ADOLESCENT IDENTITIES AND SEXUAL BEHAVIOR: AN EXAMINATION OF ANDERSON’S ‘PLAYER’ HYPOTHESIS

Abstract

We investigate the social and behavioral characteristics of male adolescents who self-identify as players, focusing particularly on Anderson’s claim that this social role is inextricably linked with poverty and minority status. Results indicate that African American respondents, those affiliated with liberal peers and young men who initially report a relatively high number of sexual partners are more likely to resonate with this identity label. Nevertheless, analyses reveal that a number of players within the sample are not disadvantaged African American youth, and there is considerable variability in their attitude and behavior profiles. Findings based on longitudinal analyses indicate that the player identity is a significant predictor of later variations in self-reported sexual behavior, net of traditional predictors, including prior behavior. Yet results of in-depth interviews conducted with a subset of the respondents complicate these quantitative findings, highlighting that young men’s perceptions of this identity are not as uniformly positive as Anderson’s depiction might lead us to expect.

Rosenberg (1981) famously argued that the self is both a “social product” and a “social force,” general observations that have received more systematic treatment within structurally oriented versions of the symbolic interactionist perspective (Burke 1980; Hunt 2003; Stryker 1980). In this analysis, we focus on adolescent sexual identities—specifically Anderson’s depiction of the player role/persona, a social type he observed in ethnographic studies of life in an inner-city neighborhood in Philadelphia. Anderson (1989) hypothesized that poverty among African American youth encourages a view of sex as a means of masculine-identity construction instead of as an integral component of a romantic relationship. He argues that among disadvantaged males, the peer group de-values relationship qualities such as love and commitment, and instead promotes the view of sex as a game, where women are tokens and the competition is against other males to gain social status.

According to Anderson, the young man who has sex with many partners gains a reputation as a ‘player,’ an identity that is cultivated due to its many positive rewards. Consistent with Rosenberg’s (1981) dictum, Anderson highlights the ways in which the basic parameters of identity have been shaped by broader structural conditions, but in turn influence the young men’s attitudes, lifestyle, and behavioral choices. Although Anderson’s ideas are well known, few studies have directly explored this identity status. Researchers have examined other social identities that develop during the adolescent period (e.g., ‘nerds,’ popular youth, brains—Eder 1985; Garner et al. 2006; Youniss, McLellan, and Strouse 1994). Our focus adds to this research tradition, however, as interest in and involvement with the opposite sex is considered a defining feature of adolescence (Sullivan 1953). More pragmatically, focusing on sexual identities such as the player is potentially important because the constellation of attitudes described (e.g., disingenuous feelings toward one’s romantic/sexual partners) and behaviors hypothesized to be associated with them (large numbers of sexual partners) may place such individuals and their partners at higher risk of HIV, STDs, and unplanned pregnancy when compared to their more typical adolescent counterparts who either have not had sex at all or who have more limited experience with a stable partner (see Carver, Joyner, and Udry 2003 for an overview of patterns of sexual behavior within a national probability sample of adolescents).

The current study relies on structured survey and qualitative data drawn from a four wave longitudinal study of adolescent romantic and sexual experiences, the Toledo Adolescent Relationships study (TARS) (n = 1,114; 538 are males). We first investigate the social characteristics of male adolescents who self-identify as players in order to evaluate Anderson’s claim that this social role is inextricably linked with poverty and minority status. These data are a useful follow-up to Anderson’s ethnography, as the TARS sample includes but is not limited to disadvantaged adolescents.

Next we examine attitude and behavioral repertoires that connect to the player identity. As an example, Anderson suggested that the player cares more about the opinions of his male peers than about heterosexual partners, but such attitudes have not been systematically investigated using a large scale survey. Similarly, we expect that the behaviors of youths who consider themselves players will include greater sexual risk-taking, but this has not been empirically established. Because Anderson hypothesized that a primary goal of the player is to have sex with as many young women as possible, we will assess the degree to which endorsement of this identity is associated with: (1) an earlier age at first sex; (2) higher numbers of sexual partners; (3) reports of concurrency (cheating); and (4) greater likelihood of engaging in ‘hook-ups’ (non-dating sexual behavior) or one-night stands.

Relying on a longitudinal assessment (n = 481), we investigate whether endorsement of the player identity predicts the number of sexual partners subsequently reported by these respondents. Much previous research has documented that past behavior is generally a solid predictor of future behavior. Thus, the respondent’s level of prior sexual experience will likely be strongly related to later number of sexual partners reported. And, consistent with this, young men who engage in player-like behavior should be more likely than others to acquire a player reputation. However, we expect to observe a less than perfect correspondence between these behaviors and this identity that potentially but not inevitably connects to them. Thus the longitudinal design will allow us to determine whether variations in degree of endorsement of the player identity itself makes a difference for understanding variations in sexual behavior, once prior number of sexual partners and traditional predictors of sexual risk-taking have been taken into account.

We suggest that finding an effect of identity on subsequent sexual behavior, net of other predictors, would provide empirical support for the symbolic interactionist emphasis on self processes, including the latter half of Rosenberg’s dictum—the “self shapes social behavior.” A contrasting perspective, however, conceptualizes self processes as largely epiphenomenal; that is, identity statuses are not viewed as particularly consequential, but merely provide labels or names to describe the behaviors involved (see e.g., Bem 1978). According to this logic, the behavior results in application of the label, while identity and labeling processes themselves are seen as having little effect on [sexual] behavior. A more contemporary symbolic interactionist approach attempts a reconciliation of these two conceptual points of view: sexually permissive behavior (itself heavily influenced by social forces) is likely to be a key but not the singular determinant of initially being considered a player type; once acquired, however, this labeling potentially has consequences for self-views, the nature of adolescents’ “hierarchies of salience,” as well as for the reactions of and treatment by others (Stryker 1981). These self and social processes serve as concrete mechanisms linking identity and subsequent behavioral choices.

As a supplement to these quantitative analyses, a final goal is to describe results of an analysis of in-depth qualitative interviews completed with a subset of the respondents who participated in the TARS study. Excerpts from the narrative data are included in order to further explore Anderson’s ideas, illuminating aspects of the player phenomenon that are more difficult to capture via typical survey methods. While the structured questions tap social dimensions of identity (i.e., whether other people would describe the youth as a player), the unstructured interviews highlight ways in which adolescents themselves give meaning to this label. The qualitative analyses support the quantitative results, but provide a somewhat more nuanced view of boys’ perspectives on the player identity. These results also fit well with a symbolic interactionist emphasis, as they point out the role of cognitions, agency, and meaning construction in the connection between labeling and identity formation processes (Musolf 2003).

Background

Research on adolescent sexual behavior has frequently focused on parent and peer influences (Borawski et al. 2003; Huebner and Howell 2003; Sieving et al. 2006) and there is also increased interest in the ways in which dating experiences influence sexual decision-making (e.g., Coontz 2006; Florsheim 2003; Giordano, Manning, and Longmore 2006). Research clearly demonstrates that adolescents are influenced by interaction and communication within these important social domains. Nevertheless, adolescents must eventually develop their own ways of relating to the opposite sex, make decisions about potential partners, and actually participate in these romantic and/or sexual liaisons (Giordano et al. 2006; Simon and Gagnon 1986). This observation may be especially important where the focus is upon adolescent sexuality, because these actions are not yet routinized and generally occur outside the immediate purview of these other reference groups. Our view, then, is that identities and the behaviors that relate to them are influenced by social factors such as friends’ attitudes toward sex, or location in the social structure (e.g., poverty), but cannot be reduced entirely to these social influence processes. Instead, these social experiences and other individual qualities and biographical characteristics combine to create self-views that vary considerably in their character and in their effects.

One aspect of the self-concept that has been studied extensively is adolescent self-esteem (Baumeister et al. 2003; Rosenberg 1979). This construct focuses on the individual’s overall sense of worth, and while early on scholars hypothesized that low self-esteem should be associated with earlier sexual debut, research has generally documented that this global evaluative dimension is not a strong predictor or consequence of sexual behavior (see the review by Goodson, Buhi, and Dunsmore 2006). Consistent with our focus here, more recent theory and research on self processes has shifted from an interest in global representations of the self to the goal of understanding the content of adolescents’ social identities (Gecas and Longmore 2003; Matsueda 1992). While the emphases within the symbolic interactionist tradition provide a general theoretical basis for exploring adolescent identities and sexual behavior, little research in this area has focused on particular ‘content’ areas (for an exception see Buzwell and Rosenthal 1996). Anderson’s qualitative research on the player identity is thus an important exception.

Anderson’s Player Hypothesis

Anderson (1989) clearly linked the player role, as briefly described at the outset, to the conditions of inner city life, particularly for young African American males. This general notion thus fits well with the idea that the self is indeed “a social product” (see also Coates 1999; Majors and Billson 1992). Anderson argued that lack of access to meaningful employment paths and careers and a general lack of opportunities for success along traditional lines serve to heighten the emphasis on ‘scoring’ with women as a way of demonstrating competence with one’s peers:

To an inner-city black male youth, the most important people in his life are members of his peer group. They set the standards for his conduct, and it is important for him to live up to those standards, to look good in their eyes. The peer group places a high value on sex, especially what many middle-class people call casual sex…Thus a primary goal of the young man is to find as many willing females as possible… the young man must also prove he is getting it. Consequently, he usually talks about girls and sex with every other young man who will listen (Anderson 1989:61).

Anderson observed these attitudes among young males living in an economically disadvantaged neighborhood in Philadelphia. However, it has not been established whether such attitudes are found only within disadvantaged contexts and/or among young African American males. For example, some other ethnographic studies of predominantly white, middle class youth have suggested that the male peer group fosters attitudes that are strikingly similar to those Anderson described, and have also posited that the opinions of male friends matter more than relationships formed with young women (e.g., Eder, Evans, and Parker 1995; Kimmel 1994; Wight 1994). Further, Anderson himself did not argue that this pattern was ubiquitous in the area he studied, recognizing that some boys did have caring feelings for their girlfriends. Nevertheless, the argument as developed and popularized hints at the predominance of this player orientation in disadvantaged neighborhoods, as reflected in Anderson’s observation that boys who did express caring sentiments were often subject to teasing or ridicule from friends. This suggests the need to explore more systematically: a) the distribution of the player identity across a more diverse sample of adolescents, b) the attitude and behavior profiles of those who have been labeled as player types, and c) whether the player identity contributes to an understanding of variations in the nature of young men’s sexual experiences, beyond the impact of their earlier behavior and other traditional predictors. Below we describe in more detail the perspective that serves as a theoretical framework for the quantitative and qualitative analyses.

Theoretical Perspective

Self as Social Product

Adding to early symbolic interactionist emphases, researchers such as Stryker (1981) have highlighted that while the self is developed through interactions with others (Mead 1934), the actor’s structural locations: a) constrain/channel the individual’s network of affiliations and b) shape the messages these others impart about what is possible, desirable or to be avoided (see also Harris 1977). Thus, it is not only employment opportunities and health care that vary by socio-economic status, but the range of “possible selves” as well (Markus and Nurius 1986). This general notion accords with recent research on the self that has forged specific links between larger systems (the stratification order, gender schemas) and identity formation processes (Brown 1999; Callero 2003; Gecas and Burke 1995; Hunt 2003), and is consistent with Anderson’s claim that conditions within disadvantaged neighborhoods foster the player ethos/identity.

While this line of theorizing and research has added much to the symbolic interactionist perspective, these elaborations were not meant to detract from more basic, distinctive features of the theory—specifically the view of the individual as an active agent:

Not only do we open the door to certain stimuli and close it to others, but our attention is an organizing process…. Our attention enables us to organize the field in which we are going to act. Here we have the organism as acting and determining its environment. It is not simply a set of passive senses played upon by the stimuli that come from without. The organism goes out and determines what it is going to respond to and organizes that world (Mead 1964:138–39).

Thus according to Mead and contemporary symbolic interactionists, mechanisms involved in identity formation include cognitive processes, human agency and elements of creativity. Accordingly, the individual does not passively receive messages from peers, but plays an active role in filtering and reacting to these communications. Corsaro and Eder emphasized this general dynamic in their elaboration of an interpretive perspective on childhood and adolescent peer cultures (Corsaro 1985; Corsaro and Eder 1990; Eder and Nenga 2003; see also Vygotsky 1978). They note that when interacting and communicating with friends, children draw on elements of the larger culture (including parental socialization efforts), but they do so creatively, inevitably constructing new peer worlds that are distinctive in multiple respects. A similar point follows in exploring links between these peer cultures and identity formation processes. As Anderson suggested, young men may develop peer norms that reward “scoring” with women, playing the field and the like, but individuals likely react to these themes selectively, considering such peer pressures alongside divergent messages from others in their networks (e.g., parents, girlfriends, minister), reflecting on their own experiences (e.g., enjoyment of the emotions associated with being in love), and other aspects of their personal biographies (whether shy or outgoing, sufficiently good looking to attract the attention of many young women and the like).

A second reason for expecting that these identities are not likely to be ubiquitous--even within disadvantaged contexts--is that the existence of the label itself connotes an element of distinctiveness to this orientation and behavior pattern. Were these features viewed culturally as routine and taken for granted, it is not likely that they would warrant a specific designation. Highlighting the idea of heterogeneity in young men’s sexual identities, we quote from a pre-test focus group session conducted at a high school serving a disadvantaged area of Toledo (Giordano, Longmore, and Manning 2001). In connection with a lengthy discussion of the respondents’ romantic and sexual experiences, Jake, a junior, contrasted his own orientation with that of a friend: “it’s a one-girl man right here [speaking of himself] and a several girl man over there. He’s [pointing to another young man participating in the same focus group] just an indiscriminant pipe layer. He just lays pipe for a living.” This example suggests that heterosexual identities are a recognized aspect of young men’s self-concepts, but does not indicate a uniform orientation with respect to these aspects of the self.

Self as Social Force

In addition to an emphasis on the social origins of the self, the symbolic interactionist position highlights that self processes themselves are consequential for understanding human behavior (Stryker 1959). Labeling theory, derived from the symbolic interactionist tradition, often focuses on topics such as mental illness, crime and other deviant behavior, emphasizing that once a label has been applied, this results in a range of self and social dynamics that matter for understanding the individual’s subsequent actions (Lemert 1975). The central notion is that the “reflected appraisals” of others may foster labels such as mentally ill, slow learner, troublemaker, or player, but the application of the label influences the subsequent reactions of others, social opportunities and constraints, and self-reflections. One of the important critiques of early labeling arguments was the tendency to focus little on the degree to which the actor’s initial behavior acted as a catalyst for these labeling processes. As Akers (1967: 46) noted, it was almost as if the individual were “minding his own business,” when suddenly he was ‘slapped’ with one of these labels. Thus it is critical to consider that mentally ill individuals, for example, do frequently engage in aberrant behaviors, the slow learner struggles in school, the troublemaker has participated in acts of delinquency, and young men who have sex with multiple partners may acquire a reputation as a player. In short, the most straightforward route to a distinctive label of this kind is undoubtedly engaging in behaviors that connect to the label in question.

Nevertheless, the symbolic interactionist perspective continues to problematize the “behavior causes the label” assertion, emphasizing the degree to which labeling processes and associated identity statuses are: a) social, b) subjective, and c) multifaceted. In the case of the player identity, others who describe a young man as a player have generally not witnessed the individual’s sexual behavior (although to be sure they may have heard his stories and seen him “talking it up” with many female acquaintances). Accordingly, some young men may gain a reputation that is not backed up by their sexual behavior, and more discreet youth may engage in sexual behavior that is not widely recognized in their social circles (see e.g., Becker’s (1963) discussion of the ‘falsely accused’ and the ‘secret deviant’ respectively). The subjective, multifaceted nature of this identity status is implicated because the player is generally associated with a cool, confident demeanor (Majors and Billson 1992), being sufficiently good-looking to attract the attention of members of the opposite sex, as well as with the expectation of sexual permissiveness. Thus, for example, it is possible that some confident, flirtatious young men have a social visibility that fosters the label, even though they are not ‘outliers’ in terms of their sexual behavior.

Recognizing the subjective, social nature of the labeling process, symbolic interaction theorists have nevertheless focused central attention on the consequences of these attributions for the individual’s life chances, views of self and subsequent behavioral choices. Indeed, this can be considered the key contribution of this theoretical perspective. Numerous studies have documented that labels, once imposed, influence the way others respond to the labeled individual (Link et al. 1989; Schur 1971; Stryker 1959). As an example, Pager (2003), recently documented, using an experimental design, that similarly qualified men were significantly less likely to be hired for a job if their application included mention of a prior felony conviction (see also Bernburg et al. 2006; Schwartz and Skolnick 1962). Other studies have focused on the effects of informal labeling, which can occur in connection with or independent of formal labeling processes. Eder (1985) explored the ways in which young women were viewed and treated by others in a middle school once they had been identified as members of the popular crowd. And consistent with our approach in the current study, Matsueda (1992) demonstrated that adolescents’ beliefs that their parents viewed them as “rule-violators” influenced later patterns of self-reported delinquency, even after initial levels of delinquency involvement had been taken into account.

Identities have a slippery, elusive quality, but play an important organizing role for the individual, as they can be viewed as a kind of cognitive filter that influences decision-making (Giordano et al. 2002; Matsueda 1992). This is especially important as the individual moves into the future and inevitably encounters novel circumstances. Situations defined as real, are real in their consequences not only because of the way others treat the individual so labeled, but because of the social underpinnings of the self-concept (Mead 1934; Thomas 1928). While some labels restrict social opportunities and increase feelings of stigma (e.g., the felon status, being labeled a special education student), other labels amplify opportunities in a particular direction. The player, as an outgoing, visible character within the school or neighborhood social scene, may be invited to more parties, continue to attract the attention of female companions, and (as Anderson suggested), find much support and reinforcement from male peers. In turn, these social opportunities may influence the individual’s “hierarchy of salience,” such that other less immediately rewarding pursuits (e.g., homework) assume less importance relative to partying and heterosexual socializing (Stryker 1981). To the degree that positive meanings are associated with the latter, this could strengthen intangible self-views such as the young man’s confidence, and in turn commitment to these lines of action.

The Current Study

Relying on a large diverse sample, we first examine the distribution/patterning of endorsement of the player identity. Is the player persona a social identity primarily adapted by disadvantaged minority males, as Anderson hypothesizes? We examine predictors of the identity such as poverty and minority status and determine how often male adolescents not living in poverty resonate with this label. Subsequently, for those respondents who do identify as players, we examine specific attitudes that Anderson hypothesized to be associated with such a self view. Our analyses compare responses of self-identified players with those of male respondents who do not endorse this identity status on: a) level of agreement about the game aspect of heterosexual socializing (i.e., enjoying the chase more than the relationship), b) degree to which such individuals frequently talk to their friends about their sex lives, and c) whether such youths are more likely to agree with the notion that friends are more important than girlfriends.

In analyses that explore the behavioral realm, we assess whether the player status is associated with: a) earlier age at first sex, b) higher number of sexual partners, c) cheating (concurrency), d) sex outside a traditional dating context and e) “one night stands.” Multivariate models are then estimated, where the focus is on number of sexual partners reported at wave four of the study, as influenced by earlier predictors (e.g., liberal peer attitudes), prior behavior (number of sexual partners reported at wave 3), and player identity. This longitudinal assessment allows us to determine whether the level of endorsement of the player identity contributes to knowledge about variations in sexual behavior once traditional predictors, including prior behavior, have been taken into account.

Our view is that a relationship between endorsement of this identity and sexual behavior in models controlling for social and demographic characteristics and prior behavior suggests a substantively distinct contribution of this identity status to variations in the nature of young men’s sexual experiences. In addition, these analyses allow us to determine whether taking into account respondents’ identification with the player role mediates any observed association between race/ethnicity and sexual risk-taking.

Anderson’s ideas were based on interviews with African American men who resided in inner city Philadelphia. Thus, as a final step in the analysis we examine the possibility of a differential impact of player status on the sexual behavior reports of African American, Hispanic, and white youths. Although Anderson did not state this explicitly, his extensive discussions suggest the possibility of a unique influence of this identity on the life course experiences of African American youths. Our analysis of the in-depth narratives elicited from a subset of the respondents sheds additional light on Anderson’s player hypothesis, by highlighting ways in which such youths understand and give meaning to their identities, as well as to their sexual choices.

Data and Method

Data

This study uses survey and narrative data from the Toledo Adolescent Relationship Study (TARS). In 2001, individual interviews were conducted with a stratified random sample of over 1,316 adolescents drawn from the year 2000 enrollment records of all youth in the 7^th , 9^th , and 11^th grades in Lucas County, Ohio. The sampling frame encompassed 62 schools across seven school districts. School attendance was not a requirement for participation in the study, and most interviews took place in respondents’ homes. Adolescents entered most responses directly on preloaded laptop computers. The sample includes oversamples of African American and Hispanic adolescents. In addition to wave 1 data, we rely on items from the third and fourth waves of interviews completed with the majority of these respondents. Wave 3 was completed in 2004–5 with 1,114 respondents or 84 percent of the original sample, and wave 4 approximately two year later (2006–7). There were 538 male respondents in wave 3. The analytic sample for cross-sectional analyses is all male respondents (n = 532) who had valid data on key focal variables. The longitudinal component focuses on male respondents who had valid data on number of sexual partners as reported at wave 4 (n = 481, or 92 percent of the wave 3 male respondents). TARS data also includes information drawn from responses to a questionnaire completed by a parent at wave 1 (typically the mother). This is the source for variables indexing the parent’s characteristics and behaviors, in this analysis limited to the item measuring parent’s education.

In addition to the survey data, we also draw on excerpts from in-depth interviews conducted with a subset (n = 51) of the male respondents who participated in the wave one structured interviews. The in-depth interviews were generally scheduled separately from the structured interview. Areas covered in general parallel the structured protocol, but allow a more detailed consideration of respondents’ complete romantic and sexual histories.

Measures

Player identity. At each wave of interviews, respondents were asked about a number of identity statuses (e.g., smart, athletic) and the degree to which others would describe them in this manner (see especially Matsueda 1992). At wave 3 an item was added to the protocol that focused on the degree to which “other people would describe [me] as a player.” The responses range from (1) strongly disagree to (5) strongly agree. Using this measure, 21 percent of our analytic sample, or 112 respondents, agree or strongly agree that people would describe them in this way.

Demographic and Social Controls¹

Race/ethnicity is self-reported. For present analyses, race/ethnicity is a four category variable coded as: white non-Hispanic (62.3 percent), African American (23.6 percent), Hispanic (10.7 percent), and ‘other’ (3.5 percent). In analyses, race/ethnicity is dummy coded with white serving as the reference category.

Age is calculated as the date of birth subtracted from the date of the interview. Respondents’ average age at wave 1 is 15.4 and the range is 12–19 years. Wave 3 average age is 18, and the range is 15–22.

High poverty neighborhood is calculated using census data. It is a dummy variable indicating whether the respondent lived, at the time of the first interview, in a neighborhood where greater than 20 percent of the neighborhood lived below the poverty level (25.9 percent).

Mother’s education refers to the highest educational level achieved by the respondent’s mother, as reported by the mother at the time of the first interview. If there is only one parent figure in the household, his or her education level is used. The response categories are less than 12 years of education (11.7 percent), high school graduate (30.9 percent), and more than 12 years (57.4 percent).

Parental monitoring is a summated five-point scale drawn from the National Longitudinal Study of Adolescent Health protocol in which respondents indicate how often parents let them make their own decisions about: (1) “the time you must be home on weekend nights;” (2) “the people you hang around with;” (3) “what you wear;” (4) “your social life;” (5) “who you can date;” and (6) “how often you can date.” High scores reflect high parental monitoring. Scores range from 6 to 30 and the mean is 22.7 indicating perceptions of high parental monitoring (alpha = .86).

Family structure is a four category variable that indicates the teen’s report of family living circumstances at the time of the first interview. The responses include both biological parents (56.5 percent), a single parent family (21.9 percent), step-parent family (12.8 percent), or some other family type (8.8 percent).

Commitment to school is indexed by the item “Good grades are important to me.” The responses range from (1) strongly disagree to (5) strongly agree. The mean value is 4.25.

Liberal peer attitudes is based on the question “My friends think it’s okay to date more than one person at a time.” A dummy variable is coded 1 if respondent agrees or strongly agrees and 0 otherwise. The percent of those who agree or strongly agree is 36 percent.

Player Attitudes

Measures of player attitudes, measured at wave 3, refer to the extent to which respondents agree or disagree on a five point scale with three items, including, Enjoy the chase: “When it comes to girls, I enjoy the chase more than the relationship” (20.3 percent agree or strongly agree). Talk to friends about sex often: How often do you talk to your friends about your sex life?” (30.9 percent agree or strongly agree) and Friends come before girlfriends: “Girlfriends come and go, but friends are always there for you” (70.7 percent).

Sexual Behavior

Age at first sexis measured by asking respondents at wave 3 whether they have ever had sexual intercourse, and if so, the age they had sex for the first time (mean age = 15.9 years).

Number of sexual partners in past 24 months At waves 3 and 4 we ask respondents: “How many sexual partners have you had in the past 2 years or 24 months? At wave 3 respondents report having an average of 2 .1 sexual partners in the past 24 months and at wave 4 2.9 sexual partners in the last 24 months. Respondents were provided an explicit definition: “Have you ever had sexual intercourse? (sometimes this is called "making love," "having sex," or "going all the way"),” in order to avoid a count that included other forms of sexual intimacy.

Number of non-dating sexual partners asks wave 3 respondents the number of sexual partners they have had in the past 24 months outside a traditional dating context. The wording of this section is identical to that used in the Add Health survey. The mean is 1.4.

Cheated is measured by the following question “Since your relationship started, how often have you gotten physically involved ("had sex") with other girls/guys?” Nearly one-fifth (20 percent) cheated or had concurrent sex in the last 24 months.

One night stands asks respondents the number of one night stands they have had in the past 24 months. Nearly one-quarter (23.2 percent) report having had a one night stand.

Analytic Strategy

We first calculate descriptive statistics for the sample as a whole, and for players and non-players separately. For this analysis we include as players those respondents who agree or strongly agree that others would describe them in this way based on the wave 3 report. Next we rely on logistic regression models to predict the odds of reporting a player identity. We estimate zero order and then full models that include sociodemographic characteristics, family factors, parental monitoring, commitment to school, liberal peer attitudes and prior behavior (previous sexual partners reported at wave 1). We present mean and percentage distributions of player attitudes (e.g., agrees to enjoying the chase more than the relationship) and behaviors (e.g., one night stands in past 24 months) for those who do/do not report having a player identity. Next, ordinary least squares are estimated predicting number of sexual partners at the time of the wave 4 interview. We include wave 3 sexual behavior as well as key control variables. We estimate models with the five-scale player identity measure to determine whether player identity is associated with reporting a higher number of sexual partners, net of these other characteristics and prior behavior. As a final step in these analyses, we estimate a series of interactions of age and player identity, race/ethnicity and player identity, and prior sexual behavior and player identity. The purpose of these interactions is to determine whether the effect of player identity status varies according to the respondent’s age, race/ethnicity, or is conditional upon level of prior sexual behavior reported.

Results

Table 1 presents the descriptive statistics for all variables included in the analyses. One fifth (21 percent) of all male respondents either agree or strongly agree that people would describe them as players. Table 2 presents results of logistic regression models that explore factors associated with endorsement of the player identity, and includes zero order and multivariate findings. Zero order results indicate that prior behavior (number of previous sex partners) is significantly related to the odds of being labeled a player. African American male youth in the sample are more likely than white respondents to agree that people they know would characterize them as a player, and residing in a disadvantaged neighborhood is also related to endorsement of this identity. Non-traditional family structure (single-parent and ‘other’), and liberal peer attitudes are also significantly associated with the player identity at the zero order. We expected older youth to be more likely to identify with the player status but we do not find a statistically significant relationship with respondent age.

Table 1.

Descriptive Statistics for the Total Sample and Separately for Players and Non-Players [about here]

	Total		Players		Non-players
	Mean/%	S.D.	Mean/%	S.D.	Mean/%	S.D.
Identity as a ‘Player’ (W3)	21.38%		--		--
Sociodemographic Characteristic s
Race/Ethnicity (W1)
White	62.26%		37.61%		68.96%
African American	23.63%		48.34%		16.91%
Hispanic	10.66%		11.19%		10.51%
Other	3.46%		2.86%		3.62%
Respondent's Age (W3)	18.14	1.61	18.29	1.69	18.10	1.64
Disadvantaged Neighborhood (W1)	25.86%		41.67%		21.57%
Family Characteristics
Mother's Education (W1)
Less than 12 years	11.66%		17.18%		10.16%
High School	30.95%		35.78%		29.46%
More than 12 years	57.39%		46.41%		60.38%
Parental Monitoring (W1)	22.68	5.77	22.60	6.24	22.70	5.54
Family Structure (W1)
Single Parent	21.92%		32.10%		19.15%
Married
Biological/Adoptive Parents	56.45%		43.42%		59.98%
Stepparents	12.82%		11.68%		13.14%
Other	8.81%		12.80%		7.73%
Commitment to School
(W1)	4.25	.69	4.18	.85	4.27	.82
Liberal Peer Attitudes
(W3)	35.36%		61.41%		28.28%
Prior Behavior
Previous Sexual Partners (W1)	.80	2.07	1.71	3.07	.56	1.64
Previous Sexual Partners
Last 2 Years (W3)	2.13	3.92	4.64	6.46	1.45	2.26
Player Identity
(Interviewer Rating) (W1)	10.49%		18.06%		8.41%
N	532		111		421

Source: Toledo Adolescent Relationship Study, Waves 1 and 3

Note: Weighted distributions.

Table 2.

Logistic Regression Models Predicting Log Odds of Player Identity [about here]

	Zero Order	Model 1
Prior Behavior (Previous Sexual Partners) (W1)	1.23***	1.14*
Sociodemographic Characteristics
Race/Ethnicity (W1)
(White)
African American	4.42***	2.88***
Hispanic	.87	.62
Other	2.02	1.55
Respondent’s Age (W3)	1.05	1.00
Disadvantaged Neighborhood (W1)	2.18***	.84
Family Characteristics
Mother's Education (W1)
Less than 12 years	1.29	1.00
(High School)
More than 12 years	.65	.73
Family Structure (W1)
Single Parent	2.13**	1.20
(Married Biological/Adoptive Parents)
Step-Parents	1.27	.76
Other	2.24*	1.06
Parental Monitoring (W1)	1.00	.99
Commitment to School (W1)	.96	.97
Liberal Peer Attitudes (W3)	3.78***	2.57***

Source: Toledo Adolescent Relationship Study, Waves 1 and 3

Note: Reference category in parentheses. N = 532. Odds ratios presented in table.

^*p < .05

^**p < .01

^***p < .001

Model 1 presents multivariate results for models including all covariates. As shown in Model 1, the racial gap in endorsement of the player identity persists net of sociodemographic background, family, school factors, liberal peer attitudes, and prior behavior. The effect is reduced substantially but remains statistically significant. Hispanic and white youth have similar odds of indicating that others would describe them as a player. The disadvantaged neighborhood covariate is not associated with the player identity in the multivariate model. The inclusion of race and ethnicity mediates the effect of neighborhood disadvantage. These data provide support for Anderson’s hypothesis. However, it is important to note the overall distributions among male respondents who participated in the TARS study, as a majority of poor African American males (56.1 percent) are not players. This is a potentially important finding, as it illustrates that the identity is not ubiquitous among disadvantaged African American males; conversely a majority of the players in the TARS sample do not have this sociodemographic profile.

Respondents who have peers with liberal sexual attitudes have significantly higher odds of the player identity in the full model. Young men who are in a social network characterized by liberal peer attitudes are more likely to report having a player identity, and consistent with our theoretical discussion, the behavior in which the young men engage significantly increases the likelihood of being accorded this social identity label. Family structure was significant at the zero order but is not significantly tied to player identity in the full model.

Player Attitudes

Next we consider the attitudinal correlates of self-identifying as a player. As the results in Table 3 indicate, players more often than their non-player counterparts do agree that they “enjoy the chase more than the relationship” (36 percent vs. 16 percent). Also, young men who adopt the player identity more often indicate that they frequently talk to their friends about their sex lives (44 percent as compared with 27 percent of non-players). We note, however, that while players compared with non players more often agree with these items, a majority of players did not agree that they enjoy the chase more than the relationship and did not indicate talking often to their friends about their sex life. This provides evidence of heterogeneity in attitudes within the player subgroup, an observation we explore in more detail in our discussion of the in-depth interview data. Focusing on the relative salience of the peer group, players and non-players did not differ significantly in endorsement of the notion that friends are ultimately more important than girlfriends. This appears to be a generally prevalent sentiment for this adolescent age-group (70 percent of males agree) perhaps reflecting the results of prior research indicating that, on average, durations of friendships most often do exceed average durations of romantic relationships (see e.g., Furman, Brown, and Feiring 1999).

Table 3.

Distribution of Attitudes and Behaviors Associated with the Player Identity [about here]

	Total		Players		Non-Players
	Mean/%	S.D.	Mean/%	S.D.	Mean/%	S.D.
Player Attitudes (W3)
Agrees to "enjoy the chase"	20.26%***		36.26%		15.90%
Talk to friends about sex life       often	30.89%***		43.68%		27.41%
Friends come before       girlfrie nds	70.73%		71.17%		70.60%
Player Behavior (W3)
Age at first sex (years)a	15.89***	2.31	15.29	1.84	16.11	2.46
Total Number of Sexual       Partners (in past 24 months)	2.13***	3.92	4.64	6.46	1.45	2.26
Number of Nondating sexual       partners (in past 24 months)	1.40***	3.22	3.14	4.23	.93	2.87
Has Cheated (in past 24   months)	18.28%***		40.91%		12.12%
One Night Stand (in past 24       months)	23.17%***		45.26%		17.16%
N	532		111 (21.38%)		421 (78.62%)

^aOnly sexually active respondents are included in this analysis (N = 354)

Source: Toledo Adolescent Relationship Study, Waves 1 and 3

Note: Mean/% = mean or percent; N = number. Weighted distributions.

^**p < .01

^***p < .001 difference between players and non-players

Sexual Behaviors

Turning to the sexual behavior experiences of respondents who agree with this self-characterization, the player identity is tied to sexual behaviors (see Table 3).² The mean age at first sex is younger for players, and these youths report a higher number of partners in the past 24 months, as well as a higher number of non-dating sexual partners. Players more often indicate that they had cheated on a partner within the past 24 months relative to young men who did not identify themselves as players (41 percent vs. 12 percent). Players also more frequently reported participating in a ‘one-night stand.’ However, again the findings highlight that not all players reported these behaviors. For example a majority indicated that they had not cheated (59 percent) and 55 percent had not participated in a ‘one-night stand.’

Longitudinal Analyses

We next estimate a series of models that include wave 4 results, in order to evaluate the degree to which the player identity is tied to later sexual behavior, once controls for traditional risk factors and prior behavior have been introduced. Table 4 presents results of regression models that predict number of sexual partners in the last 24 months as reported at wave 4. The zero-order models indicate that player identity and prior sexual behavior are both positively associated with more partners. African American, Hispanic, and youths of ‘other’ racial backgrounds report a greater number of partners than white youth. Respondents living in a disadvantaged neighborhood have more sexual partners. Youths residing in single parent or step-parent families have more sexual partners than teens living with two biological parents. Respondents with greater school commitment report a lower number of sexual partners and youth who have peers with more liberal attitudes have a greater number of sexual partners.

Table 4.

Ordinary Least Squares Regression Models Predicting Number of Sexual Partners (W4) [about here]

	Zero Order	Model 1	Model 2	Model 3
Prior Behavior (W3 sexual partners)	.45***	--	.40***	.37***
Player Identity (Self rating)(W3)	1.05***	.78***	--	.39*
Sociodemographic Characteristics
Race/Ethnicity (W1)
(White)
African American	1.94***	.89	1.15	.89
Hispanic	1.70**	1.17	.95	.92
Other	2.92**	2.68*	2.30*	2.24*
Respondent’s Age (W3)	.09	.04	−.07	−.07
Disadvantaged Neighborhood (W1)	1.00*	−.38	−.60	−.52
Family Characteristics
Mother's Education (W1)
Less than 12 years	.20	−.25	−.03	−.09
(High School)
More than 12 years	−.19	.19	.37	.37
Family Structure (W1)
Single Parent	1.54**	.50	.46	.40
(Married Biological/Adoptive Parents)   Step-Parents	3.50**	2.82***	2.65***	2.63***
Other	.41	−.63	−.59	−.68
Parental Monitoring (W1)	−.03	−.02	−.04	−.04
Commitment to School (W1)	−.51*	−.29	−.21	−.19
Liberal Peer Attitudes (W3)	2.14***	1.26**	1.12*	1.00*
Intercept	--	1.28	2.54	1.64
R2		.16	.23	.24

Source: Toledo Adolescent Relationship Study, Waves 1, 3, and 4

Note: Reference category in parentheses. N = 481.

^*p < .05

^**p < .01

^***p < .001

Model 1 in Table 4 includes the player identity along with respondent sociodemographic characteristics, family indicators, and peer attitudes. The race gap observed in sexual behavior reports in the zero order results (African American) is no longer statistically significant in the multivariate model. The race effect in this model is mediated by player identity. Model 2 replaces the player identity with a measure of prior sexual behavior. Prior sexual behavior is significantly associated with later sexual behavior and in this model once again the race differential is no longer statistically significant. The race effect in this model is mediated by prior sexual behavior. Model 3 includes both prior behavior and player identity. Identity and prior behavior are both statistically significant predictors in this model, each retaining an independent effect on number of sexual partners reported at wave 4. However, including the index of prior behavior reduces the effect of player identity by half. This indicates that much of the latter association operates through the stable behavioral trajectory in sex partners over time. Nevertheless, based on a nested F test comparing models with and without the player identity, the player identity adds significantly to explained variance in sexual behavior reports (Model 2) elicited at wave 4 (p < .04). The prior behavior measure also adds significantly to Model 1 that includes only player identity (p < .001). In the full model, ‘other’ race, step-parent family, and liberal peer attitudes are also significant predictors of number of sexual partners.³

We also estimated models that include interactions of the race/ethnicity categories with the player identity (results not shown). The interaction terms are not statistically significant and indicate that the player identity has a similar association with sexual behavior for African American, Hispanic, white, and ‘other’ race youth. We also examined whether endorsement of the player identity exerts a differential effect on reports of risk-taking behavior according to the respondent’s age. The interaction is not statistically significant, indicating a similar effect across age groups represented in the TARS sample. In addition, we estimated a model that included a player identity and prior sexual behavior interaction term. This term was also not significant, indicating a generally similar effect of player identity, regardless of whether or not the respondent reported high numbers of sexual partners at wave 3.

An Analysis of In-Depth ‘Relationship and Sexual History’ Narratives

The content of the in-depth interviews we elicited from a subset of the respondents generally accord with results of the quantitative analyses, but further complicate Anderson’s depiction of the player role. Consistent with the quantitative results, some of the African American youths in this smaller subset could be considered players, but this status is not ubiquitous among the individuals who provided these in-depth narratives; further, as several quotes below document, a number of white youths also indicate that others describe them in this fashion. These unstructured narratives also underscore that this is an identity that youths themselves understand and take into account, whether or not they believe this is an accurate description in their own case. The unsolicited references boys make are important because they signal that the label is not simply a construct developed by Anderson and appropriated by the authors, but a significant concern/issue of boys when discussing the social scene at their school, heterosexual relationships, and their own identities (e.g., I’m not some player… [Todd, 17; white]; like in school I used to kind of get a label, like a player” [Jason, 18; white]. Nevertheless, these narrative accounts also reveal a somewhat less straightforwardly positive view of boys’ of the player role, and in the process allow a more nuanced assessment of specific aspects of Anderson’s hypothesis.

Recall Anderson’s view that the key reference group for young men is their same-gender peers, where a primary concern is “looking good” in their eyes. The young men who participated in the in-depth interviews certainly reflected on social rewards from their male friends. For example, Jason, quoted above, said, the guys would say it because it’s cool… However, our analyses of the narratives hint that adolescent boys are not focused exclusively on the prestige gains they receive from other male peers. Previous research relying on the TARS qualitative data highlighted that many boys we interviewed appeared to care a great deal what girls think about them, and about the progress of their romantic endeavors (Giordano, Longmore, and Manning 2006; Giordano, Manning, and Longmore 2006). Since the results described above document that a majority of young men in the sample are not players, perhaps this is not surprising. Yet even when we examined the narratives of those who indicated that they had acquired player reputations, we find evidence of social considerations that transcend the world of the male peer group, and elements of ambiguity about this role.

Disclaimers

A number of young men who agreed with the structured item, “other people would describe [me] as a player,” within their own narratives included a range of justifications or ‘disclaimers’ about this description. As scholars of the narrative have pointed out, such justifications or “accounts” occur as a kind of linguistic interruption, highlighting potential problem areas (Scott and Lyman 1968). Thus, to the degree that these young men wholeheartedly embraced their player status and the peer group rewards that attach to it, they should not find it necessary to develop such disclaimers, apparently invoked in order to distance themselves from some of the more discrediting aspects of this role. Below we describe three such disclaimers or accounts developed by young men who had acquired the player label.

The social nature of the player role. As stated at the outset, the symbolic interactionist perspective highlights the role of interaction and communication with others as central to the development of identities, and Anderson’s treatment of the player is generally consistent with this perspective. But while Anderson focused almost exclusively on the importance of the attitudes of male peers, several of the respondents we interviewed stressed young women’s views and actions as a critical part of the social construction process:

There are some girls that would, just do anything, to be the one known to be the one that messed around with you! There’s a lot of girls that do that, they’ll use me as a trophy. [Girls] wanted to show me off, like this is who I am with… It’s just sort of cool, cause sometimes when girls like you, they want everybody to know you, they want to introduce you to everybody…they want just the hottest dude they can find, and be able to say this is mine. [Julian, 17; white]

In his own narrative account, then, Julian has framed his extensive dating life as a process heavily influenced by if not set in motion by the girls he dates, rather than by his own behavior or desire to win points in the eyes of male friends. This serves to bolster his positive self-image, and circumvent potentially negative attributions associated with the player type. This “I-can’t-help-it-if-all-the-girls-want-me” disclaimer also theorizes a level of objectification on the part of some of the girls he knows. This depiction is also of interest, because while Anderson focused more attention on young men’s perspectives, he did develop the idea of strong gender contrasts (i.e., girls were depicted as primarily interested in love and the development of a long term relationship, while boys were more likely to objectify partners, and to dupe or ‘play’ them, in order to achieve their goal of sexual conquest).

The “phase of the life course” disclaimer. Another way in which boys minimize some of the more negative connotations of their player behavior is to focus on the age-appropriate nature of their actions. As teens, several boys focused on the reality that they are too young to be married, and may extend this to include the view that long-term involvement with one girl in high school is not necessary or even appropriate. As an example, one young man who resonated with the player label told the interviewer that his uncle specifically counseled him to ‘play the field’ upon entering high school, rather than becoming tied down with one girl. Daniel, a 17 year old Hispanic youth reflected a similar sentiment, as he explained why he broke it off with a girlfriend: I was too young. I don’t want to date that long. That’s like married almost… I was too young, I was young…

The honesty disclaimer A third way in which the young men distanced themselves from the negative connotations of their player status is their focus on the association of this label with lying and other disingenuous interactions with the young women they dated. Thus, some boys agreed that they dated many girls, but to the degree that they maintained a level of honesty, went on to develop the argument that they were not ‘true’ players (here play is used as a verb, as in one who ‘plays’ girls):

Cause I’m you know, a kind of like a male whore at my school, lots of people don’t know me that good. Just cause they see all these girls that want me, they assume that I mess around with all of them and that I’m just some player… [but] I don’t think I am, no cause like I said a player is a kind of guy that would lie to a girl and lie to uh more than one. Like how they like them, I don’t do that. [David, 17; white]

People thought, you know, people heard it so much that they just thought it was true…that I was a player…That all I wanted to do was girls and such…. No I mean, they know I wasn’t going to play them, um because I wouldn’t want to be played myself, you know. [Andrew, 18; white]

David and Andrew’s narratives thus draw heavily on the distinction between their own views of self and attitudes/assumptions others make about them. Maintaining this distinction allows them to preserve a positive sense of self, as they focus in on their honest qualities. This too differs from Anderson’s depictions, however, as the disingenuous strategies involved in “getting over” on certain girls are described by Anderson as an integral and valued aspect of the game.

Temporal Shifts in Young Men’s Player Identities

In the above quotes, youths recognized their player status within the social milieus of their high schools and neighborhoods, but managed to maintain positive self-views by emphasizing their overall desirability with women, their level of honesty or up-front qualities, or by positioning this role as simply an appropriate phase for this period in their lives. However, we observe even more direct evidence of ambiguity about and discomfort with the player role in the narratives of youths who describe changes over time in their enactment of this social persona:

Like I was telling myself, when I was involved with three girls… I didn’t like it…what I was doing. I didn’t like the fact that I was hurting other girls that I liked. Kind of playing them I guess. Kind of like in school I used to kind of get a label. Like a player, and… like I didn’t like that…Like having a relationship now is pretty important to me… [Rob, 18; white]

I really didn’t like being like that…I had felt dirty and low when I was…when I had friends with benefits. It just wasn’t…it was just not me. I’ve always been raised the gentleman type and it’s just like I’m out here being like other guys just…I was looking for whatever and I felt bad about it… [Will, 17; biracial]

Yeah, in my mind yes, I knew it was wrong. [I: But you did it anyways?] Yeah. I don’t know why… [Jermaine 18; African American]

These references to shifts over time also complicate Anderson’s depiction, however, in that they point to something less than a constant proclivity that is continually bolstered by reinforcement from male companions. Certainly not all of the respondents we interviewed described changes in their player identities. Yet it is interesting to note that even the most unapologetic players within the sample frequently included some mention of a desire to eventually develop a more meaningful long-term romantic relationship:

I’d like to have a girl but I haven’t found her yet. I don’t want to get with some girl, where I’m going to want to cheat on her, and all of the other girls and all that stuff. So I guess I’m just going to wait for, if I want to be with the kind of girl you wouldn’t want to do that to, like she’s a good girl! The kind you can trust, so you want to be trusted back. [Doug, 17; white]

I got, yeah there’s like a lot of girls I could have right now, but I don’t know there’s always something wrong… I’m waiting until I find the most absolute, perfect everything in this girl. [Steve, 17; white]

Similar to the narratives of those who offer justifications for their involvement, or who indicate that they have changed over time, then, the comments of these young men reflect less than a full commitment to the enjoyable, socially rewarding aspects of their player lifestyles. Some of these respondents appeared to recognize that a continual succession of casual liaisons might not prove to be the most fulfilling strategy for their own happiness and long-term well being. These comments thus hint that while the young men have been influenced by numerous peer interactions, they are also aware of and influenced by cultural ideals that stress the promise of lasting love, and the security of a committed relationship. However, at this stage, they focus heavily on the qualities of potential partners (e.g., Steve’s suggestion that he wants the absolute, perfect everything in this girl) rather than their own actions and attitudes, as fostering this shift away from the player lifestyle.

Conclusion

In this paper, we focused on Anderson’s depiction of the adolescent player, a social type he encountered frequently in his ethnographic work in a disadvantaged area of Philadelphia. Relying on a diverse sample of teens living in a Midwestern urban/suburban area, we found that about one fifth of male respondents identified on some level with this label. The quantitative analyses document that, consistent with Anderson’s hypothesis, disadvantaged African American male adolescents are significantly more likely to agree that others would describe them as players; yet the data also highlight that this social identity transcends the confines of disadvantaged neighborhoods and minority populations—66 percent who agreed that others would describe them as players were not disadvantaged African American youths. This is a potentially important finding, given that the term ‘player’ was initially popular and may have its origins as a term used within African American communities (Anderson 1989:155).

The analyses also show that, with some exceptions, several attitudes Anderson linked to the player role are significantly associated with endorsement of this identity. Perhaps more important, youths who indicate that others would describe them in this fashion actually do engage in significantly more sexual risk-taking than their counterparts who do not resonate with this label, even after traditional predictors have been taken into account. This undoubtedly reflects a straightforward labeling process—young men who have had a succession of sexual partners are more likely than others to acquire this reputation and this label. Yet the correspondence between label and behavior is not a perfect fit. Some individuals who indicate that others consider them players do not self-report high numbers of sexual partners; conversely, some respondents who did not resonate with the player label reported a higher than average number. Consistent with prior theorizing on self (Stryker 1981) and labeling (e.g., Becker 1963) processes, however, results indicate that self-identity as a player predicts subsequent sexual behavior reports, even when sociodemographic, family, peer and prior behavior controls were introduced. Including the prior behavior measure reduces the effect of the player identity by half, but the association is still significant. This suggests that such identity labels are not imposed arbitrarily, but often derive from the individual’s own behavioral proclivities. Yet the identity measure explains additional variance in wave four reports of number of sexual partners. Thus our findings lend support for both the “self as social product’ and ‘self as social force’ emphases within the structural symbolic interactionist theoretical tradition. At the outset, we outlined some mechanisms that potentially link identity and these behavioral choices. For example, we theorized that gaining a reputation as a player might influence patterns of socializing, time spent on academic pursuits, and the adolescent’s level of social confidence.⁴ In turn, these social and self processes potentially increase the likelihood of accruing additional sexual experience. Due to the abstract and somewhat subjective nature of second order constructs such as identity, additional research is obviously needed on these and other linking mechanisms.

Interactions of race/ethnicity and the player status indicate a similar effect across race and ethnic groups in the influence of this identity on the number of recent sexual partners reported. The finding that including the player identity in these models reduces the race gap in sexual behavior reports adds to prior research on this gap that has tended to focus more on factors such as neighborhood characteristics or family structure effects (see e.g., Browning, Leventhal, and Brooks-Gunn 2005). The data also point out that a majority of African American youth who participated in the study (even focusing on those who could be classified as disadvantaged) do not endorse this identity. And, among those who do not, such youths are no more likely than their white or Hispanic counterparts to engage in high risk sexual behavior. This is congruent with Anderson’s (1990) emphasis on different adaptations and orientations within disadvantaged neighborhoods, but provides a quantitative assessment of the specific distribution of these orientations within a large, diverse sample of young male respondents.

The in-depth interview data add to this portrait by suggesting several ways in which youths maintain a positive self-image while being associated with this label. Although Anderson emphasized that boys are accorded high status when they are known as players, our own interviews suggest areas of ambivalence about the role. Some boys developed “disclaimers” about their player status, while other narratives highlighted malleability or shifts in young men’s player status over time. These areas of ambivalence and change are important to understand and research in more detail since they suggest that boys who fit this attitude/behavior profile are not destined to pursue this lifestyle indefinitely. Such findings also have implications for theory building in the area of identity-behavior connections. For example, societal reaction theories often depicted the results of labeling as powerful and irrevocable (Gove 1980). Thus, the respondents’ rather complex reflections and references to change over time serve to highlight the important role of human agency and the inherently dynamic nature of identity processes (Burke 2006; Emirbayer, and Goodwin 1994).

Limitations of the study include the local nature of the sample and that the reflected appraisal measure of player identity is only measured at the time of the third wave of interviews. It would also be useful in future research to consider several identity statuses simultaneously. For example, perhaps young men who are considered players in high school but continue to excel academically are less likely to continue a pattern of risky sexual behavior as they navigate the transition to adulthood. Another direction for future work is to develop more refined measures of attitudes associated with the player identity. Finally, we have not included young women’s views on the player role or considered issues of sexual identity from girls’ own perspectives.

Prevention efforts focused on health beliefs and knowledge have frequently been shown to be rather limited in their effectiveness (Mahoney, Thombs, and Ford 1995). The results of the current study suggest the value of focusing on issues of social reputation and identity as well as more broadly on relationship issues (e.g., cheating). Curricula should include attention to the role of peer groups, as interactions with friends and the wider network of peers clearly foster social rewards for such behaviors. Nevertheless, individuals and peer groups vary; attention to more negative social connotations and personal consequences of adapting the player role could serve to at least partially redirect adolescents’ normative beliefs about the desirability of adapting this orientation. Our research indicates that many non-players apparently do not admire this social type (see Giordano et al. 2006), and the results documented in this analysis suggest that even self-identified players are aware of some of the problematic features of this social role. It would be especially useful to focus additional research attention on those young men who distance themselves from such an orientation, and evidence a different identity profile, even within disadvantaged contexts.