaller proportion of participants who had had both main and casual partners than of those who had had only main partners reported long delays before engaging in intercourse with their most recent main partners (Χ2=12.24, df=2, p=.0023). Sixty-seven percent of the former group, compared with 1% of the latter, reported concurrent partners with their most recent main partners (Χ2=136.54, df=1, p<.0001); this difference was greater among males (81% vs. 0%) than among females (58% vs. 2%—not shown). Adolescents in the main-and-casual group intended a significantly shorter delay in initiating sex with a future main partner than those in the main-only group intended—one month vs. three months (Χ2=26.78, df=2, p<.0001). The intention to have side partners was greater in the main-and-casual than in the main-only group (F=14.28, df=5, 270, p<.0001). On average, the most recent main relationship did not differ in length between groups overall or for males, but among females, it was longer for those who had had only main partners than for those who had had both partner types (452 vs. 303; F=3.99, df=1, p=0.05—not shown).
Adolescents who had had only casual partners and those who had had main and casual partners differed in only one respect: The most recent casual relationship in the main-and-casual group was longer than the most recent relationship in the casual-only group (175 vs. 156 days—Table 2).
Participants with both main and casual partner experience reported different behaviors with each partner type. For example, their most recent casual relationship was significantly shorter, on average, than their most recent main relationship (t=1.78, p=0.007). They used condoms less often with their most recent main partners than with their most recent casual partners (Χ2=11.92, df=1, p=0.0006), and they had greater intention to use condoms with future casual than with future main partners (t=–8.03, p<0.0001). Additionally, 67% had side partners during their most recent main relationship.
Nearly half of the sample reported having had casual sexual partners; more than half of these had had both main and casual partners, but the remainder (16% of the total sample) had had sex only within casual relationships. Consistent with our prediction, adolescents with different partner-type experiences had different sexual attitudes, values, and sexual behaviors and intentions.
Those who had had main partners were primarily female adolescents, which is consistent with research showing that female adolescents are motivated to have sex by a desire for intimacy, while male adolescents are comfortable with sexual relationships in the absence of emotional closeness and are often driven by "enhancement motives" (i.e., a wish to improve their social status or conform to sexual norms).17 Females have also been found to be more likely than males to report relative monogamy patterns.18 The ethnic and racial disparities in the composition of the partner-type groups may also reflect different patterns of sexual socialization found in various cultural groups, although our findings reflected primarily that different proportions of each group identified as "other" or mixed race or ethnicity.
A number of the psychosocial differences between the partner-type groups point to potential protective and risk factors associated with partner-specific sexual decision- making. Adolescents who reported only main partner experience scored relatively high on importance placed on intimacy in relationships and on relative power in their relationships. Since both of these factors are correlated with protective behaviors (e.g., intentions to use condoms, condom use and delaying sex in new relationships) among sexually experienced adolescents,19 they may be important strengths to capitalize on in risk reduction interventions with adolescents who report only main partnership experience. Another area of intervention or education might include assisting those who have had only main partners to accurately assess their risks of contracting an STD from a main partner. Previous work indicates considerable variability in the accuracy of adolescents' perceptions of their main sex partners' risk behaviors: More than a third of one sample of sexually experienced adolescents incorrectly reported that their main partner had not engaged in a risk behavior,20 and only a quarter of those in another sample were aware that their partner had had concurrent partners.21 It should be noted that a small proportion of adolescents in our sample who reported experience with only main partners had had concurrent relationships during their most recent relationship (1%). Some of these concurrent relationships may represent the transition between main partnerships that occur among those who practice what they might consider serial monogamy.22
Adolescents who had had only casual partners placed less importance than others on intimacy in relationships. This can be thought of as a potential risk factor for this subgroup. Our results, in conjunction with previous findings,23 indicate that they would benefit from increasing the importance they place on intimacy in their sexual relationships. Several findings point to protective factors among those who had had only casual partners: Compared with the other groups, they reported more positive attitudes toward condoms' effects on trust in relationships and had a smaller proportion of females reporting pregnancy experience; compared with the main-and-casual group, they reported higher relative power in relationships. Strengthening their positive attitudes toward condoms and their sense of power in relationships, and capitalizing on their strategies to avoid pregnancy, will aid in reducing and preventing risky behavior in this group. The finding that those with only casual partner experience were less likely than others to have experienced a pregnancy is consistent with results of previous work demonstrating that those who consider themselves nonmonogamous are less likely than others to have experienced a pregnancy.24 However, those who indicate that they have had only casual sexual partners are not necessarily the same as those who consider themselves nonmonogamous, since one could engage in a series of relationships with casual partners and not have any concurrent relationships.
An interesting risk group that emerged from these data are adolescents who had had both main and casual partners. Given their reports of concurrent sexual partnerships (especially among males) and intentions to have side partners, coupled with their potentially inaccurate assessment of STD risk from sexual partners and their feelings of low relative power in relationships, this group is an appropriate target population for risk reduction interventions. Notably, we would not be aware of the special risks associated with this subgroup if we had grouped them with adolescents who had had only one type of partner. Previous work examining sexual experiences and partnerships among low-income minority youth also indicated that "the kind of partner and whether the individual has multiple partners affects condom use and should be included in analyses of…sexual behavior."25
Our findings should be interpreted in the light of a number of study limitations. Our sample was a sexually experienced adolescent group who attended an urban STD clinic in an AIDS epicenter. Consequently, our results may not generalize to adolescent populations who are not yet sexually experienced, who have access to more health care resources (and would, therefore, not attend a municipal STD clinic in a state where "sensitive services" can be obtained without parental consent at any health care setting) or who live in other geographic areas. Given the different sizes of the partner-type groups, our ability to detect significant differences may have been compromised. Although we found significant ethnic and racial differences among groups, given our sample size and makeup, we were unable to control for the effects of race and ethnicity on other variables or to account for the effects of gender within racial or ethnic groups.
We also acknowledge that the definitions of sexual partner type (main and casual) may be simplified and that the category of casual partnerships includes a variety of possible relationship contexts (e.g., one-night stands, flings or sexual acquaintances). However, qualitative and quantitative research26 suggests that the concept of main partners is robust among adolescents and that differences in attitudes, expectations and values can be found in the dichotomy of main and nonmain sexual partner types. Additionally, under our definition of side partner, participants could report concurrent sexual behavior with two main partners; however, our finding that only a small proportion of participants who reported concurrent relationships indicated this situation may be seen as evidence of the robustness of our definition of main partners. We also acknowledge that adolescents who engage in concurrent relationships during a main relationship may differ from those who engage in serial relationships with different types of partners, and both may fall into the main-and-casual group; we were unable to address this issue in the current study.
To date, little research has focused on partner-specific sexual decision-making among adolescents, and no intervention strategies include material distinguishing sexual behavior and decision-making with main sexual and casual partners, or attempt to tailor their messages to adolescents with different partner-type experiences. And in general, few interventions (even those assessed in the most recently published studies)27 use partner-specific evaluations of outcomes.
Interventions should build on previous theory-driven strategies for developing sexual risk reduction interventions,28 and should include discussion of risk-related factors that differ between adolescents who have had main partners and those who have had casual partners. They also should include direct discussion of differences in sexual decision-making and behaviors with main and new or casual partners. When partner-type experience and partner-specific factors are considered in designing sexual risk reduction interventions, research on their effectiveness can be developed that assesses the best manner of delivering relevant information, increasing motivation and teaching behavioral or communication skills. Evaluation of the effects of interventions also ought to utilize partner-specific assessment of attitudinal and behavioral outcomes.