Make your own free website on




Gary W. Harper, Ph.D., M.P.H.
Roderick J. Watts, Ph.D.
Midge Wilson, Ph.D.


   Among U.S. adolescents, street youth are disproportionately affected by HIV.   Using traditional models of HIV risk behavior to design interventions for street youth is unlikely to yield effective prevention strategies because those models fail to capture important factors that are features of the multi-level context in which street youth engage in HIV sexual risk behavior.
   Few theory-based HIV prevention programs have been rigorously evaluated.  Those that have been evaluated do not demonstrate consistent, sustained risk reduction. As a result, critics have challenged the applicability of traditional health risk reduction theories to HIV risk, and particularly to vulnerable populations such as street youth.  In step with that criticism, an alternative approach to HIV prevention strategy is evolving.  This approach recognizes the embeddedness of sexual risk behavior in larger systems, taking a broader view of context than traditional individualistic models are able to capture.  
   Contextual factors that impact the HIV-risk behavior of street youth may include gender-based power dynamics, the symbolic meaning that condoms carry, and past victimization.  These factors may be particularly salient for street youth (whose life contexts are frequently marked by abuse, emotional isolation, and drug use) and may play particularly prominent roles in the risk behavior of young women on the street.
   This study will examine the HIV sexual risk behavior of White female adolescent street youth with the goal of developing a collection of constructs that incorporate multiple levels of contextual influence.  Limiting the sample in this way will facilitate the development of constructs that are culture- and gender-specific.  Racial differences among women have been found in relation to body-image perceptions, self-esteem, gender-based power dynamics and safer-sex negotiation.  In terms of gender, males and females approach condom negotiation from different power positions, and males tend to hold more negative attitudes toward condom use than females.  
   Twelve to fifteen participants will be interviewed up to three times each.  Using grounded theory methodology, qualitative data will be analyzed, or "coded", to identify relevant concepts and form conceptual categories.   Concepts that emerge from the coding of the first two data waves will influence subsequent interview content.  Linkages between conceptual categories will be identified, forming constructs that represent precursors to a model of HIV sexual risk behavior.
   U.S. adolescents are at higher risk of contracting HIV than virtually any other subgroup of the population (Carre et al., 1997; Hendricks et al., 1998).  Of this subgroup, street youth in particular are disproportionately affected by the virus (Institute of Medicine [IOM], 1994; Sweeney, Lindegren, Buehler, Onorata, & Janssen, 1995).  They are youth who spend all or most of their time living and working on the street, often without adult supervision or a consistent source of shelter.  Overall, the rate of HIV infection among street youth has been estimated at 2 to 10 times the general adolescent rate (DiClemente, 1992; St. Louis et al., 1989).  Reducing the incidence of new HIV infections among street youth has become a public health priority (IOM, 1996).
   The goal of the present study is to contribute to the development of a model that reflects the influence of context street youth's HIV sexual risk behavior.  The introduction begins with an multi-level examination of youth street life and its potential implications for HIV risk behavior.  A review of the research literature on models of HIV sexual risk behavior and HIV prevention follows, with a discussion of the methodological, conceptual, and epistemological  limitations of  those models and programs.  This is followed by a description of the specific strategies the current study will use to address the limitations of past research.  The next section presents a rationale for the present study population and the guiding research questions.  
Street Youths' Lives
   Although developing countries face the highest levels of youth homelessness, the focus here will be on the U.S., where the number of youth living on the street is substantial and growing (Rotheram-Borus, Parra, Cantwell, Gwadz, & Murphy, 1996).  It is estimated that between 1.3 and 2.0 million U.S. youth live on the street during any given year (Athey, 1991; Rotheram-Borus, Koopman, & Ehrhardt, 1991).
   Street youth spend most or all of their time immersed in the culture of the streets, often without a consistent source of shelter or adult supervision.  Many systems have been devised in an attempt to classify these youth, whose living situations and life circumstances may vary widely.  In the international homeless youth political movement, Bond, Mazin, and Jimenez (1992) distinguish between “children on the street” and “children of the street.”  Children on the street typically work there, but have regular adult supervision available and go home to sleep at night.  Children of the street are completely isolated from their families and often live communally or “squat” and rely on each other to provide for their survival needs.  Classifications of U.S. street youth are usually based on the varied circumstances that lead youth to street life.  Bucy (1987) categorizes street youth as “homeless” (those whose families are homeless), “system kids” (those with a foster care history), “victims” (those who have experienced abuse or neglect at home), “throwaways” (those expelled from the home), and “runaways” (those who unilaterally decide to leave home).  In addition to Bucy's categories, a sixth group can be identified as street youth as well: “curb kids”.  Muir (1991) describes curb kids as youth who have a home, but spend a great deal of time working or hanging out or working on the street.  Curb kids may participate in street life either by choice, or as a result of neglect by caretakers, or both.  These six categories overlap; for instance, a youth who is expelled from the home following abuse or neglect might eventually be in foster care.  Unlike Bond's categories they do not indicate a level of housing stability or contact with caring adults.  “Homeless youth” and “curb kids”, for instance, may be housed with an intact, functional family, and “system kids” have consistent shelter and adult contact in foster care, whereas “throwaways” and “runaways” are typically isolated from their families and face unstable housing situations.  Obviously, with youth living on the street within such varied contexts, it is difficult to classify them in a way that reflects the wide range of struggles and challenges they face individually.  
   Once on the street, youth often affiliate with one or more subcultural groups that comprise the youth street culture.  While these groups tend to vary geographically, Kipke, Montgomery, Simon, Unger, and Johnson (1997) identified seven primary subcultural groups among Hollywood street youth: Gay/bisexual youth, “druggies” (drug users, dealers and buyers/runners),  “skaters/deadheads” (skaters, surfers, deadheads, and skinheads), “hustlers” (male and female sex works and drag queens), gang members, student/athletes, and “punkers” (who identified by their unique lifestyle).
Street Youth and HIV Sexual Risk
   A majority of U.S. adolescents report participating in some HIV sexual risk behaviors (Centers for Disease Control and Prevention [CDC], 1998d), but the rates of HIV risk behaviors and subsequent HIV infection are higher among street youth than among most other subgroups of adolescents (IOM, 1994; Rotheram-Borus, Feldman, Rosario, & Dunne, 1994; Rotheram-Borus et al., 1991; Sweeney, et al., 1995).  Street youth engage in injection drug use (IDU) and unprotected sex (the two primary means of HIV transmission) at higher rates than other adolescents.  Estimates of street youths' IDU rates are as high as 30%, with 59% reporting needle sharing (Harper & Carver, 1999; Kipke, O'Connor, Palmer, & MacKenzie, 1995).  In contrast, IDU rates among in-school adolescents are thought to be approximately 2% (CDC, 1998c).  Research on the sexual risk behavior of street youth has consistently found that they engage in high levels of sexual activity, but condom use is relatively low (Anderson, Freese, & Pennbridge, 1994; Rotheram-Borus, & Koopman, 1991; Sugerman, Hergenroeder, Chacko, & Parcel, 1991; Yates, Mackenzie, Pennbridge, & Cohen, 1988). They tend to initiate sexual activity earlier than other adolescents (Anderson, et al.) and they tend to have more sexual partners than their non-street peers, with as few as 15% reporting consistent condom use (Harper & Carver; Hudson, Petty, Freeman, Haley, & Krepcho, 1989; Rotheram-Borus, et al.).  
   As many as 16.7% of adolescents at homeless youth centers are infected with HIV (Allen et al., 1994; Sweeney, et al., 1995). Overall, the rate of HIV infection among street youth has been estimated at 2 to 10 times the general adolescent rate (D'Angelo, Getson, Luben, Stallings, & Gayle, 1989; DiClemente, 1992; New York City Health Department, 1990; Quinn et al., 1988; St. Louis et al., 1989).  As HIV continues to spread through the sexual network of street youth, the infection rate can be expected to rise (Service & Blower, 1995).
   Three subgroups of street youth have been identified by some researchers as being at heightened risk for HIV: victims of sexual abuse, drug abusers, and gay male youth.  As many as 60% of street youth report being sexually abused (Yates et al, 1988), and sexual abuse is associated with having a higher number of sexual partners (Rotheram-Borus, Koopman, & Bradley, 1989).  As many as 66% of street youth meet the criteria for clinical diagnosis of drug abuse, which is associated with increased HIV risk (Crumble et al., 1998; Jemmott & Jemmott, 1993; Kipke, et al., 1997; McCusker et al., 1990; Seage et al., 1997).  Estimates of the number of street youth who identify themselves as gay range from 6% (National Network of Runaway and Youth Services, 1991) to 27% (Kipke et al., 1995).  Gay male youth begin sexual activity at an early age and are likely to engage in sexual activity with gay men, who as a group have high seroprevalence rates (Rotheram-Borus, 1991).  
HIV Risk in The Nested Contexts of Street Life
   The preceding statistics reflect the prevailing epidemiological approach to research on HIV and street youth.  They are useful in identifying segments of the street youth population that may be most vulnerable to HIV, and in indicating the types of sexual behaviors that place them at risk.  However, such findings merely catalogue risk behaviors without explaining why they occur.  As a result, they are of limited use in designing effective interventions, since “without and understanding of the context that leads to risk, the responsibility for reducing the risk is left with the individual, and nothing is done to alter the more fundamental factors that put people at risk of risks.” (Link & Phelan, 1995, p. 85).  
   In a 1994 report which reviewed federally funded AIDS-related interventions, the Institute of Medicine admonished that the HIV risk behavior of individuals cannot be accurately analyzed apart from the larger environmental structures in which it is embedded  (IOM, 1994).  It is an observation that is repeated frequently by critics of traditional approaches to analyzing HIV risk behavior (Choi & Coates, 1994; DiLorenzo & Hein, 1995; Kashima, Gallois, & McCamish, 1993; Parker, 1996;  Satcher, 1996; Tawil, Verster, & O'Reilly, 1995; Muir, 1991).  In order to understand the HIV-risk behavior of street youth, it is first important to examine the context of these young people's lives and the myriad of environmental challenges that they face.
   Bronfenbrenner's (1993) ecological perspective of human development provides a starting point for appreciating the scope of interrelationships among environments and their impact on individual behavior.  Following a brief description of that perspective, the context of street life and implications for HIV risk behavior will be examined at each level of analysis that Bronfenbrenner outlines.  A full appreciation of the impact of divergent levels of influence on HIV risk behavior will require a constructivist approach to research, focusing on qualitative reports from street youth themselves.  However, the discussion will include attempts to identify obvious interrelationships among levels of influence as they relate to HIV sexual risk behavior.
   Bronfenbrenner conceptualizes the “environment” as four nested interdependent levels (Appendix A).  The most basic unit is “proximal level” or the microsystem which is a pattern of roles, activities and interpersonal relations experienced in a given face-to-face setting with particular social and symbolic features.  Microsystem settings for most adolescents include the family, school, and other peer group settings, as well as the settings in which sexual encounters occur.  A fundamental assumption is that events in each microsystem are interrelated to and interact with events in other mircrosystems (Bronfenbrenner, 1993).  
   Several microsystems in interaction constitute a mesosystem.  The richness or impoverishment of the mesosystem depends on the number and quality of the interrelating links among microsystems and the values they do or do not share (Muus, 1996).  Within the mesosystem, an individual plays different roles, which change as a function of time and maturation and as a function of moving from one microsystem context to the next (Bronfenbrenner, 1979).  
   The exosystem is the structure of the larger community, which includes infrastructure resources, community organizations, employers and local government.  Although the individual usually does not participate in exosystem processes, the results of those processes exert a one-way influence on the resources available to the individual and can impoverish or enrich the quality of micro- and mesosystem interactions.  
   The macrosystem is the fourth and largest level of the environmental system.  It contains the “overarching societal groundplan for the ecology of human development” (Bronfenbrenner, 1993), and consists of the pattern of subordinate systems characteristic of a given general culture.  This includes the general cultural, social and economic values as well as public policy (Muus, 1996).  Embedded in the macrosystem are opportunity structures and life course options (Bronfenbrenner).
   Microsystem influences on HIV risk behavior.  The decision to use a condom is negotiated within a microsystem, the complexity of which has been underestimated in HIV prevention research (Afifi, 1999; Amaro, 1995; Kashima et al., 1993; Soet, Dudley, & Dilorio, 1999; Worth, 1989).  Condom use negotiation is influenced by microsystem features such as the relative power of each partner and the symbolism that condoms carry in the context of that particular relationship.  Often, the position of women in regard to these microsystem features make it particularly difficult for them to engage in HIV protective behavior.  In the context of street life, young women may face additional obstacles to condom use.
   Gender-based power dynamics play an integral role in condom negotiation.  Men are more likely to hold negative attitudes toward condom use than women, tending to agree with statements such as “condom use reduces pleasure” and “condoms are uncomfortable” (Cline & McKenzie, 1994).  Since the actual use of a condom is clearly under male control, a woman's ability to enact safer sex often depends on her ability to influence male partners to use a condom (Soet et al., 1999).   
   Yoder and Kahn (1992) characterize this type of interpersonal influence as “power over” or dominance.  Unfortunately, condom negotiation is typically embedded in a context of socially sanctioned inequality between heterosexual partners  in which male dominance prevails and women are socialized to be sexually passive (Amaro, 1995).  Gender inequality not only skews the power balance of condom negotiation, but may also prevent it entirely.  Harlow and colleagues (Harlow, Quina, Morokoff, Grimley, & Rose, 1992) found that women who anticipated negative partner reaction toward condom use were more likely to engage in unsafe sexual behaviors.  Heterosexual women in disadvantaged socioeconomic contexts (i.e., poverty and homelessness), often further subordinate their desire to use a condom in order to maintain access to the survival resources that male sexual partner bring to the relationship (Worth, 1989).
   For street youth, the power imbalance in sexual microsystems is even more pronounced in the context of survival sex.  With limited economic options, street youth are often forced to sell their bodies in order to survive. As one young woman describes it: “I couldn't get a job, and I couldn't get a place.  I was just stuck ... I was living on the streets and getting propositions when I decided to do it because I didn't feel like there was any other choice at that time” (San Francisco Task Force on Prostitution, 1996).  Up to half of street youth report trading sex for money, food, drugs, or a place to stay (Kipke et al., 1995; Rotheram-Borus & Koopman, 1989; Yates et al., 1988).  The ramifications for HIV risk behavior are clear -- street youth are not necessarily in a position to control condom use in the context of survival sex.  Refusing to engage in unsafe sex, for these youth, may mean going without necessities such as food, shelter, or protection.  Among young women on the street who engage in survival sex, a dual power disadvantage exists when negotiating condom use in terms of both gender dynamics and the dynamics of exchanging sex for survival resources.
   The fact that condoms are not symbolically neutral further complicates condom use negotiation (Afifi, 1999).  Condoms are imbued with meanings that can threaten both partners' identity and the sanctity of their relationship (Sobo, 1995).  For many individuals, two central messages are implied by requests to use condoms:  the requestor is sexually promiscuous (and potentially diseased) and/or believes that her/his partner is.  Women may avoid asking a male partner to use a condom because of the negative message it may imply about her.  In the context of traditional gender discourse, which dictates that women are to have fewer sexual partners than men (Holloway, 1984), a woman who requests that her partner use a condom risks being seen as sexually active, implying that she is `available' for sex, and that she is `seeking' sex (Worth & Rodriguez, 1987).  This perception is most salient in sexual encounters occurring early in a relationship, when image management is particularly important (Afifi).
   In established sexual relationships, the stakes rise.  Engaging in sex without a condom communicates trust and condoms become associated with infidelity and deceptive behavior (Sobo, 1995).  Once condom use is abandoned or limited to contraceptive purposes, reintroducing them threatens the emotional comfort and security of the “idealized union,” characterized by faithfulness and trust (Sobo, 1993).   Young women engage in this form of idealization earlier in relationships than do young men.  They tend to interpret casual sexual encounters as more meaningful than do young men in order to maintain a “respectable” self-image (as defined by traditional gender discourse).  Therefore, they tend to be willing to abandon condom use early in relationships in the interest of establishing intimacy and trust (Moore & Rosenthal, 1998).
   Issues surrounding the symbolic meaning of condoms may take on more salience in the sexual relationships of street youth.  Their opportunities to form healthy, trusting relationships are often limited or non-existent, so they may crave the emotional security of a stable, committed, monogamous relationship. Although no research exists on the topic, street youth may engage in unprotected sex at an early stage in relationships and may tend to over-idealize sexual relationships in an attempt to create bonds that meet their unfulfilled emotional needs, in spite the HIV risk involved.
   Mesosystem influences on HIV risk behavior. The microsystem in which condom negotiation occurs is not self-contained.  It is influenced by a vast accumulation of experiences in other microsystem-level relationships.  In healthy environmental systems, there is congruence among the developmental demands of various microsystem encounters.  Individuals draw on past experiences and capacities developed in one microsystem to help them meet the demands of another.  However, the environmental system of street youth is impoverished by fragmentation.  They are prematurely confronted with adult roles and responsibilities that are incongruent with their developmental stage.  Given this mesosystem disorganization, events in microsystem processes between street youth and their sexual partners must be understood in the context of the roles they play and the resources they develop within other microsystems, particularly family relationships, street peer relationships, and relationships in which they have been sexually victimized.
   The impact of family experiences on street youths' HIV behavior is not well understood, but it is recognized as a potentially important influence needing further study (McCarthy & Hagan, 1992).  For most adolescents, the family is their primary microsystem in which positive processes such as emotional support and guidance are necessary for healthy development (Bronfenbrenner, 1993).  As adolescents mature, they develop a sense of efficacy and esteem by assuming meaningful but manageable responsibilities.  However, as they develop autonomy, the security of home and parents remains an important resource for youth (Mechanic, 1991).  However, street youth prematurely confront developmentally challenging responsibilities, such as independently providing themselves with food and shelter.  They are not able to rely on the security of a healthy family microsystem to help them successfully negotiate the normative developmental challenges of adolescence, let alone the difficulties of surviving on the street (Family and Youth Services Bureau [FYSB], 1995).  Nor do they typically have nurturing family experiences upon which to draw as emotional resources.  Their family experiences are typically marked by dysfunction, neglect, and abuse.  Approximately a quarter of street youth report violence from other family members, the absence of a caretaker or parental drug or alcohol abuse (Bass, 1992).  In fact, lack of a supportive, functional family is the factor most commonly associated with adolescent homelessness (Rotheram-Borus, 1993; U.S. General Accounting Office, 1989).
   Sexual powerlessness and victimization, often by family members, is another disturbing feature of many street youths' mesosystems.  As many as 60% report having been sexually abused (Barden, 1990; Chelimsky, 1982; Rothman & David, 1985; Yates et al., 1988).  Among runaways at emergency shelters, 70% were victims of sexual assault or severe physical abuse (Kennedy, 1991).   The interrelation between abusive microsystem experiences and later sexual microsystem events is firmly supported by the literature on sexual victimization.  A history of sexual victimization was a significant predictor of unsafe sexual behaviors (Harlow et al., 1992), and sexual abuse was associated with having a higher number of sexual partners (Rotheram-Borus et al., 1989) as well as a higher probability of being re-victimized later in life (Russell, 1986).  Early sexual exploitation may also be linked to later engaging in illicit sexual transactions.  Survey research indicates that large proportions of sex-trade workers report a history of early sexual victimization (Wright, 1997; Wyatt, Newcomb, & Reiderle, 1993).  In addition, clinical observations suggest that sexually expolited children come to define their sexuality as “not belonging to them ... as being a commodity” (Wright, p. 1).  Among street youth, sexuality may be the only commodity they have to sell.
   Research has not specifically examined the association between street youth's HIV sexual risk behavior and their history of trauma or victimization.  But intuitively, if street youths' past microsystem experiences are marked by victimization, powerlessness and the inhibition of their developing sense of esteem and efficacy, the role competencies and resources that they can be expected to bring to sexual encounters may be limited and unlikely to foster HIV protective behaviors.  Investment in one's own well-being is an essential personal resource in motivating HIV protective behaviors.  However, street youth who experience microsystem events that communicate to them that they are not valued and that they have little control over events in their lives may have difficulty making that investment in themselves.
   In addition to family disorganization and victimization, a third important mesosystem influence on street youths' HIV risk behavior is peer group affiliation.  Youth tend to associate with peers whose sexual behavior is similar (Dolcini & Adler, 1994; Jessor & Jessor, 1977).  Kipke et al.'s research with subcultural groups of street youth (Kipke et al., 1997) found that condom use is influenced by peer group affiliation.  While “hustlers” and gay/bisexual youth were significantly more likely than other peer groups to engage in protected survival sex, “punkers” and gang members were significantly more likely to engage in unprotected sex, indicating that peer norms may function as both a risk and protective factor for HIV-related behavior.  In order to better understand the nature of these peer groups, and the ways in which peer values, beliefs and behavioral norms influence risk behaviors, more research is needed (Kipke et al.).
   Exosystem influences on HIV risk behavior.  At the exosystem level, there is no sanctioned role for street youth.  Those who are minors, unless legally emancipated, are excluded from access to the basic societal resources that they need, such as employment, housing and education.  The absence of legitimate employment opportunities may force minor street youth to engage in illicit transactions, such as survival sex or criminal activity, in order to provide for their own basic needs (Greenblatt & Robertson, 1993; San Francisco Task Force on Prostitution, 1996).  However, regardless of whether they have the money or how they earned it, minor street youth cannot buy shelter on their own, since they are too young to legally rent a place to live or even register at a motel.  Consequently, stable housing is available by either depending on an adult who can obtain and share it with them (which diminishes their power in that relationship, already skewed by an age difference), or by “squatting” illegally in abandoned buildings (which exposes them to numerous health and safety risks).  
   At 18, although finally old enough to sign a lease, many street youth cannot legitimately earn enough money to afford decent housing because they lack a high school diploma -- parents or guardians typically must be involved in registering for school (National Coalition for the Homeless, 1999; San Francisco Task Force on Prostitution, 1996).  One survey of street youth indicated that as few as 5% attended school regularly (Harper & Carver, 1999).  So, ironically, the same exosystem that denied street youth access to basic resources as minors thwarts their transition from illicit activities to mainstream adulthood for lack of the very resources they were denied.
   With so few opportunities to legitimately meet their own basic needs, survival on the street is a day-to-day struggle.  Street youth are forced to focus on obtaining not only shelter, but food, clothing, and medical care when needed.  Most carry weapons in order to protect themselves from victimization, and some resort to scavenging in dumpsters to find food and other necessities (Greenblatt & Robertson, 1993; Lundy, 1995).  Providing for immediate survival needs often takes precedence over long-term concerns such as HIV (Luna, 1991).  As a result, the value that street youth assign to the threat of HIV infection cannot be assumed to parallel that of other adolescents, nor can it be assumed to be take precedence over short-term concerns that are subjectively less “serious” than contracting HIV.  Messages about AIDS prevention are easily disregarded in the context of more immediate crises such as finding a meal and a safe place to sleep (Muir, 1991).  As one youth put it, “Why should I care about dying ten years from now when I do not know where I will sleep and how I will get food tomorrow?”  (Rotheram-Borus et al., 1991, p. 1191).  
   Macrosystem influences on HIV risk behavior.  At the macrosystem level street youth live on the fringes of the larger sociocultural system, embedded in which are the lifecourse options and patterns of social interchange that are characteristic of our society.  Their status is  “[marginalized] with respect to the dominant culture and their lives unfold on the edge of society and experience” (Lundy, 1995).  Lacking ties to the fundamental units of mainstream social organization, such as school, family and work, they have limited opportunities to develop the basic skills that mainstream culture rewards, including planning skills and consistent work habits.  A sense of alienation and frustration at their lack of the types of resources that would be required to integrate into society may contribute to many street youths' feelings of hopelessness (Cwayna, 1993).  
   At the policy level, response to the needs of street youth has been inadequate (Rotheram-Borus et al., 1991).  Services available to street youth have not grown with their growing numbers (IOM, 1994), and the number of foster care homes have decreased while the number of street youth continues to increase (Rotheram-Borus et al.).  Although street youth can access services designed for their needs through community-based agencies, they typically avoid such contact for various reasons, including fear of being returned to family or foster care and general distrust of adults who represent “the system” (Harper & Carver, 1999; Robertson, 1988).  Overall, the limited response of society at the macrosystem level to the needs of street youth ultimately constrains the life options available to them.
The Psychological Toll of Street Life
   At every level of environmental analysis, from the microsystem to the macrosystem, the context of street life is marked by challenges, traumas, and unmet needs that impact physical and emotional wellbeing.  The psychological toll of street life is immense and cannot be ignored in attempts to understand HIV risk decision-making.  Cut off from the resources and pathways that mainstream society offers for achieving status and security, street youth are unlikely to view their future optimistically (Cwayna, 1993; Greenblatt & Robertson, 1993; Luna & Rotheram-Borus, 1992).  Without a sense of hope for one's future, an individual is unlikely to act on standard risk reduction recommendations (Cochran & Mays, 1989; Mays & Cochran, 1988).  In fact, variations in degrees of future time orientation have shown reliable links with various measures of condom use (Agnew & Loving, 1998).  
   Studies of the mental health status of street youth reveal that feelings of hopelessness and despair are common.  Street youth experience higher levels of depression, and engage in more self-destructive behavior, including suicidal behavior, than non-street youth (Booth & Zhang, 1996; Caton, 1986; FYSB, 1995; Greenblatt & Robertson, 1993; Shaffer & Caton, 1984).  A suicidal or depressed youth is unlikely to change her behavior in order to promote a healthy future (Luna & Rotheram-Borus, 1992).  As one youth put it, “I live on the street.  I don't really have anyone.  I die every day.  I'm not afraid of death.  I'm afraid of life” (Luna, 1991, p. 513).
   Faced with the multiple stressors of street life, street youth as a group use drugs and alcohol at alarming rates -- many in an effort to self-medicate in order to cope (FYSB, 1995).  In a study by Koopman, Rosario, and Rotheram-Borus (1994), 33% of runaway youth reported using drugs specifically as a way of dealing with problems and 29% reported that it was a form of escapism.  In a peer-conducted survey of suburban street youth, 92% reported using marijuana, 47% reported amphetamine use, and 45% reported using LSD (Harper & Carver, 1999).  Kipke et al. found similarly high use rates among Hollywood street youth for marijuana use (93%) and LSD (61%) (Kipke et al., 1997).  In that sample, 89% reported ever having been drunk and 31% meet diagnostic criteria for having both an alcohol and drug abuse disorder.  The use of alcohol and noninjecting drugs does not represent a direct route of HIV transmission, but substance use is thought to disinhibit sexual risk behavior, posing an indirect risk (Miller, Turner, & Moses, 1990).  In fact, among heterosexual runaways, frequency of substance use was inversely related to frequency of condom use (Koopman et al.).
    Since street youth face a host of stressors unique to their life situation, standard HIV prevention programs designed for general adolescent populations are unlikely to be effective in changing their HIV risk behavior (DiClemente, 1992; Rotheram-Borus et al., 1994). Addressing the challenge of HIV prevention with street youth will require innovation the design of interventions as well as the in the theories that guide intervention development.
Review and Critique of the Literature on HIV Risk and Prevention
   Some grounded theory methodologists recommend against conducting a literature review in the substantive area of the study prior to data collection (Glaser, 1992; Rennie, Phillips, & Quartaro; 1988).  The concern is that becoming steeped in existing theories of the phenomenon being studied might reduce the investigator's openness to new concepts that emerge from data collection.  Despite these concerns, reviews of the models traditionally used to explain HIV risk behavior were useful in the present study for (1) developing a foundation for understanding HIV risk behavior of street youth from the perspective of individualistic models, (2) identifying areas of limited knowledge requiring further investigation, and (3) generating initial areas of inquiry based on points of incongruence between current theories and the realities of street youths' lives.
Trends in HIV Research and Intervention
   Over the course of the AIDS epidemic, public health strategy has shifted from a broad prevention focus to a more targeted one (CDC, 1994).  Early in the epidemic, when the AIDS virus was being identified primarily in homosexual men, the gay community responded with community level prevention efforts.  Rates of new infection decreased, primarily through social diffusion of safer sex practices as a result of the well-organized social structure of those communities (Phillips, 1991).  As new AIDS cases arose in heterosexual communities, the major strategy for limiting the spread of AIDS was through public education about HIV, its routes of transmission and ways of preventing it (Linville & Fischoff, 1993).  That approach succeeded in raising the general level of knowledge of HIV/AIDS and prompting an increase in condom use in the general population (CDC).  As a result, the rate of new infection in the general population plateaued and declined (CDC, 1997).  However, certain subgroups remain disproportionately affected by the epidemic (CDC, 1998b).  Epidemiologic studies show that the HIV epidemic actually consists of “multiple sub-epidemics that are constantly evolving and are intertwined with other significant social problems such as poverty [and] homelessness” (CDC, 1998b)  Street youth are one subgroup that has experienced nondeclining rates (Rotheram-Borus & Koopman, 1991), as are other marginalized groups, such as women and ethnic minorities (CDC 1998c; Rosenberg & Biggar, 1998).
   In response to this pattern, current HIV prevention efforts typically target high-risk segments of the population (CDC, 1994).  With this shift, research began to focus on identifying high risk groups and the specific behaviors that place them at risk through epidemiological studies (Prochaska, Redding, Harlow, Rossi, & Velicer, 1994).  Epidemiology, however, can only suggest “who” is in need of intervention and “what” behaviors represent risk, providing no guidance as to “how” to intervene.  That is the role of theory.
   Interventions that target high risk groups, when based on theory, have drawn on traditional models of health behavior. The most innovative population-specific programs tailor program content (derived from traditional models) to the culture, norms, and beliefs of the target group in an effort account for “context”.  This approach facilitates appropriate communication of intervention content, but the content does not necessarily address the contextual whole that gives rise to the risk behavior.  Although it has been moderately successful in changing the HIV risk behavior of certain groups (CDC, 1999; Fisher, Fisher, & Rye, 1995; Jemmott, Jemmott, & Fong, 1998; National Institutes of Mental Health [NIMH] Multisite HIV Prevention Trial Group, 1997), it is unclear whether those changes are clinically significant.  Further study is needed to determine whether such “tailored” programs produce long-term behavior change in high-risk populations.  
   Recently, criticism of traditional individualistic models suggests that they may have reached their useful limit in HIV prevention with the populations that are currently most vulnerable to HIV infection because they fail to account for contextual influences on risk behavior (Choi & Coates, 1994; DiLorenzo & Hein, 1995; Kashima et al., 1993; Parker, 1996; Tawil et al. 1995).   In step with that criticism, a third phase of HIV prevention strategy seems to be evolving.  In contrast to the individualistic view of traditional models, this approach recognizes the embeddedness of sexual risk behavior in larger systems, taking a broader view of context than traditional individualistic models are able to capture.  
Traditional Models of Health Behavior
   Recent theory-based HIV prevention programs have drawn primarily on three models of health behavior: the Health Belief Model (Lux & Petosa, 1998), the Theory of Reasoned Action/Theory of Planned Behavior (Fisher, et al., 1995; Jemmott et al., 1998; NIMH Multisite Prevention Trial, 1997; St. Lawrence et al., 1995), and Social Cognitive Theory (Jemmott et al., 1998; NIMH Multisite Prevention Trial).  These models have been vigorously and successfully applied to a broad range of target behaviors and populations, including mammography use (Clanan, 1984; Taplin & Montaño, 1993), alcohol abuse in early adolescence (Williams, 1995), fruit and vegetable intake among children (Domel, 1993), and weight reduction with overweight adults (Mischel, 1973).  Together, they form the bedrock of modern public health HIV prevention theory (Fishbein & Guinan, 1996).
   The Health Belief Model (HBM: Becker, 1974; Rosenstock, 1974) was initially advanced in a Finnish experiment on cardiovascular disease.  It proposes that the likelihood of taking a recommended preventive health action is directly influenced by “perceived threat” of disease and the valence of “perceived benefits” (beliefs regarding effectiveness of the action in avoiding the disease) minus “perceived barriers” to taking to action (such as expense and embarrassment).  Perceived threat is impacted by whether one feels personally at risk of contracting the disease (“perceived susceptibility”) and feelings concerning the seriousness of the disease (“perceived severity”), as well as environmental cues from the environment that support taking the action, such as media information and advice from others (“cues to action”).  Demographic and sociopsychological variables such as race, age, and peer pressure are thought to modify perceived threat, susceptibility and severity, as well as benefits and barriers to action (See Appendix B).  
   The original HBM was later modified to include two additional constructs:  self-efficacy and fear.  Bandura (1977) defined self-efficacy as the belief that one can successfully execute a behavior required to produce a particular outcome.  Lack of self-efficacy is thought to be a barrier to action.  Rogers' (1975) Protection Motivation Theory described the role of fear-arousing cues to action in increasing perceptions of severity, susceptibility and threat.
   HIV interventions based on the HBM attempt to modify behavior by increasing perceptions of susceptibility to HIV and severity of the threat, often by arousing moderate levels of fear, reinforcing images associated with the positive benefits of prevention, and providing information and skill training to boost self-efficacy.  
   According to the Theory of Reasoned Action (TRA: Fishbein & Ajzen, 1975), the most important determinant of behavior is behavioral intention, or the perceived likelihood of performing the health behavior.  Behavioral intention, in turn, is determined by attitudes toward the behavior (belief that the behavior is linked with certain outcomes and evaluation of their value) and subjective norm (belief about the value that important referent persons attach to the behavior and motivation to do what each referent thinks).  The TRA's focus is broader than the HBM's.  It incorporates the role of peer norms which are thought to have a strong influence on adolescent risk behavior (Brown, DiClemente, & Reynolds, 1991).  
   The TRA was extended by Ajzen and Madden's (1986) Theory of Planned Behavior (TPB), which added the construct of perceived behavioral control to the model.  TPB states that, independent of the subjective norm and attitude toward the behavior, behavioral intention is directly affected by perceived behavioral control.  The effect was found to be strongest for behaviors that are not fully under volitional control and weakest when an individual perceives that they have more control over a behavior than they actually have (Madden, Ellen, & Ajzen, 1992).  
   A strength of the TRA/TPB is its flexibility.  TRA/TPB models are tailored to particular populations and target behaviors by exploring how the constructs may apply to the target population (Middlestadt, Bhattacharyya, Rosenbaum, Fishbein, & Shepherd, 1996).  Through open-ended elicitation, the components of the model take shape based on the attitudes, beliefs and norms that members of the target population describe as important determinants of their behavior.  Like the HBM, however, TRA/TPB is an individualistic theory.  The “context” of behavior resides within the individual, limited to one's attitudes, beliefs and perceptions of norms.  An additional weakness is that when the TRA/TPB is applied to condom use behavior, it is too simplistic.  The attitudes, beliefs and norms that contribute to condom use behavior are likely to vary depending upon the setting, type of partner, and type of sexual act involved.
   Social Cognitive Theory (SCT) is a synthesis of a number of disparate theories of health behavior, including the Health Belief Model and the Theory of Reasoned Action (Bandura, 1986).  The core concepts of SCT are reciprocal determinism and self-efficacy.  Self-efficacy is the confidence that a person feels about performing a certain behavior, including confidence in one's ability to overcome barriers performing that behavior (Bandura, 1982; 1986).  Reciprocal determinism is the notion of dynamic interaction between personal determinants (cognitive, affective and biological factors), environmental influences, and behavior (Bandura, 1986).  Bandura stresses that behavior itself is influential:  “Behavior, internal personal factors, and environmental influences all operate as interlocking determinants of each other” (Bandura, 1978, p. 346).  By introducing the notion of reciprocal determinism, Bandura reframed the tension between environmental versus individual influences on behavior, placing the individual in a more active role in determining his own environment.  “It is largely through their actions that people produce environmental conditions that affect their behavior in a reciprocal fashion” (Bandura, 1978, p. 345).  
   The goals of SCT-based HIV prevention programs are often to  (1) provide HIV information to increase people's awareness and knowledge of health risks, (2) develop social and self-regulative skills needed to translate informed concerns into effective prevention action, (3) enhance skills and build self-efficacy by providing opportunities for guided practice and corrective feedback, and (4) develop social supports for the desired HIV-protective behavior (Bandura, 1994).   The notion of “environmental influences” is thought to extend causal factors in behavior beyond the individual.  However, since reciprocal determinism is considered a principle or postulate, it is not submitted to empirical tests.  As a result, critics of SCT complain that the specific relationships among its elements (person, environment, and behavior) have not been sufficiently defined (Baranowski, Perry, & Parcel, 1997).  Indeed, the nature of reciprocal relationships between the person, the environment, and behavior are inherently complex.  They are likely to vary across different populations and for different health behaviors.  Beyond postulating that these complex relationships exist, SCT does not provide a framework for empirically testing them or systematically mobilizing or controlling them for health promotion.
   Furthermore, SCT encompasses eight major concepts in addition to self-efficacy and reciprocal determinism (e.g., behavioral capacity, expectancies, self-control, etc.), without specifying the nature of any relationships among them.  In sum, SCT has been criticized as being so vague and so broad that it is difficult to falsify, since it explains “everything and nothing” (Baranowski et al., 1997).  While it appears that SCT-based approaches which teach social and self-management skills may help change the HIV risk behavior of a number of different populations (Fisher & Fisher, 1992; Kim, Stanton, Li, Dickersin, & Galbraith, 1997; Peterson & DiClemente, 1994), this is not surprising, since SCT is such a broad amalgamation of other models.
Efficacy of Theory-Based HIV Interventions
   Bridging the gap between theory and practice has been an ongoing struggle for HIV researchers and interventionists (Choi & Coates, 1994; Glanz, Lewis, & Rimer, 1997; Shriver et al., 1998).  For several reasons, it is difficult to draw conclusions about the relative efficacy of various theoretical approaches to HIV prevention.  Often, the match between the theoretical models used in research and actual prevention programs is tenuous.  Sometimes it is altogether absent.  In a review of the public health literature, Glanz and Oldenburg (1996) found that slightly less than half of the articles relevant to health behavior and health promotion reported on explicit use of one or more theories. Those programs that do apply health behavior models rarely utilize the full, formal theoretical model in their design.  Typically, program designs are based on a few concepts from one or more models (CDC 1996; IOM, 1994; Maticka-Tyndale, 1995).
   Ideally, competing theories might be compared by examining outcome evaluations of the interventions based upon them. However, in spite of the voluminous behavioral literature on AIDS, there have been relatively few published evaluations of explicitly theory-based HIV preventive interventions with adolescents (Damond, Breuer, & Pharr, 1993; Jemmott & Jemmott, 1992; Jemmott et al., 1998; Jemmott, Jemmott, & Spears, 1992; Kipke, Boyer, & Hein, 1993; Rickert, Gottlieb, & Jay, 1990; St. Lawrence et al., 1995; Slap, Plotkin, Shalid, Michelman, & Forke 1991; Walter & Vaughan, 1993).  Among these evaluations, methodological limitations and design issues make it difficult to compare the relative efficacy of various theories.  For example, data on long-term behavior change is often not collected, so the stability of outcomes across models can rarely be assessed.  Causal inferences are often weakened by failure to randomly assign participants to study conditions.  Direct cross-study comparison of outcomes is difficult because there is no consensus regarding outcome measurement (Kim et al., 1997; Pinkerton et al., 1998), and the outcomes reported sometimes exclude measurements of behavior, focusing instead on changes in theoretical constructs such as attitude or intention.  These methodological and design-related problems have made it difficult to perform meaningful meta-analyses of HIV prevention programs.
   In response to the need for a centralized source of rigorous intervention data, the CDC began developing the Prevention Research Synthesis (PRS) Database.  The PRS includes national  and international studies of HIV sexual risk behavior interventions conducted from 1988 to the present.  Each study meets the CDC's criteria for relevance (e.g., behavioral or HIV/STD outcomes) and methodological rigor (CDC, 1998a).  Sogolow et al. (1998) examined the 22 PRS studies that reported sufficient safer sex data for meta-analysis (of the 72 US-based studies in the database at that time) and found an overall 14% risk difference favoring those who received an intervention.  Using the same database, with somewhat stricter inclusion criteria, Neuman et al. (1998) found an average of 18% reduction in sexual risk behavior in the 10 studies of heterosexual adults that they examined.  Neither study attempted to compare the relative efficacy of different theoretical approaches.  
     Beyond methodological and design issues, the merit of published HIV program  evaluations must be assessed in terms of significance of the reported outcomes.  Statistically significant differences between control and intervention groups do not necessarily mean that the findings carry any practical clinical significance, or that the intervention is useful from a public-health standpoint.  As Choi and Coates (1994) observe, “[T]here is no standard to determine whether program results have public-health significance in terms of probable reductions in HIV incidence.  Analyses are needed and consensus is required to provide standards for the kinds of effects needed in order to consider a program useful from a public-health standpoint” (Choi & Coates, p. 1385).
   A well-designed, methodologically rigorous study by Jemmott et al. (1998) illustrates the problem of “significance.”  They conducted a randomized, controlled trial of a safer sex intervention based on social cognitive theory and the theory of reasoned action.   The results at 3-month follow-up were statistically significant: among those who received the intervention, 65.6% (21/32) reported using condoms every time they had sex following the intervention, compared with 36.1% (13/36) of the control group (p <.02).  Presumably, between 11 and 17 of the intervention participants would have consistently used condoms at follow-up if they had not received the intervention based on the condom use rates reported by the control group (36.1% and 51% reported consistent condom use at 3- and 12-month follow-up, respectively).  Among the intervention group, 21 reported consistent condom use following the intervention.  In practical terms, the intervention appears to have influenced the behavior of as few as 4 or as many as 10 participants (12.5% to 31.2% of the intervention group).   
   This intervention had an important impact even if only 4 participants changed their behavior, since any shift toward HIV protective behavior is a desirable outcome.   However, as Choi and Coates (1994) observed, no standard exists by which to assess the potential public health significance of such an impact.  Currently, an alpha level of .05 is the only threshold beyond which an intervention is deemed “effective.”  From a public health perspective, though, beyond that threshold is a wide range effectiveness levels.  In judging program effectiveness, the oversimplified dichotomy between statistically significant and non-significant behavioral change may invite complacency.  It can support the illusion that a traditional theory-based intervention represents the “state of the art” when in fact a large proportion of variance goes unexplained, indicating that the theory should be improved upon.  Undoubtedly, if a traditional medical intervention were successful 12.5 to 31.2 percent of the time, research would intensively pursue more effective approaches.
        The stability of behavioral change over time is a second critical aspect of program effectiveness that is too rarely assessed.  When longitudinal behavior change data is collected, program impact often appears short-lived by statistical standards.  In the Jemmott et al. (1998) study, at 6-month follow-up, there was no statistical difference in consistent condom use between the intervention and control groups (50% versus 37.5%, p = .29) and 12-month follow-up group differences were even smaller (62.5% v. 51.2%, p <.35).  The intervention effect appeared to be transient, suggesting that this intervention would not be useful in producing a long-term shift to consistent condom use among African-American adolescents at risk for HIV through sexual transmission.  
   Among studies that do report significant behavior change at follow up, the tendency to focus on statistically significant group differences can distort an intervention's practical impact.  St. Lawrence et al. (1995) predicted that low-income African-American adolescents who received SCT-based behavioral skills training in additional to an HIV educational program (behavioral skills training “BST” group)  would report more condom-use behavior at follow-up than those who only received the HIV educational program (education control condition “EC” group).  At 6- and 12-month follow-up, the BST group did report a significantly higher percentage of condom-protected intercourse occasions than the EC group (F [1,134] = 5.94, p <,05), however the effect must be attributed to a decline in condom use among the EC group rather than to behavioral gains in the BST group.  The practical impact of the intervention on the BST group was essentially a return to pre-intervention levels by 12-month follow-up.  At best, the intervention averted the decline in protective behaviors that was seen in the EC group.  The study's conclusion that “[Y]outh who were equipped with information and specific skills lowered their risk to a greater degree... (St. Lawrence et al., p. 221)” is not supported by the follow-up data on the study's central measure of long-term program effectiveness.  Of course, the short-term behavior change that BST produced is more desirable than none at all, but to advance BST as a superior intervention approach is to overstate its practical long-term impact.
Review of HIV Preventive Interventions
   Only a limited number of programs targeting youths' sexual HIV risk behavior have been empirically evaluated (St. Lawrence et al., 1995).  Even fewer programs for street youth have empirical support.  As we briefly examine the research, it is important to keep in mind the limitations of published evaluations, specifically, that “program effectiveness” and “statistical significance” are synonymous throughout these studies.
   Interventions targeting adolescents.  Of the programs that have demonstrated reduced HIV risk behavior among adolescents, most have been based on the principles of Social Cognitive Theory.  Kirby and Coyle (1994) reviewed 11 published evaluations of school-based HIV prevention programs and found that, of the five programs that were effective, all were SCT-based.  Among the 15 HIV prevention programs included in the Program Archive on Sexual Health & Adolescence (PASHA: Card, Niego, Malliari, & Farrell, 1996 ), seven use strategies to enhance self-efficacy, and all of them included a behavioral skills development component.  Choi and Coates' 1994 review found that, among the four evaluations of classroom-based interventions published between 1991 and 1994, only the program that provided AIDS education and skills training reported statistically significant improvements in risk behavior scores at 3-month follow up.
   In addition to pointing to the current widespread use of SCT, reviews of this literature illustrate the extent of the methodological problems described earlier.  In Choi and Coates' 1994 review of HIV interventions for young adults, of the nine interventions reviewed, only three used random assignment to groups, only one assessed behavior change, and that study included no long-term behavior change data.  In Stanton, Kim, Galbraith, and Parrott's 1996 review of adolescent HIV-risk reduction interventions, 5 of the 21 studies examined explicitly stated a theory upon which the intervention was based.  Only one of those studies provided follow-up at six months.  The median duration of follow-up was one week.  Given these methodological problems and the paucity of long-term behavioral measures, it is difficult to assess whether the theories they are based on produce long-term behavior change among adolescents.
   Interventions targeting street youth.  A number of HIV prevention interventions with street youth have been published (e.g. Goulart & Madover, 1991; Huba, 1998; Schneir, Kipke, Melchoir, & Podschun, 1993; Tenner, Trevithick, Wagner, & Burch, 1998; Woods et al. 1998), however, they are typically descriptive only, usually do not identify a theoretical foundation, and rarely used behavioral measures evaluate program efficacy.  For example, Tenner et al. describe Seattle's YouthCare, one of 10 HIV/AIDS “Special Projects of National Significance” supported by the U.S. Health Resources and Services Administration.  The program provided services to youth “on the margins”, including homeless and runaway youth.  YouthCare ran “peer-based programs which provide support and education” for youth at high risk of HIV infection.  Only formative evaluation data was reported.  No outcomes were examined.  Similarly, the Boston HAPPENS Program (another “Special Project of National Significance”) was merely described by Woods et al.  No evaluation is reported.  Goulart and Madover described HIV prevention “strategies” of the Larkin Street Youth Center in San Francisco:  providing information on safe sex practices, dispensing condoms and bleach and assessing HIV/AIDS knowledge.  Again, no evaluation was done.  Podschun's description of the Teen Peer Outreach-Street Work Project in San Diego focused primarily on the role of outreach workers.  Podschun cited budget constraints as the reason no outcome evaluation was conducted.  Schneir et al. described the evaluation of another “Special Project of National Significance” through Children's Hospital Los Angeles (CHLA).  Measuring sexual risk behavior change (or any behavior change) was not one of the eight stated goals of the evaluation.  In spite of the absence of efficacy data, one of the eight goals was “to document the CHLA service model so that successful aspects may be disseminated and replicated” (Schneir et al., p. 640).
   Only one published program has empirically demonstrated effectiveness at changing HIV risk behaviors among street youth.  This intensive program used SCT as an underlying conceptual framework (Rotheram-Borus, Koopman, Haignere, & Davies, 1991).  An integral component of the intervention was providing the youth with referrals to comprehensive care, including food, shelter, and mental health services.  These elements of the intervention, while not directly related to HIV risk behaviors, altered the basic circumstances of the youths' lives and were considered critical to the effectiveness of the program (Rotheram-Borus, Koopman & Ehrhardt, 1991).  As the success of that intervention suggests, providing street youth with a stable, supportive environment in which their basic needs are met is an effective way to change their HIV-risk behavior (Anderson et al., 1994; Luna & Rotheram-Borus, 1992; Rotheram-Borus et al., 1994), since HIV-risk behavior is embedded in the context of street life.  Eliminating the context of youth street life is a legitimate long-term public policy goal, but not enough resources are currently available to provide all street youth with what appears to be the most potent element of this intervention -- a stable, supportive environment (Rotheram-Borus, 1991).
   In sum, although traditional models of health risk behavior have successfully been used to reduce a variety of health risk behaviors, they have fallen short of demonstrating consistent, significant, changes in HIV risk behavior.  Furthermore, theory-based studies that have demonstrated statistically significant HIV-related behavioral changes (such as Jemmott et al., 1998) have not necessarily demonstrated sustained, practical significance.  A number of potential reasons for this shortfall have been advanced, most of which focus on the unique characteristics of HIV itself and the unique qualities of the groups to which traditional models have been applied.  Many of these critiques become more salient and valid when viewed through the lens of the context of street life.
The Limitations of Traditional HIV Research
   Traditional models of HIV risk behavior focus extensively on intra-individual factors and fail to capture larger contextual factors that contribute to risk behavior.  In this regard, two lines of criticism of the models themselves have been prominent: (a) that generalizing traditional models to HIV is inappropriate because of the unique features of HIV as a health risk, and (b) that traditional models are not necessarily appropriate representations of the HIV risk behavior of the groups that are most vulnerable to HIV.   In a larger sense, traditional models must be recognized as the product of the prevailing Western approach to science.  As a result, they are limited by the epistemological assumptions inherent in that approach.  The limitations of traditional models of HIV risk behavior will be examined on both of these levels.
Conceptual Limitations of HIV Risk Models
   The use of traditional health risk models to address HIV risk behavior has been widely criticized based on the differences between HIV risk behavior and other forms of health risk behavior.  Four central differences set HIV risk behavior apart from other health risk behavior: the seriousness of HIV, the ambiguity of the consequences of risk behavior, the social and emotional complexity of sexual behaviors, and the unique qualities of sexual decision-making. (Kashima et al., 1993; Satcher, 1996)
   AIDS is among the most serious of health threats, and all of the models reviewed above rely on perceived threat to mobilize HIV-protective behavior.  The link between perceived threat and behavior has been supported in the literature, primarily in reviews by Becker (Becker, 1974; Janz & Becker, 1984).  However, in the studies reviewed, evidence of the link between perceived threat and preventive behavior was weaker in the context of extreme health threats, such as uncontrolled type I diabetes, than in the context of less serious health threats (Montgomery et al., 1989).  So, while there is general agreement that perceived threat does predict a variety of preventive behaviors, there is also consensus that theories that rely upon it were designed to address non-fatal, reversible health threats, and therefore probably do not generalize to HIV risk behaviors (Emmons et al., 1986; Gerrard, Gibbons, Warner, & Smith, 1994).
   In addition, unlike various other health threats to which traditional models have successfully been applied, the consequences of HIV risk behavior are ambiguous and not immediate.  HIV transmission is a function of the nature and frequency of sexual contact, the infectivity of the donor, and characteristics of the recipient (Lawrence, Jason, Holman, & Murphy, 1991).  Given this array of variables, there is not a one-to-one relationship between exposure and infection.  However, unlike other health risks such as smoking or unhealthy eating, HIV can be transmitted with relatively few instances of risky behavior (Satcher, 1996), but avoiding HIV demands highly consistent behavior change with very little margin for error, making intervention “a challenge virtually unprecedented in the behavioral sciences” (Kelly, Murphy, Sikkema, & Kalichman, 1993, p. 1023).  Adding to the ambiguity of HIV risk, the delay between HIV infection and the emergence of symptoms can be up to 10  years (Carre et al., 1997; Hendricks et al., 1998). Such a long delay between a risk behavior its consequence can contribute to an illusion that past risk behavior was without consequence.  In sum, the threat of HIV is neither certain nor proximal to the risk event, and is dependent on fewer risk episodes than other health threats.  As a result, the motivational properties of perceived threat that have been shown to operate in the context of other health risks are likely to be eroded in the context of HIV risk (Gerrard et al., 1994).
   A third weakness of traditional models is that condom use occurs in a context that is more socially and emotionally complex than other health-related behaviors.  Individualistic models cannot account for the fact that the decision to use a condom is a negotiated one, involving more than one person (Kashima et al., 1993).  In a larger sense, dominant models that focus narrowly on individual, intrapsychic processes do not easily accommodate contextual personal and sociocultural variables such as gender roles, cultural values, and the nature of the relationship in which sexual activity occurs (Dolcini et al., 1993; IOM, 1994; Peterson, Catania, Dolcini, & Faigeles, 1993; van Campenhoudt, Cohen, Guizzardi, & Hausser, 1997).  Narrowing the focus to individual-level phenomena also obscures factors that are likely to affect the HIV risk behavior of women disproportionately.  From the perspective of multilevel analysis, women are at a disadvantage in enacting HIV protective behavior in terms of power in heterosexual relationships (Amaro, 1995; Soet et al., 1999; Worth, 1989)  internal pressure to conform to the dictates of gender discourse (Moore & Rosenthal, 1998), and the effects of earlier sexual coercion and victimization (Harlow et al., 1992).  
   Another common criticism is that traditional models represent HIV-risk decision-making as a rational, linear process.  Inherent in traditional models is the assumption that sexual encounters are regulated by self-formulated plans of action, by individuals who are acting in a volitional, intentional manner (IOM, 1994).  However, sexual behavior is often impulsive and, at least in part, physiologically motivated.  A well-formulated plan of action that is the product of a careful weighing of potential harms and benefits can be dismissed in the context of a passionate sexual encounter when competing proximal goals (i.e., sexual gratification) offset well-informed intentions (i.e., condom use) (IOM).  Among women, behavioral intention can also be offset by an inability to convince a partner to use a condoms (Soet et al., 1999).  Among women with limited access to resources, such as street youth, competing goals may include survival goals if she relies on her sexual partner for essential resources (Worth, 1989).  
   In addition to the preceding arguments against applying traditional models to HIV risk behavior, another line of criticism focuses on the appropriateness of applying the specific constructs involved to certain populations --- particularly adolescents and marginalized groups, and especially street youth.  When viewed through a contextual lens, many of the assumptions underlying traditional health risk constructs appear to be incongruent with the lives of street youth.  
   No existing model of health behavior reflects the unique cognitive, maturational and developmental influences that are critical to understanding adolescent behavior (Brown et al., 1991).  For example, current models presume that decisionmakers of all ages are capable of a rational thought process in which an array of relationships between behavior and consequences are weighed and evaluated.  This process is characteristic of formal operational thought which is the last, most mature stage of cognitive development.  Only adolescents who have reached cognitive maturity can conceptualize the many possibilities in a system and reason on the basis of his or her own mental constructions (Elkind, 1967).  Environment plays an important role in an adolescent's cognitive development (Rogoff, 1993; 1997; Sentrock, 1998).   General cognitive development is associated to some extent with the degree to which one's cultural environment provides opportunities to practice cognitive skills (Sentrock) and the development of formal operational thought has been found to relate to adolescents' schooling experience (Rogoff, 1993, 1997).  General cognitive development may be fostered through “cognitive socialization”, which promotes cognitive growth through guidance from individuals who are skilled in the culture's tools.  For some people, the transition from concrete to formal operations is delayed until late adolescence or beyond.  This may be particularly true of youth who have spent considerable time on the street, out of school, and focused on immediate, concrete survival tasks.
   Not only do many adolescents approach decisionmaking with a limited set of cognitive skills, they are also more susceptible than adults to the “personal fable”, a form of cognitive bias with significant implications for the process of HIV-risk decision making (Elkind, 1967; Moore & Rosenthal, 1991).  The personal fable is a form of cognitive egocentrism that permits adolescents to perceive themselves as invulnerable to the consequences of their own risk behavior in spite of recognizing that others are vulnerable (Gruber & Chambers, 1987).  This sense of invulnerability can provide adolescents who continue to engage unsafe sexual activity with an illusion that they are protected from infection (Greig & Raphael, 1989).  In street youth, this tendency may be even more pronounced.  A strong sense of invulnerability may be a necessary, even adaptive response to the uncompromising challenges of street life.  
   Among marginalized groups that are at highest risk for HIV, the appropriateness of applying models that rely on self-efficacy as a pivotal construct is questionable.  For these vulnerable populations, such as ethnic minorities, women, and street youth, who experience oppression and powerlessness, self-efficacy is linked in complex ways to the context of their lives.  Traditional models do not account for a more generalized level of self-efficacy that may be the product of marginalization.  Kelly (1995) defines this generalized self efficacy as “whether or not one believes that any personal steps will change any significant life circumstances, which may be necessary to motivate taking action across a variety of life problem areas, including HIV risk behavior “ (p. 127).  Among street youth, low generalized self-efficacy is likely to be common.  As one social service provider observed, “Some of these kids don't feel they have any control over their lives, like they're just floating down the river of life with no pilot (Sherman, 1995, p. 2)
   A related assumption of most health behavior models is that behavior change occurs in the context of a positive future orientation.  Presumably, an investment in future well-being underlies decision making and, ultimately, behavior change.  However, as discussed earlier, street youth are cut off from the resources and pathways that mainstream society offers for achieving status and security, and are unlikely to view their future optimistically (Cwayna, 1993; Greenblatt & Robertson, 1993; Luna & Rotheram-Borus, 1992).
   Another tacit assumption, shared by theories that rely on notions of susceptibility or vulnerability, is that the value of a given health threat is absolute, rather than relative.  Even youth who have a positive future orientation must weigh the threat of HIV infection within the context of other threats to their health and safety.  Among street youth, providing for immediate survival needs often takes precedence over long-term concerns such as HIV (Luna, 1991).  Health behavior theories lack a dimension that would weigh the subjective threat of HIV in the context of other, perhaps more salient health and survival risks that many street youth regularly face.
Epistemological Limitations of Current Research
   Despite increasing recognition that HIV sexual risk behavior involves complex and multidimensional processes, current models provide very limited insight into how HIV sexual risk is experienced, understood and managed by those individuals who are most vulnerable to HIV infection.  At the core of the various criticisms of traditional models is a basic difficulty with their philosophical and methodological approach to understanding psychological phenomena.  
   The philosophical problem is that traditional models of HIV sexual risk behavior reflect a trait-oriented philosophy.  The unit of analysis is the individual and the primary determinants of psychological functioning are person-oriented variables such as self-efficacy and perceived threat, which have virtually no relationship to physical settings or social contexts (Altman & Rogoff, 1987).  This individualistic perspective is consistent with the prevailing Western positivist approach to science: the individual “self” is detached from the world (Reason, 1994).
     Methodologically, the problem has been that much of the research on HIV risk-taking behavior does not appear to be adequately grounded in the perspectives and experiences of the marginalized groups that are most at risk for HIV.  Empirical support for the HBM, TRA and SCT is based predominantly on quantitative studies of non-marginalized populations and often examines health threats other than HIV.  In practice, the implication for prevention is that those models provide a priori hypotheses about factors that contribute to health risk behavior that can be “translated” into HIV interventions that are appropriate for marginalized groups.  This practice of theory generalization is supported by the positivist assumption that, until falsified, theories represent an objective “truth statements” that describe a knowable reality which is independent of time and context (Lincoln & Guba, 1985).
Rationale for the Present Study
   The purpose of the present study was to contribute to the development of a more accurate model of HIV sexual risk and protective behavior among female street youth by (1) examining HIV sexual risk behavior among street youth from a philosophical perspective that would link behavior to the context in which it occurs; and (2) applying a methodology that would allow  street youth to articulate the processes relevant to their HIV sexual risk behavior.  The product of this research will be a collection of constructs that incorporate multiple levels of contextual influence on the HIV sexual risk behavior of White female adolescent street youth.
   The philosophical foundation of this research is an ecological perspective, which maintains that human behavior occurs in a context and can only be understood in relation to that context (Kelly, 1979).  A central tenant of the ecological perspective is that people negotiate and construct their realities in reciprocal interaction with multiple levels of their context (Bronfenbrenner, 1979).  The current study utilized a grounded theory development methodology, a qualitative research approach.  Grounded theory provides a systematic set of procedures to develop an inductively grounded theory about a phenomenon, based on the detailed reports of members of the population of interest (Strauss & Corbin, 1990).  This methodology provided a vehicle for investigating phenomena from a “emic” perspective, that would emphasize the attributions, meanings, and experiences of the participants, rather than the “etic” viewpoints of the observer (Strauss & Corbin, 1994; Watts, 1999).  A transactional approach provided additional philosophical structure for the analysis of the qualitative data.  Based on this approach, unit of analysis will be discrete events that are intrinsically embedded in historical, situational and physical contexts.  Holism is the essence of the transactional approach, in that events represent a confluence of people, space and time which are meaningful only in relation to each other (Altman & Rogoff, 1987).  A transactional approach is particularly compatible with grounded theory methodology, in that both seek to identify underlying patterns across similar events and maintain an openness to emergent explanatory principles (Altman & Rogoff; Strauss & Corbin, 1990).
   The study will look specifically at the HIV sexual risk and protective experiences of White female adolescents who are without a stable source of shelter at the beginning of the study.  The study will be limited to participants of this gender and race because the literature suggests that gender and culture may be particularly relevant to understanding the context of condom use behavior.  The study will be limited to participants without a stable source of housing because the term “street youth” (or more commonly “homeless youth”) typically includes subgroups whose housing situation is actually stable (i.e., youth in foster care).  Limiting the sample this way will allow a more focused examination of HIV sexual risk and protective behaviors in the context of housing instability.
   The present study's focus on young women is based on the premise that gender-related issues play a crucial role in heterosexual condom use (Cochran & Mays, 1993; Worth, 1989).  The ability of women to negotiate when and how they engage is sexual activities is embedded in culture-bound gender roles and gender-based power relations. As a result, the goals, meanings, experiences and expectations that females draw upon in approaching condom use are likely to differ in important ways from those of males.
   In addition, the empirical literature indicates that males and females differ on variables related to condom use.  Perceiving that one's partner has a positive attitude toward condom use is associated with greater use of condoms by women than men, as is communication about condoms with one's partner (Sheeran, Abraham, & Orbell, 1999).  While men are more likely to purchase and carry condoms, they tend to rely on their female partners to negotiate condom use (Carter, McNair, Corbin, & Williams, 1999).  Consistent gender differences have also been found in attitudes toward condom use (Campbell, Peplau, & DeBro, 1992; Helweg-Larsen, & Collins, 1994; Sacco, Levine, Reed, & Thompson, 1991).
Cultural differences in psychosocial factors such as gender power roles (De La Cancela, 1989) and negotiation of safer sexual practices (Worth, 1989) may be major determinants of HIV sexual risk behaviors.  Therefore, it is important to understand HIV risk behavior of specific cultural groups in order to avoid over generalization and to support the development of culture-specific HIV preventive interventions.  Since the interviewer was European American, White Non-Hispanic youths were selected in order to minimize the effect of intercultural differences on the research process.
Guiding Research Questions
    The following broad research questions were generated by the preceding review of the existing literature on HIV sexual risk behavior, descriptive accounts of the lives of street youth, and through the researcher's prior contact with and qualitative interviews of street youth.  Specific interview items associated with each broad research question appear in the interview guide (Appendix C).
    Research Question I:  What gender-based power dynamics occur during condom negotiation? The goal of this question is to understand whether and how participants experience differences in power between themselves and their partners, and to explore how they manage such differences in the process of condom negotiation.
    Research Question II:  In what ways do condoms function symbolically in sexual encounters?  What is the relationship between these symbolic functions and condom use?  This research question will broadly address the potential meaning of condom use in a variety of relationship contexts, both specifically (in the context of a particular relationship) and globally (across different relationships).  In exploring condom symbolism, there will be an emphasis on whether respondents consider the notion of the “idealized union” (the cultural ideal of a trusting, monogamous relationship) to be a salient feature of condom symbolism and negotiation.
    Research Question III: What other features of the immediate context are evident in the condom negotiation process reported by respondents?  The focus of this research question is on discrete sexual encounters and their specific features.  “Context” here is broadly defined, and may include intra-individual factors such as those posited by traditional models of health risk behavior (i.e., perceived threat, behavioral intention, etc.), and factors that influence the participant's state of mind, such as mood, level of sexual arousal, or drug use, as well as concrete features of the environment, such as the availability of condoms.  Interview questions will not specifically address such factors, but participants will be encouraged to articulate them if they emerge.
    Research Question IV: How is prior personal and/or vicarious experience of sexual or physical victimization related to assertiveness in condom negotiation?  How is the participant's perception of possible negotiating options influenced by prior personal and/or vicarious experience of sexual or physical victimization?  This research question will focus on whether and how past exposure to sexual or physical victimization might be associated with (a) the range of negotiating stances (i.e., assertiveness, passivity, withdrawal from negotiation) that participants perceive as being available to them; (b) the reactions that participants anticipate from their partners in response to behaving assertively during condom negotiation; (c) participants' perception of other potential consequences or outcomes of various negotiating stances. Participants will be encouraged to articulate any past personal and/or vicarious experiences that contribute to their perceptions surrounding these issues.
Research Participants
   The participants will be 12 to 15 White female adolescents between the ages of 14 and 21 who report having had heterosexual anal for vaginal sex and meet the criteria for “being on the street.”  “Being on the street” will be defined being currently without a consistent  place to sleep, not including shelters, jails or “squats” (self-appropriated space that provides shelter, such as an abandoned building, to which one has no legal claim).  “Consistent” will be defined as a place that one can  reasonably certain of occupying for the coming 30 days.  
   Subjects will be recruited primarily through Neon Street Services, a Chicago organization that provides services to homeless/street youth.  The use of a random sampling strategy would be highly impractical.  Many street youth avoid contact with social service agencies such as Neon Street because they represent “the system,” which they perceive to be indifferent to their needs and interests.  Initially, a convenience sampling strategy will be used, with Neon Street staff will helping the researcher to identify potential participants based on the study criteria.  In order to ensure that the sample includes participants who do not receive services from Neon Strreet,  participants will also be recruited through a “snowball” sampling strategy (Patton, 1990).  To facilitate snowball sampling, participants will be given several business-size cards at the end of each interview that briefly describe the study and provide a way to contact the researcher.  They will be encouraged to share the cards with White female friends.
   There will be three possible ways for a participant to become involved in the research:  (1) by indicating to Neon Street Staff when they will be available to meet with the researcher; (2) through the researcher approaching them to schedule an interview; and (3) by calling the researcher based on a referral from a friend.  
   On initial contact with the researcher, potential participants will be briefly be screened to determine whether they meet the criteria for being on the street.  Participants who are 14 to 21 years old will be asked to participate in the study if they:  (a) indicate that they have engaged in vaginal or anal sex, (b) identify themselves as White (or “Caucasian,” or “European-American,” or indicate European national origins), and (c) state that they do not know where they will be living in 30 days.  Those who are interviewed in person and do not meet the criteria for study participation will be paid $5 for their time.  Participants who are screen by telephone will not be compensated.  
   Following the initial screening, consent will be obtained from each participant (See Informed Consent, Appendix D).  The researcher will read the consent form aloud before asking the participant to sign the form.  Both the researcher and the participant will be required to sign the form before the study interview begins.  Participants will be given a copy of the consent form to keep for future reference.  Although some of the participants will be under 18, parental consent will not be obtained.  Not only would it be impractical, but lack of consistent contact with parents or guardians is a defining feature of this population.  It has become common practice to waive parental consent requirements in studies of street youth.
   All interviews will be conducted in one of three rooms a the Neon Street Drop-in Center.  The rooms were chosen because they are quiet and provide privacy.  The equipment used for the interview session will include a tape recorder that uses a standard audiocasette tape located between the participant and the interviewer.  The length of the interview will vary widely across participants, but is expected to take between an hour and 90 minutes.  The interview guide (Appendix C), which is based on the guiding research questions of this study, will be used initially.  As data is collected and hypotheses are formed, the interview guide questions will be modified to explore emerging themes more closely.  Most questions will be open ended, and probe questions will be used to elicit details of the youth's experience.
   Following the interview, a date, time and location will be set for a follow-up interview in approximately 1-2 months.  Participants will be given a reminder card as well as referral cards to facilitate “snowball” recruitment.  They will be paid in cash for their participation on the following scale: $10 for the first interview, $15 for the second and $20 for the third.  After the participant leaves, the researcher will tape record notes of behavioral observations, her own responses to the content of the interview, and any unusual events or circumstances around the interview session.  The audiotape will be transcribed by the researcher into a format compatible with the NUD*ISTsoftware used for data analysis.
   It is hoped that each participant will participate in two to three interview sessions over the course of approximately six months.  The re-interview process in integral to grounded theory methodology (See Chapter III, Data Analysis).  It may be difficult to maintain contact with participants across time, since street youth are typically highly mobile.  
   To facilitate subsequent follow-up contact, participants will be asked to provide the names and addresses and phone numbers of friends, family members, and any other individuals with whom they tend to maintain contact, as well as social service agencies that they intend be in contact with.  They will also be asked to identify places that they frequent (i.e., neighborhoods, restaurants, other businesses).  Participants will also be asked about their plans for the next several months to provide “hints” about where they may be contacted.  Out of concern for participants' privacy, they will be assured that contacts will only be told that they are participating in a research project through DePaul University.   Participants will be encouraged to contact the researcher (at a phone number which has code-access voicemail) at any time.  These tracking methods have been successfully used in longitudinal studies of high-risk adolescents (Gwadz & Rotheram-Borus, 1992; Harper & Carver, 1999).  As an incentive for participants to sustain their involvement in the research, the rate of compensation will increase across the three interviews: from $10 to $15 to $20.
   Using grounded theory methodology, three waves of qualitative data will be analyzed.  The product of the analysis will be a collection of HIV sexual risk behavior contructs that include contextual influences on condom use behavior.  The constructs represent the precursors of a model of HIV sexual risk behavior.
Data Analysis
   Grounded theory development is a systematic method of research whose purpose is to generate rather than test theory.  According to Strauss and Corbin (1990), this methodology aids researchers in producing theory that is "conceptually dense" (i.e., a theory that is not merely descriptive but delineates a dense network of concepts and the relationships among them). Analysis in grounded theory development occurs concurrently with data collection, thus creating an iterative process that "grounds" theory in reality.  The concepts that emerge from the coding of initial interviews influence subsequent interview content.  The ongoing interplay between data collection and analysis is congruent with two central goals in grounded theory development: to sample events and incidents that are indicative of theoretically relevant concepts and to maximize variability in emergent categories by sampling data sources whose attributes may qualify the developing theory (Strauss & Corbin, 1990).
   In the current study, the emerging framework will be developed by alternating between two analysis or "coding" processes.  Each wave of data will be subject to open coding followed by axial coding.  Open coding involves fracturing the data into "meaning units" or concepts with descriptive names and specified properties.  Concepts that appear to pertain to the same phenomena will be grouped together to form categories, each represented by a unique code.  Axial coding involves putting the data back together in new ways by proposing connections between each category (or phenomenon) and its subcategories.  More specifically, each category will be broken up into subcategories via specification of:  (1) its properties, (2) the causal and intervening conditions that affected it, (3) the context in which it was embedded, (4) action strategies by which it was handled or carried out, and (5) the consequences of those action strategies (Strauss & Corbin, 1990).  
   In analyzing each of the three data waves, the researcher will alternate between open and axial coding as she searches for patterns of relationships between categories.  The initial open coding scheme that results from analysis of the first wave of data is subject to change as categories become more refined with each successive data wave.  The third and final phase of grounded theory data analysis, selective coding, typically occurs after data collection has ended.  It involves identifying a core category and linking all major categories to it.  Selective coding will not be performed in the present study.
   During open coding analysis of the first wave of data, a team approach will be utilized to reduce researcher bias in generating codes and establish coding reliability for the initial open coding scheme.  The team will consist of the researcher and two graduate students in psychology (Kin King, M.A., and an additional graduate student who will be provided by Karen Budd, Ph.D.).  First, the researcher will open code all of the data and develop an open coding scheme.  She will then meet with the other team members, explain her coding scheme, and solicit feedback.  The open coding scheme will be supplemented and revised as necessary based on feedback from the team.  
   The team will then use the revised open coding scheme to practice coding interviews.  Given various units of interview data, team members will use a separate coding sheet to circle any of the open coding categories that they believed were present in the data.  It may be possible that all codes, some codes or no codes will be circled for a given data unit.  During these practice sessions, inter-rater reliability will be calculated after each practice analysis.  Using simple agreement, the results for each team member will be checked against each other and the researcher.  If both agree that a particular code was present or absent for a unit, that will be considered a match.  Coding discrepancies will be discussed as they arise and the coding scheme will be revised further if necessary.  Practice will continue until at least a 75% level of agreement is reached across several consecutive unit analyses.  
   In order to formally establish the reliability of the resulting open coding scheme, each of the three team members will independently code the same 3 to 5 randomly selected interviews (from among the interviews that were not used for practice coding) without discussing their coding with other team members.  The unit of analysis will be at least a full page of text, but will be determined on the basis of natural shifts in the flow of the interview.  Shifts, which will be defined by the researcher, will include the end of the participant's response or spontaneous shift in the partipant's focus from one theme to another, unrelated theme.  For example, if a participant's description of an event starts at the beginning of a page and ends in the middle of the next page, that page and a half will be considered one unit of analysis.  The next unit of analysis will begin there and end as close to the middle of the following page as possible.  
   The minimum acceptable level of inter-rater reliability will be 75% (again, calculated using simple agreement), meaning that 75% of the two-coder judgements ("code absent" / "code present") must match. If a 75% level of agreement is not reached, the team will discuss discrepancies, revise the open coding scheme as needed, and use the revised scheme to code 3 more randomly selected interviews from among those not previously coded by the team.
   Afifi, W.A. (1999).  Harming the ones we love: Relational attachment and perceived consequences as predictors of safe-sex behavior.  The Journal of Sex Research, 36(2), 198-206.

   Agnew, C.R., & Loving, T.J. (1998).  Future time orientation and condom use attitudes, intentions and behavior.  Journal of Social Behavior and Personality, 13, 755-764.

   Ajzen, I., & Madden, T.J. (1986).  Prediction of goal-directed behavior: Attitudes, intentions, and perceived behavioral control.  Journal of Experimental Psychology, 22, 453-474.

   Allen, D.M., Lehman, J.S., Green, T.A., Lindgren, M.L., Onorato, I.M., & Forrester, W. (1994).  HIV infection among homeless adults and runaway youth, United States. 1989-1992.  AIDS, 8, 1593-1598.

   Altman, I., & Rogoff, B. (1987).  World views in psychology: trait, interactional, organismic, and transactional perspectives.  In D. Stokols & I. Altman (Eds.)., Handbook of environmental psychology.  NY: Wiley.

   Amaro, H. (1995).  Love, sex, and power.  American Psychologist, 50, 437-447.

   Anderson, J.E., Freese, T.E., & Pennbridge, J.N. (1994).  Sexual risk behavior and condom use among street youth in Hollywood.  Family Planning Perspectives, 26, 22-25.

   Athey, J.L. (1991).  HIV infection and homeless adolescents.  Child Welfare, 70, 517-528.

   Bandura, A. (1977).  Self-efficacy: Toward a unifying theory of behavior change.  Psychological Review, 84, 191-215.

   Bandura, A. (1978).  The self system in reciprocal determinism.  American Psychologist, 33, 344-358.

   Bandura, A. (1982).  Self-efficacy mechanism in human agency.  American Psychologist, 37, 122-147.

   Bandura, A. (1986).  Social foundations of thought and action: A social cognitive theory.  Englewood Cliffs, N.J.: Prentice Hall.

   Baranowski, T., Perry, C.L., & Parcel, G.S. (1997).  How individuals, environments, and health behavior interact: Social Cognitive Theory.  In K. Glanz, F.M. Lewis & B.K. Rimer (Eds.), Health Behavior and Health Education: Theory, Research, and Practice, 2nd Ed. (pp. 153-179).  San Francisco: Jossey-Bass.

   Barden, J.C. (1990, February 5).  Strife in families swells tide of homeless youths.  The New York Times, pp. A1, B8.
   Bass, D. (1992).  Helping vulnerable youths: Runaway and homeless adolescents in the United States.  Washington, DC: NASW Press.

   Becker, M.H. (1974).  The health belief model and personal health behavior.  Thorofare, NJ: Charles B. Slack.

   Bond, L.S., Mazin, R., & Jimenez, M.V. (1992).  Street youth and AIDS.  AIDS Education and Prevention, Supplement, 14-23.

   Booth, R.E., & Zhang, Y. (1996).  Severe aggression and related conduct problems among runaway and homeless adolescents.  Psychiatric Services, 47(1), 75-80.

   Bronfenbrenner, U. (1979).  The ecology of human development.  Cambridge, MA: Harvard University Press.

   Bronfenbrenner, U. (1993).  The Ecology of Human Development: Research Models and Fugitive Findings. In R.H. Wozniak & K.W. Fischer (Eds.) Nature, nurture & psychology (pp.313-324).  Washington, DC: American Psychological Association.

   Brown, L.K., DiClemente, R.J., & Reynolds, L.A. (1991).  HIV prevention for adolescents: The utility of the health belief model.  AIDS Education and Prevention, 3, 50-59.

   Bucy, J. (1987).  Prepared statement of June Bucy, Executive Director, the National Network of Runaway and Youth Services, Washington, DC.  In The crisis in homelessness: Effects on children and families.  Hearing before the select Committee on Children, Youth and Families, House of Representatives.  Washington, DC., U.S. Government Printing Office.

   Campbell, S.M., Peplau, L.A., & DeBro, S.C. (1992).  Men, women and condoms: The attitudes and experiences of heterosexual college students.  Psychology of Women Quarterly, 16, 273-288.

   Card, J.J., Niego, S., Malliari, A., & Farrell, W.S. (1996).  The Program Archive on Sexuality, Health & Adolescence: Promising “prevention programs in a box”.  Family Planning Perspectives, 28, 210-220.

   Carre, N., Prins, M., Meyer, H., Goldberg, D.J., Zangerle, R., Coutinho, R.A., & van den Hoek, A. (1997).  Has the rate of progression to AIDS changed in recent years?  AIDS, 11, 1611-1618.

   Carter, J.A., McNair, L.D., Corbin, W.R., & Williams, M. (1999).  Gender differences related to heterosexual condom use: the influence of negotiation styles.  Journal of Sex and Marital Therapy, 25(3), 217-225.

   Caton, C.L.M. (1986).  The homeless experience in adolescent years.  In E.L. Bassuk (Ed.), The mental health needs of homeless persons (pp 63-70). San Francisco: Josey Bass.
   CDC (Summer, 1994).  Prevention program recommendations -- Advisory committee presents final report.  HIV/AIDS Prevention, 5(1), 1.

   Centers for Disease Control and Prevention  (1996).  New CDC Data on Youth HIV/AIDS risk: Wide Range Prevention Messages Needed [On-line].  Available:

   Centers for Disease Control and Prevention (1997).  HIV and AIDS Trends.  Washington, DC: Government Printing Office.  

   Centers for Disease Control and Prevention (1998a, June).  CDC Works to Synthesize HIV/AIDS Prevention Intervention Research.  HIV/AIDS Prevention, 12.

   Centers for Disease Control and Prevention (1998b, March).  Focus on CDC research to support HIV prevention goals.  HIV/AIDS Prevention Newsletter, March 1998, 1.

   Centers for Disease Control and Prevention (1998c).  The HIV/AIDS Epidemic in the United States [On-line].  Available:

   Centers for Disease Control and Prevention (1998d).  Youth Risk Behavior Surveillance
-- United States, 1997.  MMWR, 47(SS-3).

   CDC (1999).  Trends in HIV-related sexual risk behaviors among high school students -- Selected U.S. Cities, 1991-1997.  MWWR, 48(21), 440-443.

   Chelimsky, E. (1982, May 5).  The problem of runaway and homeless youth.  In Oversight hearing on runaway and homeless youth program.  Washington, D.C:  U.S. House of Representatives, Subcommittee on Human Resources, Committee on Education and Labor.

   Choi, K., & Coates, T.J. (1994).  Prevention of HIV infection.  AIDS, 8, 1371-1389.

   Clanan, M. (1984).  The Health Belief Model and participation in programmes for the early detection of breast cancer: A comparative analysis.  Social Science and Medicine, 19, 823-830.

   Cline, R.J.W., & McKenzie, N.J. (1994).  Sex differences in communication and the construction of HIV/AIDS.  Journal of Applied Communication Research, 22, 322-337.

   Cochran, S.D., & Mays, V.M. (1989).  Women and AIDS related concerns: Roles for psychologists in helping the worried well.  American Psychologist, 44, 529-535.

   Cochran, S.D., & Mays, V.M. (1993).  Applying social psychological models to predicting HIV-related sexual risk behaviors among African Americans.  Journal of Black Psychology, 19, 142-154.

   Crumble, D. Wagstaff, D.A., Pinkerton, S.D., Sikkema, K.J., Kelly, J.A., & Cargill, V.A. (1998).  Substance use and HIV risk among low-income adolescents.  Int. Conf. AIDS, 12, 219-20.    

   Cwayna (1993).  Knowing where the fountains are.

   Damond, M.E., Bruer, N.L., & Pharr, A.E. (1993).  The evaluation of setting and a culturally specific HIV/AIDS curriculum: HIV/AIDS knowledge and behavioral intent of African American adolescents.  Journal of Black Psychology, 19, 169-189.

   D'Angelo, L.J., Getson, P.L., Luben, N., Stallings, E., & Gayle, H. (1989).  HIV infection in adolescents: Can we predict who is at risk? In Abstracts of the V International Conference on AIDS (p. 712).  Ottowa, Ontario, Canada: International Research Development Center.

   DiClememte, R.J. (1992).  Adolescents and AIDS.  Newbury Park: Sage.

   DiLorenzo, T., & Hein, K. (1995).  Adolescents: The leading edge of the next wave of the HIV epidemic.  In J.L. Wallander & L.J. Siegel (Eds.) Adolescent Health Problems: Behavioral Perspectives (pp. 117-140).  NY: Guilford.

   Dolcini, M., & Adler, N.E. (1994).  Perceived competencies, peer group affiliation, and risk behavior among early adolescents.  Health Psychology, 13(6), 496-506.

   Dolcini, M., Catania, J., Coates, T., Stall, R., Hudes, E.,  Gannon, J., & Pollack, L. (1993).  Demographic characteristics of heterosexuals with multiple partners: The national AIDS behavioral surveys (NABS).  Family Planning Perspectives, 25, 208-214.

   Domel, S.B. (1993).  Development and evaluation of a school intervention in increase fruit and vegetable consumption among 4th and 5th grade students.  Journal of Nutrition Education, 25, 345-349.

   Elkind, D. (1967).  Egocentrism in adolescence.  Child Development, 38, 1025-1034.

   Emmons, C., Joseph, J.G,. Kessler, R.C., Wortman, C.B., Montgomery, S.B., & Ostrow, D.G. (1986).  Psychosocial predictors of reported behavior change in homosexual men at risk for AIDS.  Health Education Quarterly, 13, 331-345.

   Family and Youth Services Bureau (1995).  Youth with Runaway, Throwaway, and Homeless Experiences...Prevalence, Drug Use, and Other At-Risk Behaviors.  FYSB Update, October 1995, 1-11.

   Fishbein, M., & Ajzen, I. (1975).  Belief, attitudes, intention and behavior: An introduction to theory and research.  Reading, MA: Addison-Wesley.

   Fishbein, M., & Guinan, M. (1996).  Behavioral science and public health: a necessary partnership for HIV prevention.  Public Health Reports, 111(S), 5-10.
   Fisher, J.D., & Fisher, W.A. (1992).  Changing AIDS-Risk Behavior.  Psychological Bulletin, 111, 455-474.

   Fisher, W.A., Fisher, J.D., & Rye, B.J. (1995).  Understanding and promoting AIDS-preventive behavior: Insights from the theory of reasoned action.  Health Psychology, 14(3), 255-264.

   Gerrard, M., Gibbons, F.X., Warner, T.D., & Smith, G.E. (1994).  Perceived vulnerability to HIV infection and AIDS preventive behavior: A critical review of the evidence.  In J.B. Pryor & G.B. Reeder, The Social Psychology of HIV Infection, Hillsdale, NJ: Lawrence Erlbaum Associates.  59-85.

   Glanz, K., Lewis, F.M., & Rimer, B.K. (1997).  Linking theory, research, and practice.  In K. Glanz, F.M. Lewis, & B.K. Rimer (Eds.), Health Behavior and Health Education, 2nd Ed..  San Francisco, CA: Jossey-Bass.  

   Glanz, K., & Oldenburg, B. (1996).  Relevance of health behavior research to health promotion and education.  In D.S. Gochman (Ed.) Handbook of health behavior research, Vol. 4.  New York: Plenum.

   Glaser, B. G.  (1992).  Emergence vs. forcing:  Basics of grounded theory analysis.  Mill Valley, CA:  Sociology Press.

   Greenblatt, M., & Robertson, M.J., (1993).  Life-styles, adaptive strategies, and sexual behaviors of homeless adolescents.  Hospital and Community Psychiatry, 44, 1177-1180.

   Greig, R., & Raphael, B. (1989).  AIDS prevention and adolescents.  Community health Studies, 13, 211-219.

   Gruber, E., & Chambers, C.V. (1987).  Cognitive development and adolescent contraception: Integrating theory and practice.  Adolesence, 22(87), 661-670.

   Goulart, M., & Madover, S. (1991).  An AIDS prevention program for homeless youth.  Journal of Adolescent Health, 12, 573-575.

   Holloway, W. (1984).  Gender difference and the production of subjectivity.  In J. Henriques, W. Holloway, C. Urwin, C. Venn, & V. Walkerdine (Eds.) Changing the subject; Psychology, social regulation and subjectivity (pp. 227-263).  London: Methuen.

   Harlow L.L., Quina, K., Morokoff, P.J., Grimley, D.M., & Rose, J.S. (1992).  Predicting HIV Risk in Women: A Multivariate Structural Model.  Paper presented at the annual meeting of the Society of Multivariate Behavioral Psychology, Cape Cod, M.A.

   Harlow, L.L., Rose, J.S., Morokoff, P.J., Quina, K., Mayer, K., Mitchell, K., & Schnoll, R. (1998).  Women HIV sexual risk takers: related behaviors, interpersonal issues, and attitudes.  Women's Health, 4(4), 407-39.
   Harper, G.W., & Carver, L.J. (1999).  “Out-of-the-Mainstream” youth as partners in collaborative research: exploring the benefits and challenges.  Health Education & Behavior, 26(2), 250-265.

   Helweg-Larsen, M., & Collins, B.E. (1994).  The UCLA multidimensional condom attitudes scale: Documenting the complex determinants of condom use in college students.  Health Psychology, 13, 224-237.

   Hendricks, J.C., Satten, G.A., van Ameijden, E.J., van Druten, H.A., Coutinho, R.A., & van Griensven, G.J. (1998).  The incubation period to AIDS in injecting drug users estimated from prevalent cohort data, accounting for death prior to an AIDS diagnosis.  AIDS, 12(12), 1537-44.

   Hudson, R., Petty, B., Freeman, A., Haley, C., & Krepcho, M. (1989).  Adolescent runaways' behavioral risk factors, knowledge about AIDS and attitudes about condom usage.  In Abstracts of the V International Conference on AIDS, (Abstract No. TDP 11). Ottowa, Ont.: International Research Development Center.

   Institute of Medicine (1994).  AIDS and Behavior.  Washington, DC: National Academy Press.

   Institute of Medicine (1996).  The hidden epidemic: Report in brief [On-line].

   Janz, N.K., & Becker, M.H. (1984).  The health belief model: A decade later.  Health Education Quarterly, 11, 1-47.

   Jemmott, L.S., & Jemmott, J.B. (1992).  Increasing condom-use intentions among sexually active black adolescent women.  Nursing Research, 41(5), 273-279.

   Jemmott, J.B., & Jemmott, L.S. (1993).  Alcohol and drug use during sexual activity: Predicting the HIV-risk-related behaviors of inner-city black male adolescents.  Journal of Adolescent Research, 8, 41-57.

   Jemmott, J.B., Jemmott, L.S., & Fong, G.T. (1998).  Abstinence and safer sex HIV risk-reduction interventions for African-American adolescents: A randomized controlled trial. JAMA, 279, 1529-1536.

   Jemmott, J.B., Jemmott, L.S., Spears, H., Hewitt, N., & Cruz-Collins, M. (1992).  Self-efficacy, hedonistic expectancies, and condom-use intentions among inner-city Black adolescent women:  A social cognitive approach to AIDS risk behavior.  Journal of Adolescent Health, 13, 512-19.

   Jessor, R., & Jessor, S.L. (1977).  Problem behavior and psychological development: A longitudinal study of youth.  New York: Academic Press.

   Kashima, Y., Gallois, C., & McCamish, M. (1993).  The theory of reasoned action and cooperative behavior: It takes two to use a condom.  British Journal of Social Psychology, 32, 227-239.

   Kelly (1979). Ecological perspective.

   Kelly, J.A. (1995).  Changing HIV risk behavior: Practical strategies.  New York: Guilford Press.

   Kelly, J.A., Murphy, D.A., Sikkema, K.J., & Kalichman, S.C. (1993).  Psychological interventions to prevent HIV infection are urgently needed: New Priorities for behavioral research in the second decade of AIDS . American Psychologist, 48, 1023-1034.

   Kennedy, M.R. (1991).  Homeless and runaway youth mental health issues: No access to the system.  Journal of Adolescent Health, 12, 576-579.

   Kim, M.A., Stanton, B., Li, X., Dickersin, K., & Galbraith, J. (1997).  Effectiveness of 40 adolescent AIDS-risk reduction interventions: A quantitative review.  Journal of Adolescent Health, 20, 204-215.

   Kipke, M.D., Boyer, C., & Hein, K. (1993).  An evaluation of an AIDS risk reduction education and skills training (ARREST) program.  Journal of Adolescent Health, 14, 533-539.

   Kipke, M.D., Montgomery, S.B., Simon, T.R., Unger, J.B., & Johnson, C.J. (1997).  Homeless youth: Drug use patterns and HIV risk profiles according to peer group affiliation.  AIDS and Behavior, 1 (4), 247-259.

   Kipke, M.D., O'Connor, S., Palmer, R., & MacKenzie, R.G. (1995).  Street youth in Los Angeles: Profile of a group at high risk for human immunodeficiency virus infection.   Archives of Pediatric & Adolescent Medicine, 149, 513-519.

   Kipke, M.D., Simon, T.R., Montgomery, S.B, Unger, J.B., & Iversen, E.F. (1997).  Homeless youth and their exposure to and involvement in violence while living on the streets.  Journal of Adolescent Health, 20, 360-367.

   Kirby & Coyle (1994) reviewed 11 published evaluations of school-based HIV prevention programs and found that . . .

   Koopman, C., Rosario, M., & Rotheram-Borus, M.J. (1994).  Alcohol and drug use and sexual behaviors placing runaways at risk for HIV infection.  Addictive Behaviors, 19 (1), 95-103.

   Lawrence, D., Jason, J., Holman, R., & Murphy, J. (1991).  HIV transmission from hemophilic men to their heterosexual wives.  In N. Alexander, H. Gabelnick, G. Hodgen, & R. Spieler (Eds.), The heterosexual transmission of AIDS: Proceedings of the CONRAD 2nd International Workshop.  New York: Alan R. Liss.
   Lincoln, Y.S., & Guba, E.G. (1985).  Naturalistic inquiry.  Beverly Hills, CA: Sage.

   Link, B.G., & Phelan, J. (1995).  Social Conditions as Fundamental Causes of Disease.  Journal of Health and Social Behavior, Extra Issue, 80-94.

   Linville, P.W., & Fischoff, B. (1993).  AIDS risk perceptions and decision biases.  In J.B. Pryor & G.D. Reeder (Eds.), The social psychology of HIV infection.  Hillsdale, NJ: Lawrence Erlbaum Associates.  

   Luna, C.C. (1991).  Street youth: Adaptation and survival in the AIDS decade.  Journal of Adolescent Health, 12, 511-514.

   Luna, G.C., & Rotheram-Borus, M.J. (1992).  Street youth and the AIDS pandemic.  AIDS Education and Prevention (Suppl. 1), 1-13.

   Lundy, K. C. (1995).  Sidewalks talk: A naturalistic study of street kids. Garland: New York.  

   Lux, K.M., & Petosa, R. (1998).  Preventing HIV infection among juvenile delinquents: educational diagnosis using the health belief model.  In D. Buchanan & G. Cernada (Eds.),  Progress in preventing AIDS?: Dogma, dissent, and innovation (pp. 55-73).  New York: Baywood.  

   Madden, T.J., Ellen, P.S., & Ajzen, I. (1992).  A comparison of the Theory of Planned Behavior and the Theory of Reasoned Action.  Personality and Social Psychology Bulletin, 18 (1), 3-9.

   Maticka-Tyndale, E. (1995).  Can we? Have we? Prevention of sexual transmission of HIV.  The Canadian Journal of Human Sexuality, 4 (2), 79-102.

   Mays, V.M., & Cochran, S.D. (1988).  Issues in the perception of AIDS risk and risk reduction activities by Black and Hispanic/Latina women.  American Psychologist, 43, 949-957.

   McCarthy, B., & Hagan, J. (1992).  Surviving on the street: The experiences of homeless youth.  Journal of Adolescent Research, 7 (4), 412-430.

   McCusker, J., Westenhouse, J., Stoddard, A.M., Zapka, J.G., Zorn, M.W., & Mayer, K.H. (1990).  Use of drugs and alcohol by homosexually active men in relation to sexual practices.  Journal of Acquired Immune Deficiency Syndrome, 3, 729-737.

   Mechanic, D. (1991).  Adolescents at risk: New directions.  Journal of Adolescent Health, 12, 638-643.

   Middlestadt, S.E., Bhattacharyya, K., Rosenbaum, J., Fishbein, M., & Shepherd, M. (1996).  The use of theory based semistructured elicitation questionnaires: Formative research for CDC's prevention marketing initiative.  Public Health Reports, 111 (Suppl. 1), 18-27.

   Mischel, W. (1973).  Toward a social cognitive learning reconceptualization of personality.  Psychological Review, 80, 252-283.

   Miller, H.G., Turner, C.F., & Moses, L.E. (1990).  AIDS: The second decade.  Washington, DC: National Academy Press.

   Montgomery, S.B., Joseph, J.G., Becker, M.H., Ostrow, D.G., Kessler, R.C., & Kirscht, J.P. (1989).  The health belief model in understanding compliance with preventive recommendations for AIDS: How useful?  AIDS Education and Prevention, 1, 303-323.

   Moore, S.M., & Rosenthal, D.A. (1991).  Adolescent invulnerability and perception of AIDS risk.  Journal of Adolescent Research, 6, 164-180.

   Moore, S.M., & Rosenthal, D.A. (1998).  Contemporary youths' negotiations of romance, love, sex, and sexual disease.  In V.C. deMunck (Ed.), Romantic love and sexual behavior: perspectives from the social sciences (pp. 203-232).  Westport, CT: Praeger.  

   Muir, M. (1991).  The environmental context of AIDS.  Praeger: New York.

   Muus, R.E., Velder, E., & Porton, H. (1996). Urie Bronfenbrenner's ecological perspective of human development.  In R.E. Muus, E. Velder, & H. Porton (Eds.), Theories of adolesence (pp. 312-338).  New York:McGraw-Hill.

   National Coalition for the Homeless (1999).  NCH Fact Sheet #11: Homeless youth [On-line].  Available: internet:

   National Institutes of Mental Health Multisite Prevention Trial Group (1997).  Demographic and behavioral predictors of sexual risk in a multisite HIV prevention trial.  AIDS, 11 (S2), S21-27.

   National Network of Runaway and Youth Services (1991).  To whom do they belong?  Runaway, homeless and other youth in high-risk situations in the 90's.  Washington, DC: Author.

   Neumann, M., Johnson, W.D., Semaan, S.V., Sogolow, E.D., Ramirez, G., Strouse, D., & Hedges, L. (1998).  Synthesis of HIV prevention intervention research for heterosexual populations.  12th World AIDS Conference, Abstract #23157.

   New York City Health Department (1990, January).  AIDS surveillance report.  New York: New York City Health Department, AIDS Surveillance Unit.

   Parker, R.G. (1996).  Empowerment, community mobilization and social change in the face of HIV/AIDS.  AIDS, 10 (Suppl. 3), S27-S31.

   Peterson, J., Catania, J., Dolcini, M., & Faigeles, B. (1993).  Multiple sexual partners among Blacks in high-risk cities.  Family Planning Perspectives, 25, 263-267.

   Peterson, J.L., & DiClemente, R.J. (1994).  Lessons learned from behavioral interventions.  In R.J. DiClemente & J.L. Peterson (Eds.), Preventing AIDS: Theories and methods of behavioral interventions (pp. 319-322).  New York: Plenum Press.

   Phillips, K. (1991).  The primary prevention of AIDS.  In M. Pitts & K. Phillips (Eds.) The psychology of health (pp. 139-155).  New York: Routledge.  

   Pinkerton, S.D., Holtgrave, D.R., Leviton, L.C., Wagstaff, D.A., Cecil, H., & Abramson, P.R. (1998).  Toward a standard sexual behavior data set for HIV prevention evaluation.  American Journal of Health Behavior, 22 (4), 259-266.

   Podschun, G.D. (1993).  Teen Peer outreach-Street Work Project: HIV prevention education for runaway and homeless youth.  Public Health Reports, 108, 150-155.

   Prochaska, J.O., Redding, C.A., Harlow, L.L., Rossi, J.S., & Velicer, W.F. (1994).  The transactional model of Change and HIV Prevention: A review, Health Education Quarterly, 21 (4), 471-486.

   Quinn, T.C., Glasser, D., Cannon, R.O., Matuszak, D.L., Dunning, R.W., Kline, R.L., Campbell, C.H., Israel, E., Fauci, A.S., & Hook, E.W. (1988).  Human immunodeficiency virus infection among patients attending clinics for sexually transmitted diseases.  New England Journal of Medicine, 318, 197-203.

   Reason, P. (1994).  Inquiry and alienation.  In P. Reason (Ed.), Participation in human inquiry (pp. 9-15).  Thousand Oaks, CA: Sage.

   Rennie, D. L., Phillips, J. R., & Quartaro, G. K.  (1988).  Grounded theory:  A promising approach to conceptualization in psychology?  Canadian Psychology, 29, 139-150.

   Rickert, V.I., Gottlieb, A., & Jay, M.S. (1990).  A comparison of three clinic-based AIDS education programs on female adolescents' knowledge, attitudes, and behavior.  Journal of Adolescent Health Care, 11, 298-303.

   Rogoff, B. (1993).  Children's guided participation and participatory appropriation in sociocultural activity.  In R. Wozniak & K. Fischer (Eds. ), Development in context: Acting and thinking in specific environments.  Hillsdale, NJ: Erlbaum.

   Rogoff, B. (1997).  Cognition as a collaborative process.  In D. Kuhn & R.S. Siegler (Eds.), Handbook of Child Psychology (5th ed.).  New York: Wiley.

   Robertson, J.M. (1988).  Homeless adolescents: a hidden crisis.  Hospital Community Psychiatry, 39, 475.

   Rogers, R.W. (1975).  A protection motivaton theory of fear appeals and attitude change.  Journal of Psychology, 91, 93-114.

   Rosenberg, P.S., & Biggar, R.J. (1998).  Trends in HIV incidence among young adults in the United States.  JAMA, 279, 1894-1899.

   Rosenstock, I.M. (1974) . Historical origins of the health belief model.  Health Education Monographs, 2, 328-335.

   Rotheram-Borus, M.J. (1991).  Serving runaway and homeless youths.  Fam. Community Health, 14 (3), 23-32.

   Rotheram-Borus, M.J., (1993).  Suicidal behavior and risk factors among runaway youths.  American Journal of Psychiatry, 150 (1), 103-107.

   Rotheram-Borus, M.J., Feldman, J., Rosario, M., & Dunne, E. (1994).  Preventing HIV among runaways: Victims and victimization.  In R.J. DiClemente & J.L. Peterson (Eds.), Preventing AIDS: Theories and methods of behavior interventions.  New York, NY: Plenum Publishing Corporation.  

   Rotheram-Borus, M.J., & Koopman, C. (1989).  AIDS and adolescents.  In R. Lerner, A. Peterson, J., & Brooks-Gunn, J. (Eds.), The encyclopedia of adolescence, New York: Garland Press.

   Rotheram-Borus, M.J., & C. Koopman (1991).  Sexual risk behaviors, AIDS knowledge and beliefs about AIDS among runaways.  American Journal of Public Health, 81, 208-210.
   Rotheram-Borus, Koopman & Bradley (1989).  [sexual abuse is associated with having a higher number of sexual partners].

   Rotheram-Borus, M.J., Koopman, C., & Ehrhardt, A.A. (1991).  Homeless youths and HIV infection.  American Psychologist, 46, 1188-1197.

   Rotheram-Borus, Koopman, Haignere,  & Davies (1991).  Reducing HIV sexual risk behaviors among runaway adolescents.  JAMA, 266, 1237-1241.

   Rotheram-Borus, M.J., Parra, M., Cantwell, C., Gwadz, M., & Murphy, D.A. (1996).  Runaway and homeless youths.  In R.J. Di Clemente, W.B. Hansen and L.E. Ponton (Eds.), Handbook of Adolescent Health Risk Behavior.  NewYork: Plenum Press.

   Rothman, J., & David, T. (1985).  Status offenders in Los Angeles County: Focus on runaway and homeless youth (A study and policy recommendation).  Los Angeles: University of California, School of Social Welfare.

   Russell, D.E.H. (1986).  The secret trauma: Incest in the lives of girls and women.  New York:Basic Books.

   Sacco, W.P., Levine, B., Reed, D.L., & Thompson, K. (1991).  Attitudes about condom use as an AIDS-relevant behavior: Their factor structure and relation to condom use.  Psychological Assessment, 2, 265-272.

   St. Lawrence, J.S., Brasfield, T.L., Jefferson, K. W., Alleyne, E., O'Bannon III, R.E., & Shirley, A. (1995).  Cognitive-behavioral intervention to reduce African American adolescents' risk for HIV infection.  Journal of Consulting and Clinical Psychology, 63, 221-237.

   St. Louis, M.E., Hayman, C.R., Miller, C., Anderson, J.E., Petersen, L.R., & Dondero, T.J. (1989).  HIV infection in disadvantaged adolescents in the U.S.: Findings from the Job Corps Screening Program.  Abstracts of the V International Conference on AIDS (p. 711).  Ottowa, Ontario, Canada: International Research Development Center.

   San Francisco Task Force on Prostitution (1996).  Final report: 1996 [On-line].  Available: 10youth.html.

   Satcher, D. (1996).  The importance of behavioral science in HIV prevention.  Public Health Reports, 111 (Suppl. 1), 1-2.

   Schneir, A., Kipke, M.D., Melchior, L.A., & Huba, G.J. (1998).  Childrens Hospital Los Angeles:  A model of integrated care for HIV-positive and very high-risk youth.  Journal of Adolescent Health, 23S, 59-70.

   Santrock (1998).  Adolescence (7th ed.).  NewYork, NY:McGraw-Hill.

   Seage, G.R., Mayer, K.H., Lenderking, W.R., Wold, C., Gross, M., Goldstein, R., Cai, B., Heeren, T., Hingson, R., & Holdberg, S. (1997).  HIV and hepatitis B infection and risk behavior in young gay and bisexual men.  Public Health Reports, 112, 158-167.

   Service, S.K., & Blower, S.M. (1995).  HIV transmission in sexual networks: an empirical analysis.  Proceeding of the Royal Scoiety of London, Biological Sciences, 250, 237-244.

   Shaffer, D., & Caton, D. (1984).  Runaway and homeless youth in New York City:  A report to the Ittleson Foundation.  New York: The Ittleson Foundation.  

   Sheeran, P., Abraham, C., & Orbell, S. (1999).  Psychosocial correlates of heterosexual condom use: a meta-analysis.  Psychological Bulletin, 125 (1), 90-132.

   Sherman, B. (1995).  Interview with Sandy Cooke, District Supervisor, Vancouver Adolescent Street Unit, Ministry of Social Services.  British Columbian Institute Against Family Violence Newsletter [On-line].  Available:

   Shriver, M., DeBurger, R., Brown, Christopher, Simpson, H.L., & Myerson, B. (1998).  Bridging the gap between science and practice: Insight to researchers from practitioners.  Public Health Reports, 113 (S1), 189-193.

   Slap, G.B., Plotkin, S.L., & Khalid, N. (1991).  Human immunodeficiency virus peer education program for adolesdent females.  Journal of Adolescent Health, 12, 434-442.

   Sobo, E.J. (1993).  Inner city women and AIDS: The psycho-social benefits of unsafe sex.  Culture, Medicine and Psychiatry, 17, 455-485.

   Sobo, E.J. (1995).  Choosing unsafe sex: AIDS risk and denial among disadvantaged women.  Philadelphia:University of Pennsylvania Press.    

   Soet, J.E., Dudley, W.N., & Dilorio, C. (1999).  The effects of ethnicity and perceived power on women's sexual behavior.  Psychology of Women Quarterly, 23, 707-723.

   Sogolow, E.D., Semaan, S., Johnson, W.D., Neumann, M., Ramirez, Sweat, M., & Doll, L.S. (1998).  Effects of US-based interventions on safer sex: Meta-analyses, overall and for populations, age groups, and settings.  12th World AIDS Conference, Abstract #14283.

   Stanton, B., Kim, N., Galbraith, J., & Parrott, M. (1996).  Design issues addressed in published evaluations of adoelscent HIV-risk reduction interventions: A review.  Journal of Adolescent Health, 18, 387-397.

   Strauss, A.L., & Corbin, J. (1990).  Basics of qualitative research: Grounded theory procedures and techniques.  Newbury Park, CA: Sage.

   Strauss, A.L., & Corbin, J. (1994).  Grounded theory methodology: An overvivew.  In N.K. Denizen & Y.S. Lincoln (Eds.), Handbook of qualitative research (pp. 273-295).

   Strecher, V.J., & Rosenstock, I.M. (1997).  The health belief model.  In A. Baum, S. Newman, J. Weinman, R. West, & C. McManus (Eds.), Cambridge Handbook of Psychology, Health and Medicine (pp. 113-116).  New York: Cambridge University Press.

   Sugerman, S.T., Hergenroeder, A.C., Chacko, M.R., & Parcel, G.S. (1991).  Acquired immunodificiency syndrome and adolescents: knowledge, attitudes and behaviors of runaway and homeless youths.  American Journal of Disesases in Children, 145, 431-436.

   Sweeney, P., Lindegren, M.L., Buehler, J.W., Onorata, I.M., & Janssen, R.S. (1995).  Teenagers at risk of human immunodeficiency virus type 1 infection.  Archives of Pediatrics and Adolescent Medicine, 149, 521-528.

   Taplin, S.H., & Montaño, D.E. (1993).  Attitudes, age and participation in mammographic screening: A prospective analysis.  Journal of the American Board of Family Practice, 6, 3-23.

   Tawil, O., Verster, A., & O'Reilly, K.R. (1995).  Enabling approaches for HIV/AIDS prevention:  Can we modify the environment and minimize the risk?  AIDS, 9, 1299-1306.

   Tenner, A.D., Trevithick, L.A., Wagner, V., & Burch, R. (1998).  Seattle YouthCare's prevention, intervention, and education program: A model of care for HIV-positive, homeless, and at-risk youth.  Journal of Adolescent Health, 23 (Suppl. 2), 96-106.

   U.S. General Accounting Office (1989).  Homelessness: Homeless and runaway youth receiving services at federally funded shelters.  Washington, DC:Author.

   van Campenhoudt, L., Cohen, M., Guizzardi, G., & Hausser, D. (1997).  Sexual interaction and HIV risk.  Bristol, PA: Taylor & Francis.

   Walter, H.J., & Vaughan, R.D. (1993).  AIDS risk reduction among a multi-ethnic sample of urban high school students.  JAMA, 270, 725-730.

   Watts, R.J. (1999).  Unpublished lecture in Qualitative Research Methods.

   Williams, C. (1995).  A home-based prevention program for sixth grade alcohol use:  results from Project Northland.  Journal of Primary Prevention, 16, 125-147.

   Woods, E.R., Samples, C.L., Melchiono, M.W., Keenan, P.M., Fox, D.J., Chase, L.H., Tierney, S., Price, V.A., Paradise, J.E., O'Brien, R.F., Mansfield, C.J., Brooke, R.A., Allen, D., & Goodman, E. (1998).  Boston HAPPENS Program: A model of health care for HIV-positive, homeless, and at-risk youth.  Journal of Adolescent Health, 23 (Suppl. 2), 37-48.

   Worth, D. (1989).  Sexual decision making and AIDS: Why condom promotion among vulnerable women is likely to fail.  Studies in Family Planning, 20, 297-307.

   Worth, D., & Rodriguez, R. (1987).  Latina women and AIDS.  SIECUS Report XV, 3,

   Wright, D. (1997).  Acknowledging the continuum from childhood abuse to male prostitution..  British Columbian Institute Against Family Violence Newsletter [On-line].  Available:

   Wyatt, G.E., Newcomb, M.D., & Reiderle, M.H. (1993).  Sexual abuse and consensual sex: Women's developmental patterns and outcomes.  Newbury Park: Sage.

   Yates, G., Mackenzie, R., Pennbridge, J., & Cohen, E. (1988).  A risk profile comparison of runaway and non-runaway youth.  American Journal of Public Health, 78, 820-821.

   Yoder, J.D., & Kahn, A.S. (1992).  Toward a feminist understanding of women and power.  Psychology of Women Quarterly, 16, 381-388.

   Appendix A.   Bronfenbrenner's Four Ecological Levels     
    Appendix B.   The Health Belief Model     
    Appendix C.  Interview Guide     
    Appendix D.  Consent Form