Minor and Adult Domestic Sex Trafficking Risk Factors in Ohio
Abstract
Objective: Sex trafficking is a pressing issue in Ohio—a major node in national sex trafficking networks; however, no quantitative analysis of risk factors specific to Ohio victims has been conducted. Method: Using a survey conducted by the Ohio Human Trafficking Commission representing Ohio urban, street-based sex workers (N = 328), this study uses a life-course theory framework to identify and measure the effect of direct risk factors on domestic minor and adult sex-trafficking outcomes using multivariate logistic regressions. Variables of specific interest include peer influence, experiences during transience, prior minor victimization, and a lack of support while attempting to escape. Results: Survival sex and peer influence increase the odds of minor victimization, and a lack of available support and resources for victims while attempting to escape significantly increase the odds of adult victimization. Conclusion: This analysis produces a better understanding of risk factors facing Ohio’s sex trafficking victims and provides broader insights about the determinants of domestic sex trafficking. In addition, these results have been used in a report to an Ohio Senate Caucus promoting a federal bill regarding missing children.
Human trafficking is the world’s second largest criminal enterprise, claiming approximately 12.3 million victims worldwide, an extrapolative estimate of cases reported by the International Labour Office (Belser, 2005). Women are estimated to comprise between 50% and 80% of victims worldwide, of whom nearly 70% are trafficked into the sex trade (U.S. Department of State, 2007; U.S. Department of Justice, 2004). UNICEF reported more than 2 million children were trafficked globally (U.S. Department of State, 2010). The nongovernmental organization (NGO) End Child Prostitution, Child Pornography and Trafficking of Children for Sexual Purposes (ECPAT) approximated 300,000 to 400,000 U.S. youth (individuals younger than 18 years) are trafficked in the United States annually (Willis & Levy, 2002; Spangenberg, 2001). The enumeration measures and methodologies used to approximate these hidden and underground populations, if elucidated, are by no means definitive and are thus subject for debate; therefore, these estimates should be considered with caution.
This article focuses on domestic sex trafficking in the United States, which is defined by the U.S. Victims of Trafficking and Violence Protection Act of 2000 as an inducement into committing commercial sex acts by force, fraud, or coercion, or when persons induced to perform sex acts are under 18 years old. The term commercial sex act refers to sexual acts exchanged for valued resources, encompassing a range of sexual acts such as prostitution, exotic dancing, pornography, sexual entertainment, sexual servitude, and servile marriage (Logan, Walker, & Hunt, 2009). Under the federal law, those who profit from commercial sex acts performed by persons under their control by means of force, manipulation, or coercion, traditionally known as pimps, are now referred to as traffickers (Williamson, Karandikar, et al., 2010). In this study, respondents were identified as trafficking victims if they affirmatively responded to the following: (a) currently or previously self-identified as a trafficking victim, (b) was forced into the sex industry by a trafficker or agent acting in their place, such as a boyfriend or family member, or (c) sold sex at age 18 years or younger, but did not self-identify as a trafficking victim.
This study analyzed direct sex trafficking risk factors in Ohio within a life-course theory framework. Life-course theory was used to identify and systematize direct minor and adult sex trafficking risk factors, whose impact on outcomes were measured using multivariate logistic regressions on a representative sample of Ohio urban, street-based sex workers. Results, policy implications, and future research suggestions are discussed.
Background
Ohio hosts an extensive commercial sex market, including domestic and international trafficking victims, due in large part to its strategic location as an interstate transportation node. The intragovernmental anti-trafficking initiative, Operation Cross Country IV, found all Ohio’s major cities host at least one Asian massage parlor, which frequently use illegal immigration statuses or large smuggling debts as leverage to force women to serve as sex workers. Further, Ohio contains concealed sex markets operating at migrant camps, pornography studios, and truck stops. The state also has the fifth highest number of strip clubs in the country (Krauss, 2014; National Human Trafficking Resource Center, n.d. Williamson, Karandikar, et al., 2010). Ohio is not a destination state for international trafficking victims; rather, the state serves as a source for recruits and profitable market for trafficked minors in large part due to its position along traffickers’ circuits between Chicago and Detroit (Williamson, Karandikar, et al., 2010). Ohio’s cities are important nodes in domestic sex trafficking supply and transport networks because of the state’s accessibility to interstate highways and water transit routes (i.e., Lake Erie and the Ohio River), proximity to Canada, sizable immigrant population, and clusters of impoverished people (Williamson, Karandikar, et al., 2010). Toledo is Ohio’s fourth most populous city and a notorious source of sex trafficking victims. In fact, since the passage of the Victims of Trafficking and Violence Protection Act of 2000, at least one victim in every national sting was from Toledo (Perdue, Williamson, Billings, Schart, & Boston-Gromer, 2011).
The Polaris Project (n.d.), a national anti-trafficking NGO, ranked Ohio ninth on a list of all U.S. states regarding the inability to take sufficient action to combat human trafficking through programming, legislation, or policy (Rocha, 2012). The state responded in 2010 by passing Senate Bill 235, which rendered human trafficking a standalone felony offense and established provisions for educating law enforcement on human trafficking issues, such as requiring data collection on convictions and providing protection for victims (Ohio Senate Bill 235, 2010; Perdue et al., 2011). In June 2012, the Ohio Legislature enacted safe harbor laws as part of House Bill 262, meaning minors arrested for selling sex were no longer treated and prosecuted as criminals, but considered victims of child abuse (Ohio House Bill 262, 2012).
Despite the passing of these anti-trafficking laws, as well as the committed work of activists, domestic sex trafficking persists in Ohio. The Ohio Attorney General’s Trafficking in Persons Study Commission’s Research and Analysis Sub-Committee (Williamson, Karandikar, et al., 2010) estimated approximately 1,000 female and disproportionately minority Ohio girls (aged 12 to 17 years ) are trafficked into the sex trade each year. Further, an additional 3,016 Ohio girls are at risk of trafficking; this risk is especially high for runaways who have been missing from home for more than 2 weeks (Williamson, Karandikar, et al., 2010). Approximations of at-risk female and male populations were calculated by modifying Estes and Weiner’s (2001) population extrapolation method, which multiplies the general risk population (e.g., girls, runways, homeless youth) by an empirical rate based on actual incident counts (see Williamson, Karandikar, et al., 2010, pp. 38–43 for full discussion of this method). The number of trafficked girls were estimated using rates recorded by Northwest Ohio’s Innocence Lost Task Force. Estimates for Ohio women trafficking victims remain elusive (Williamson, Karandikar, et al., 2010).
Theoretical Foundation and Hypotheses
This study sought to measure the effect of direct risk factors for minor and adult domestic sex trafficking on the odds of trafficking outcomes. This investigation was motivated by Reid’s (2010, 2011) structural modeling approach, which differentiated between direct and indirect risk factors. Reid separated these two types of risk factors by developing a structural equation causal pathway model of sex trafficking victimization, beginning with caregiver strain (e.g., parental mental health issues, domestic violence), which led to risk-inflating factors (e.g., running away from home, early onset of drug use). Moreover, negative emotions arising from past trauma directly increased vulnerability to victimization (Reid, 2010). Specifically, Reid (2010) identified a statistically significant casual pathway from children running away from home to minor sex trafficking victimization, beginning with childhood maltreatment, neglect, and sexual abuse. This model was constructed within the framework of Sampson and Laub’s (1993) age-graded developmental theory of informal social control—an amalgamation of life course and social control theory—recommended for application to sex trafficking research by Reid (2012a, 2012b).
The basic life course structure considers four directing principles influencing human development: (a) human agency, (b) time and place, (c) event timing, and (d) linked lives (Elder, Johnson, & Crosnoe, 2003). Human agency refers to an individual’s decisions regarding behavior within the context of personal opportunities and historical, social, and geographic contexts (Reid, 2012a). The time and place of life experiences entail the effect of historical events (e.g., the Great Depression) on a population cohort, and the geographic context involves the life opportunities presented by the individual’s location (Reid, 2012b). The timing of life events is termed transitions; the times between transitions are known as durations (Elder et al., 2003). The timing of life events is a particularly important component of life-course theory because immediate situational adaptations potentially place individuals on life-long trajectories. Linked lives refer to the social embeddedness of individuals within others’ lives, such as the existence social ties (Elder et al., 2003).
Sampson and Laub’s (1993, 2005) aged-graded informal social control theory posits dynamic change is motivated by individual responses to social events, decisions, and outcomes; that is, social events are major catalysts determining different life outcomes among individuals (Reid, 2012b). Structural factors are also an important component of the theory; for example, financial hardship, mobility, family disruption, and interactions with family members and peers—mitigated by an individual’s temperament or constitution (Reid, 2012a). Antisocial behaviors (e.g., delinquency) persisting from childhood into adulthood also influence the process of how individuals make decisions and relate to their social environment (Sampson & Laub, 1993). Social capital (or the lack thereof) motivates or dissuades antisocial behavior. For example, stable employment diminishes the attractiveness of theft or financially-motivated crime, whereas chronic unemployment amplifies the appeal of informal and underground occupations.
Risk Factors for Sex Trafficking
The theoretical framework suggests minor’s trafficking risk factors are largely structural; however, other commonly-identified nonstructural risk factors are summarized in Table 1. The main structural factor affecting minor trafficking outcomes is negative childhood and adolescent informal social control processes, which are managed by personal traits. Negative childhood and adolescent informal social control processes include poor school performance, family dysfunction, running away or being forced from home, homelessness, engaging in survival sex, having family or friends in the sex trade, juvenile detention, childhood abuse, early drug or alcohol use, and placement in foster care (Reid, 2012a).
| Minor Victim Risk Factors | n (%) | Adult Victim Risk Factors | n (%) | Shared Risk Factors | n (%) |
|---|---|---|---|---|---|
| Spent time in juvenile detention before involvement in sex work | 44 (13%) | No one would help or believe you when you tried to get help out of sex work | 61 (19%) | Homeless before involvement in sex work | 117 (36%) |
| Engaged in survival sex while running away | 82 (25%) | Did not seek help out of sex work | 79 (24%) | Substance abuse disorder after involvement in sex work | 59 (18%) |
| Friends involved in the sex trade (bought sex or sold others for sex) | 66 (20%) | Adults respondents who were minor trafficking victims | 59 (18%) | Frequently used drugs and alcohol before involvement in sex work | 158 (48%) |
| Ran away from home before involvement in sex work | 138 (42%) | Diagnosed with a mental health disorder after involvement in sex work | 27 (8%) | ||
| Diagnosed with depression before involvement in sex work | 85 (26%) | Responsible for dependent child | 103 (31%) | ||
| Difficulty in school before involvement in sex work | 105 (32%) | Dropped out of school | 126 (38%) | ||
| Conflict with parents before involvement in sex work | 101 (31%) | ||||
| Familial poverty before involvement in sex work | 110 (33%) | ||||
| Placement in foster care | 46 (14%) |
Adult risk factors are structurally motivated by a lack of social capital, independent from negative informal social controls (Reid, 2012a). Domestic adults often accept victimization as a means to support dependents (particularly children from teenage pregnancies) or as the result of coercion or abandonment by abusive romantic partners. Further, adult trafficking risk is often elevated because prospects for formal employment and accumulation of social capital are hindered by low educational attainment, few job skills, drug addictions, and mental health disorders (Holger-Ambrose, Langmade, Edinburgh, & Saewyc, 2013). Adult sex trafficking victims also victimized or homeless as minors are at a greater disadvantage due to the aged-graded social control principle of contiguity, which states continuous disruptions in education, family ties, and friendships hinders the development of positive adult social bonds (Sampson & Laub, 1993).
The analysis involved testing several hypotheses based on the theoretical framework and two less-directed suppositions, or subhypotheses, created to test commonly-held logics regarding the role of race and age on sex trafficking outcomes.
• Hypothesis 1: If a minor engages in survival sex, then he or she is at greater odds to be a minor trafficking victim, even when accounting forrunning away from home and homelessness.
• Hypothesis 2: If an individual has friends involved in buying sex or trading others for sex prior to personal involvement in the sex trade, then he or she is at increased odds of becoming a minor sex trafficking victim.
• Hypothesis 3: If a sex worker attempts to quit sex work, but is rejected or ignored, then he or she has greater odds of being an adult sex trafficking victim than those who found help.
• Hypothesis 4: If an adult sex worker was also victimized as a minor, then he or she has greater odds of currently being a sex trafficking victim.
▪ Subhypothesis 1: The odds of a sex worker becoming a minor or adult trafficking victim differ according to the individual’s race/ethnicity.
▪ Subhypothesis 2: The odds of being a sex trafficking victim decrease with age.
Method
This study relies on two multivariate logistic regression models to test the hypotheses while accounting for other theoretical minor- and adult victim-specific risk factors using a representative sample of urban, street-based sex workers in Ohio. A list of risk factors available for the minor and adult models is presented in Table 1.
Data and Sample
The analysis in this study used data obtained from a U.S. Department of Justice-funded survey administered by the Domestic Sex Trafficking in Ohio Research and Analysis Sub-Committee of the Ohio Human Trafficking Commission in 2011. The dataset is unique for its large size (N = 328), thoroughness (more than 100 questions), and representativeness of urban street-based Ohio sex workers (Williamson, Perdue, Belton, & Burns, 2012). The majority of the questionnaire was formatted as binary responses (i.e., yes/no; have/have not), covering demographics, sex work history, abuse, runaway episodes, transportation routes taken by traffickers, among other topics. A full descriptive analysis of the survey results can be found in Williamson, Karandikar, et al. (2010). Data collectors worked with trained advocates at collection sites in private and secure library study rooms. Contributors younger than 18 years were reported to child welfare. Questions were reworded and repeated during the survey to assess consistency and reliability; those with vast discrepancies were removed. A demographic overview of the survey respondents can be found in Table 2.
| Demographic | All Respondents | Minor Victims | Adult Victims |
|---|---|---|---|
| Total | 328 | 61 (18.6%) | 54 (16.5%) |
| Female (%) | 232 (70.7%) | 43 (70.5%) | 49 (90.7%) |
| Male (%) | 96 (29.3%) | 18 (29.5) | 5 (9.3%) |
| White (%) | 106 (32.2%) | 14 (23%) | 29 (53.7%) |
| African American (%) | 183 (55.7%) | 39 (63.9%) | 19 (35.2%) |
| Hispanic or Latino/a (%) | 6 (3.1%) | 2 (3.3%) | 5 (9.3%) |
| Multiracial | 17 (5.2%) | 5 (8.2%) | 1 (1.9%) |
| Median age | 37 years | 35 years | 30 years |
Respondents were obtained using respondent-driven sampling (RDS), which involves initial participants passing the survey onto others within their cohort (i.e., sex workers), circumventing the challenging and time-consuming task of developing rapport with large numbers of hard-to-reach respondents. Social network connections between respondents are recorded to calculate the amount of homophily (i.e., “birds of a feather flock together”) among the initial respondents and their referrals, which is used to statistically compensate for the bias produced by the non-random sampling approach. The first step of an RDS survey is to identify “seed” participants; in this case, individuals known to be involved in various forms of sex work. Each surveyed individual was compensated $10 and provided five referral coupons; another $10 was awarded for each additional referral. Coupons could be redeemed at convenient locations within each city on specified days and times.
The survey was designed to be geographically comprehensive and included eight of Ohio’s principal cities: Toledo, Dayton, Columbus, Cleveland, Cincinnati, Youngstown, Lima, and Chillicothe. Regrettably, data from Youngstown and Chillicothe were unable to be collected because of a lack of seed referrals. The lone respondent from Lima was counted as a member of the Dayton group because of the city’s small sample size and proximity to Dayton. The survey represented 52% of Ohio’s 20 largest cities’ populations. Columbus was home for 44% of the respondents, and Toledo and Cleveland each comprised roughly 15%of the sample. Cincinnati and Dayton covered 13.7% and 11.9% of the respondents, respectively. This dataset is assumed to be geographically representative of sex workers in Ohio’s urban commercial sex markets because the survey covers more than half of Ohio’s urban population centers.
The typology of sex workers represented in the survey was ascertained by examining their self-identified work environments and income levels. Of this sample, 53% of respondents reported performing sexual services in a motel or a bar, 41% via street work and 18% in a strip club or lingerie store (i.e., locations used by street-level workers). However, 60% of respondents reported performing acts in locations where both street-based and higher-end sex workers perform services, including clients’ homes, the sex worker’s home, or an arranged location.
Admittedly, work environments do not completely define the type of sex worker; therefore, self-reported annual household income was used as an additional criterion. The majority of respondents (73%) reported annual incomes of less than $10,000, below the 2011 individual poverty level defined by the U.S. Department of Health and Human Services (U.S. Census Bureau, n.d.). Approximately 10% of respondents earned between $11,000 and $39,000 annually; 1.2% of respondents recorded annual incomes between $40,000 and $99,999; and approximately 1% earned more than $100,000 annually. Lower income respondents are likely represent informal or street-based workers earning low and subsistence-level incomes, whereas the uppermost income class probably represents high-end escorts. This sample is skewed towards lower-income respondents, which may accurate depict the empirical distribution of sex worker stratums (Magnani, Sabin, Saidel, & Heckathorn, 2005).
False responses are often elicited when participants are confronted with questions on sensitive topics or queries that present conflicting interests, such as abuse or criminal activity. To mitigate this problem, questions regarding trauma most would feel uncomfortable discussing or disclosing with strangers were avoided, particularly rape, childhood sexual abuse, and abuse from traffickers or clients.
The survey questions used to create the models’ variables are found in Table 3. All variables are affirmative responses to the respective event or descriptor, dummy-coded as 1. Age was the only continuous variable. Variables with the temporal designation “before involvement in sex work” are events that transpired within the year or more than 1 year before involvement in sex work. Conversely, events that transpired within-the year and more-than 1 year-after timeframes are labeled “after involvement in sex work.”
| Constructed variable | Survey questions (affirmative responses) |
|---|---|
| Minor sex trafficking victims* | • Affirmed they were forced into sex work and were currently younger than 18 years old, or • Affirmed they were minors in both of the following categorical questions: “At what age did you first receive money for sexual services?” and “At what age did you first sell sexual services?” |
| Adult sex trafficking victims* | • Affirmed they were forced into sex work and currently older than 18 years, or • Affirmed “Why are you unable to leave sex work: Someone forced you to continue to participate” or • “Do you currently have a pimp?” |
| Engaged in survival sex while running away | • Affirmed “I entered sex work because I was teenage runaway and traded sex for things I needed” or • Affirmed “When I ran away, I traded sex for food, clothing, and shelter.” |
| Friends involved in the sex trade (bought sex or sold others for sex) | • Affirmed “I had friends who purchased sex within a year/more than a year before I entered sex work.” or • Affirmed “I had friends who sold others for sex within a year/more than a year before I entered sex work.” |
| Did not seek help out of sex work | • Affirmed “I did not seek help out of sex work.” • The omitted, reference category was “Yes, someone showed an interest in helping me or seeing that I got help.” |
| No one would help or believe you when you tried to get help out of sex work | • Affirmed “No one would help you or believe you when you tried to leave sex work.” • The omitted reference category was created from the question: “Yes, someone showed an interest in helping me or seeing that I got help.” |
| Adults respondents who were trafficking victims when minors | • Variable created by choosing respondents categorized as minor trafficking victims who were currently older than 18 years. |
Analytic Strategy
The analysis used two multivariate logistic regression models. The dependent variables were dichotomous (i.e., victim/not a victim); consequently, the classical linear regression assumption of homoscedasticity and a normally distributed error term becomes untenable, as does the model’s ability to predict values exceeding 1 or under 0. Therefore, either logistic or probit regression models should be considered. Logistic regression was chosen because its coefficients are more easily interpretable as odds ratios than probit marginal effects, and because no theoretical foundation exists to select the probit link function, which assumes the latent factors generating the dependent variable are normally-distributed.
Baseline models consisted of the hypothesized risk factors and control variables; additional risk factors were included based on their statistical fit in the model. Logistic regression generally abides by the heuristic rule of 10, meaning 10 dependent variable outcome events are required for each explanatory variable. Vittinghoof and McCulloch (2007) discovered the rule of 10 can be relaxed to five events per-variable with minimal effects on bias and efficiency. Compromising at seven events per-variable, as many as nine variables were permitted in the minor trafficking risk model and eight variables in the adult trafficking risk model. Backwards stepwise selection (selection threshold p = 0.1) was used to construct the models; however, variables essential to the hypotheses remained in the model regardless of statistical significance. Correct classification percentages and the predicted area under the receiver operating characteristic (ROC) curve were used as a relative performance measure assessing the model’s ability to correctly predict sex trafficking outcomes. Missing data were evident, but not a major concern. Two respondents did not report their current age, and therefore, those observations were excluded from the adult trafficking dataset.
To ensure the veracity of the models, coefficients were estimated with robust standard errors and multicollinearity, goodness-of-fit, and model specification tests were performed. Potential correlation among same-city respondents jeopardizes the logistic regression assumption of an independent error term; therefore, intra-city correlation was accounted for using STATA’s clustering option, which produces robust standard errors (Nichols & Schaffer, 2007). Cities potentially differ according to demographics, policing policies, and other unobservable disparities, such as data collectors’ interactions and relationships with respondents. Multicollinearity was managed by ensuring the variable inflation factor was less than five and the condition number was under15.
Wald tests determined whether the model statistically significantly differed from complete randomness. Additionally, model specification was tested using Stata’s linktest command, which essentially tested whether any statistically significant explanatory variables could be added to the model other than by chance (StataCorp, 2011). In addition, each model was required to pass the Hosmer-Lemeshow goodness-of-fit test, as recommended by Peng, Lee, and Ingersoll (2002).
Results
The results of the minor and adult sex trafficking models explained approximately 18% and 28% of the variation in the data (based on Nagelkerke R2) and supported most of the hypotheses while passing the robustness and specification tests (see Tables 4 and 5). The following five factors increased the odds of being forced into minor sex trafficking: Engaging in survival sex, having friends who bought or sold others for sex, difficulty in school before sex work, spending time in juvenile detention, conflict with parents before involvement in sex work, and ascribing as multiracial. Engaging in survival sex while running away had the greatest effect, increasing the odds of being forced into sex work as a minor by approximately 2.6 times (160%), compared with those who did not engage in survival sex while running away. Respondents having friends who bought or sold others for sex had 2.16 (116%) greater odds of victimization than respondents without such friends. Previous conflict with parents increased the odds of minor victimization by 1.97 (97%). When compared with African American respondents, the multiracial control variable increased the odds of being a minor trafficking victim by 1.84 (84%).
| Covariate | Odds ratio | p | Robust SE | Lower Bound | Upper Bound |
|---|---|---|---|---|---|
| Ran away from home | 0.72 | 0.099 | 0.14 | 0.49 | 1.06 |
| Homeless before sex work | 1.05 | 0.896 | 0.42 | 0.48 | 2.30 |
| Survival sex while running away | 2.61 | 0.004 | 0.87 | 1.36 | 5.03 |
| Friends bought sex and sold others for sex | 2.16 | 0.059 | 0.88 | 0.97 | 4.79 |
| Difficulty in school before involved in sex work | 2.14 | 0.002 | 0.53 | 1.31 | 3.49 |
| Did not get along with parents before sex work | 1.97 | 0.000 | 0.32 | 1.43 | 2.71 |
| Spent time in a juvenile detention center before involved in the sex trade | 2.02 | 0.012 | 0.56 | 1.17 | 3.48 |
| White | 0.62 | 0.194 | 0.23 | 0.30 | 1.28 |
| Multiracial | 1.84 | 0.000 | 0.31 | 1.32 | 2.57 |
| Hispanic | 1.58 | 0.605 | 1.40 | 0.28 | 9.00 |
| Covariate | Odds ratio | p | Robust SE | Lower Bound | Upper Bound |
|---|---|---|---|---|---|
| Homeless before involved in sex work | 1.80 | 0.018 | 0.45 | 1.11 | 2.93 |
| No one would help or believe you when you tried to get help out of sex work | 2.63 | 0.036 | 1.21 | 1.07 | 6.46 |
| I did not seek help out of sex work | 0.58 | 0.066 | 0.18 | 0.30 | 1.04 |
| Forced as a minor, now an adult | 1.72 | 0.227 | 0.78 | 0.71 | 4.17 |
| Current age | 0.94 | 0.000 | 0.02 | 0.91 | 0.97 |
| Female | 4.71 | 0.027 | 3.30 | 1.20 | 18.56 |
| White | 2.46 | 0.041 | 1.09 | 1.04 | 5.84 |
| Multiracial | 0.44 | 0.421 | 0.45 | 0.06 | 3.25 |
| Hispanic | 3.75 | 0.035 | 2.36 | 1.10 | 12.84 |
The adult sex trafficking risk model produced seven statistically significant coefficients: homelessness before involvement with the sex trade, unsuccessfully attempting to exit sex work, not trying to exit sex work, current age, and two of the race control variables (White and Hispanic). Adult trafficking victims who attempted to leave sex work but did not find help in doing so were at 2.63 times (163%) greater odds of being coerced than those who found help. Victims who did not seek help were at 0.56 times (56%) lower odds of being currently coerced into sex work than those who successfully found help. Previous sex trafficking victimization as a minor did not significantly affect adult victimization outcomes (p = 0.227). The odds of being currently trafficked decreased by approximately 0.06 (6%) for every 1-year increase in age. The odds a White adult is currently being trafficked are about 2.46 times (146%) greater than the odds for African Americans, and 3.75 times (275%) greater for Hispanics than African Americans.
Discussion
This study used a representative sample of Ohio urban, street-based sex workers to measure the effect of direct risk factors for minor and adult sex trafficking on empirical victimization outcomes. The estimates provide a suitable inferential basis to discuss and prioritize sex trafficking risk factors, particularly regarding the role of survival sex on minor victimization and the availability of assistance for escaping victims on adult victimization, which can be used to direct future research and potentially motivate anti-trafficking policy and programming in Ohio.
Both sex trafficking models produced statistically significant variables and validated the majority of the hypotheses. Hypotheses 1 and 2 were supported; that is, engaging in survival sex and having peers involved in the sex trade were statistically significant risk factors for minor sex trafficking. Respondents’ experience leaving sex work was found to be a magnitudinous, positive correlate for adult trafficking victims, supporting Hypothesis 3. Hypothesis 4 was not supported; sex trafficking victimization as a minor had no statistically significant effect on adult trafficking outcomes.
Engaging in survival sex while running away from home produced a positive, statistically significant coefficient; however, homelessness was statistically insignificant and running away from home was marginally statistically significant and actually lowered the odds of minor victimization—the literature cites both as a risk factor (Cobbina & Oselin, 2011; Edwards, Iritani, & Hallfors, 2006; Halcón & Lifson, 2004; Reid, 2012b; Roe-Sepowitz, 2012; Saewyc, MacKay, Anderson, & Drozda, 2008). The compounding effect of other negative informal social controls closely related to transience, such as conflict with parents and difficulty at school, was likely the reason why homelessness and running away from home before sex work were statistically insignificant.
The effect of survival sex on victimization is unsurprising because engaging in survival sex increases sex market exposure by attracting repeat buyers and eventually sex traffickers, concurring with Johnson’s (1992) claim that as many as 90% of minor females who sell sex are trafficked. Alternatively, survival sex could be a manifestation of relational shame, defined as self-denigrating beliefs and behaviors pertaining to romantic/sexual relationships, which Reid (2010) found to positively contribute to minor victimization, but also confound the effect of running away from home. Nevertheless, this result suggests the qualitative nature of the runaway or homeless experience (i.e., selling sex for basic necessities) is indeed the risk behind youth transience. However, the potential endogenous relationship between relational shame and survival sex should be addressed in future research.
Prospective efforts to reduce minor sex trafficking victimization should focus on runaways and homeless children, including establishing greater number of safe youth shelters and providing basic resources such as food, water, shelter, and medical supplies (Kotrla & Wommack, 2011; Reid, 2012b). At-risk children who have access to basic necessities are less likely to resort to survival sex (see Greene, Ennett, & Ringwalt, 1997; Whitbeck, Hoyt, Johnson, Berdahl, & Whiteford, 2002). Another approach to reduce survival sex incidence is by educating runaway and homeless youths about the potential dangers of street life and offering guidance on how to avoid high risk situations. Ideally, outreach groups would provide this information (digitally and in physical form) along with information and resource maps of charitable services where basic necessities can be obtained. For example, a directory of resources for youth in the Cleveland area has been implemented by the local United Way 2-1-1; resource information can be accessed 24/7 by phone or online (
The consequential role of peers involved in buying sex or selling others for sex supported Hypothesis 2 and is consistent with the findings of other research on the topic (Cobbina & Oselin, 2011; Kennedy et al., 2007; Loza et al., 2010). Further examination using a survey question that identified victims’ traffickers offered clues on the potential dynamics of respondents’ otherwise unknown social networks. An overview of these relationships can be found in Table 6. Traffickers defined as “an unrelated male who first acted as a friend or boyfriend or a family member” typically target female victims (Azaola, 2000; Clawson, Dutch, Solomon, & Grace, 2009; Curtis, Terry, Dank, Dombrowski, & Khan, 2008).
| Trafficker (person forced by) | Minor Victims | Adult Victims |
|---|---|---|
| An unrelated male who first acted like a boyfriend or friend | 23% | 41% |
| An unrelated female who first acted like a girlfriend or friend | 31% | 13% |
| An unrelated female or male who first acted like a girlfriend, boyfriend, or friend | 54% | 54% |
| A family member | 13% | 6% |
Males referenced in Table 6 were most likely traffickers; however, it is possible some worked within the trafficker’s network as connectors or recruiters. Female “friends” may have possibly also worked as connectors or recruiters for their ability to quickly develop ties with women and underage girls. Connectors are individuals who profit from introducing individuals to a recruiter or trafficker, but are otherwise generally unaffiliated with the sex trade (Williamson & Prior, 2009). Recruiters work with traffickers peripherally (e.g., dealing drugs) or can be trafficking victims who the trafficker forces to convince “fresh faces” to expand the trafficker’s stable, a practice referred to as knocking. The importance of network positions and influence in victims’ social networks is noteworthy and should be the considered in future research on victim recruitment.
The regression model indicated difficulty in school doubled the odds of being trafficked into sex work as a minor. Kaestle (2012) found school connectedness enhances self-confidence and lowers the risk of minor sex trafficking as well as contributing to lower rates of delinquency, conduct problems, drop out, teenage pregnancy, mental health issues, and substance abuse. Estes and Weiner (2001) suggested difficulty in school lowered self-esteem, and similarly, Hawkins and Weiss (1985) found poor academic performance contributed to feelings of isolation and frustration among youth. Consequently, such youth will often seek validation and acceptance elsewhere—potentially sex traffickers who discourage school attendance and enable the youth to disregard school.
Spending time in a juvenile detention center almost doubled the odds of being trafficked into sex work as a minor. Detention in a juvenile corrections facility suggests the individual lacks impulse control and has low self-esteem (Twill, Green, & Traylor, 2010). These traits, particularly low self-esteem, render individuals especially susceptible to traffickers’ false promises of friendship, security, and excitement (Flowers, 2001). Williamson and Prior (2009) found rehabilitation-related stress caused some detainees to shut down or run away, suggesting ineffective, high-stress rehabilitation strategies rewarding immediate results might actually be detrimental long-term. Such therapy strategies might prove effective in the short-term; however, these strategies largely overlook underlying issues that might be addressed with more intensive treatment such as, psychodynamic psychotherapy. Exposure to correctional detainees involved with sex traffickers increases the risk of recruitment among incarcerated adult females, and the risk apparently persists among girls in juvenile corrections—a topic worthy of examination in future research (Meekins, 2013).
Conflict with parents before sex trafficking victimization increased the odds of a minor becoming a trafficking victim by about 97%; however, parental conflict was nebulously defined by the survey, meaning the nature of these disputes is unknown. Nevertheless, children engaged in parental conflict are particularly susceptible because traffickers often promise a simulated family structure and the order and security they did not receive from their parents (Jesson, 1993; Martin, Hearst, & Widome, 2010; Walker, 2002). In addition to parental conflict, Nadon, Koverola, and Schludermann (1998) found trafficked minors were also more likely to have low family cohesion; however, although this variable characterizes negative informal social control, it also may confound running away from home, peer influence, or difficulty in school. Surprisingly, the only statistically significant (p = 0.05) Pearson correlation between the aforementioned variables was running away from home (r = 0.17). Based on Pearson correlation tests and the simultaneous statistical significance of parental conflict, the variable’s was likely psychologically manifested through personal traits contributing to poor decision-making skills, and the need for external validation or stability, in addition to further-reaching impacts, such as lowered academic performance and positive peer influence.
Although minority and indigenous groups have been identified as at greater risk of becoming victims of domestic minor sex trafficking victimization, the coefficient estimates indicated multiracial respondents (n = 17) were the only group for which the risk of minor sex trafficking reached the level of statistical significance (66% greater odds than the African American reference group; Acharya, 2009, 2010; Clawson et al., 2009; Deer, 2010; Saewyc et al., 2008). Most of the multiracial survey respondents originated from Columbus and Toledo, and reported a racial/ethnic combination of African American, White, or Hispanic. It is difficult to ascertain any empirically proven psychological issues explaining why multiracial individuals are at greater risk of sex trafficking because, as a group they generally do not devalue their racial identity as much as monoracial minorities and exhibit higher self-esteem scores (Hanna, 2002; Shih & Sanchez, 2005). Multiracial individuals’ higher risk might be the result of a combination of their marketability (due to their typically lighter complexions) and the negative structural externalities of multiracial population clusters in Ohio (typically highly impoverished areas; Acharya, 2010; Flowers, 2001; Tyler, 2009). In fact, spatial correlation exists between the U.S. Census’ 2010 American Community Survey 5-year estimates for the percentage of families in poverty and those identifying as two or more races, among Ohio Census Tracts. A bivariate local Moran’s I, which measures whether similar values of each variable correlate in space (on a scale from -1 to 1, with values closer to 1 indicating high levels of spatial clustering) produced a modest Moran’s I of 0.19 (p = 0.001). Further, local clusters of high-poverty Census Tracts with a large proportion of multiracial residents were overwhelmingly located in the state’s largest inner cities, which is indicative of multiracial individual’s exposure to the underground sex trade. The bivariate Moran’s I statistic was computed using GeoDa (Anselin, Ibnu, & Youngihn, 2006).
The adult trafficking risk model produced seven statistically significant variables, validating 2 of 3 hypotheses. Adult trafficking victims unable to find help to leave sex work have roughly 180% greater odds of currently being an adult trafficking victim than those who found help to escape, supporting Hypothesis 3. Victims’ difficulty or inability to find assistance might be related to community risk factors observed by other scholars, such as deficient trust with social services (Clawson et al., 2009), the overall lack of anti-sex trafficking resources (Estes & Weiner., 2001), and an inept police response to customer and trafficker violence (Deer, 2010; Williamson & Cluse-Tolar, 2002). Nevertheless, this finding is supported by the informal social control theory framework because the inability to locate the adequate resources and/or failure to convince others to assist is a definite indicator of social capital deficiency.
The alternate response, “I did not seek help out of sex work” decreased the odds of being victimized as an adult by 40%. This finding is the likely result of affirmative responses from independent sex workers who never sought to leave their occupation. However, it is difficult to distinguish between victims who did not seek help and independent workers with no desire to quit; therefore, this variable is interpreted with uncertainty and understood as a necessary component of the categorical survey question from which the variable was created. The lack of community resources, anti-trafficking services, and a questionable sense of security for victims from their traffickers while escaping suggest policy research should focus on sex trafficking awareness media and programming in communities known to host underground sex markets, locating outreach organizations closer to high-risk areas, and protecting victims from reprisals.
Minor sex trafficking victimization did not statistically significantly increase the odds of victimization as an adult, invalidating Hypothesis 4. This finding could be explained by one or both of the following: (a) reporting bias caused by traffickers’ greater physical or emotional control over long-time victims, thus reducing victims’ ability or incentive to take the survey compared with relatively new, lesser-entwined victims, or (b) most of the survey’s minor trafficking victims are able to escape by adulthood. Only 13 minor victims (24%) were trafficked as adults; the youngest adult victim trafficked as a minor was 21 and the next youngest was 24—the oldest was 47. Older respondents with minor trafficking histories might have once escaped, but returned to traffickers because of low social capital, financial hardship, or continuing antisocial behaviors established as a minor (Zimmerman et al., 2003). Responsibility for dependent children appeared to be an additional intuitive adult trafficking risk; in fact, 77% of current adult victims trafficked as a minor had children, compared with 30% of current adult victims without minor trafficking histories (Cobbina & Oselin, 2011; Loza et al., 2010). This finding might also be evidence of a control tactic used by traffickers that involves impregnating the victim and using the child as leverage by threatening physical harm or refusal to grant the victim custody if she threatens to leave or escape (Flowers, 2001; Johnson, 1992). In addition, minor trafficking victims might be reluctant to escape if they develop loyalty and love for their trafficker, especially those promoted as their trafficker’s bottom, which is the woman in closest rank with the trafficker who is assigned to manage other victims (Hanna, 2002; Williamson & Cluse-Tolar, 2002).
Adult trafficking victims who experienced homelessness before entering sex work have a higher probability of unemployment, drug addiction, and low education attainment; thus, motivating a reliance on informal occupations. Moreover, these victims have low social capital due to disruptive socialization during their youth and few available gainful employment options, rendering the traffickers’ promises of financial security an attractive option (Clawson et al., 2009; Cobbina & Oselin, 2011; Estes & Weiner, 2001). Aside from providing temporary employment opportunities or job training, improving the self-efficacy of homeless individuals—perhaps by conducting outreach activities centered on recognizing and deterring sex traffickers—may also prove beneficial.
The adult victimization model also supported subhypothesis 1—the risk of adult trafficking differs among racial and ethnic groups. White and lighter-skinned women are in higher demand by the sex market than darker-skinned women (Acharya, 2010; Flowers, 2001; Tyler, 2009). Therefore, unsurprisingly, White and Hispanic respondents produced much greater odds of being trafficked as an adult than darker-skinned African Americans. The curiously high coefficient for Hispanics might reflect involvement with Hispanic trafficking networks, which cater exclusively to Hispanic men through outcall escort services or brothels frequently located near cantinas (Polaris Project, 2011). The Hispanic adult trafficking victims resided in Cleveland and Toledo, cities home to Ohio’s two largest Hispanic populations. Interestingly, no Hispanic victims reported working at migrant labor camps that primarily service transient Hispanic men.
Subhypothesis 2 was also validated because the odds of victimization decreased with advancing age. The model indicated every additional year of age decreases the odds of adult victimization by 6%. The predicted probability of adult victimization peaks at about age 35 years; however, the peak in the raw data occurs around age 30 years. This result is realistic because older, victims, especially those with a history of ongoing substance abuse, eventually become unprofitable due to lack of marketability or productivity and are typically discarded by traffickers (Williamson & Cluse-Tolar, 2002). One potential actionable result of this finding is targeting a female demographic younger than 35 years when planning adult trafficking research surveys and outreach efforts in Ohio.
Conclusion
Ohio is a well-known sex trafficking recruiting state; therefore, measuring direct trafficking risk factors is important for developing and enacting policies that reduce the number of sex trafficking victims in the state and surrounding region. Minor and adult sex trafficking victimization models were based on Reid’s (2012a) recommendation of Sampson and Laub’s (1993, 2005) aged-graded informal social control theory within a life course framework. Four hypotheses and two subhypotheses regarding risk factors of particular interest were constructed and tested along with several other relevant risk factors. The multivariate logistic regression models’ coefficient estimates validated most of the hypotheses and found significant several other direct risk factors promoting minor and adult sex trafficking outcomes among Ohio urban, street-based sex workers. The results of this analysis were used in a report to an Ohio Senate caucus promoting a federal bill regarding missing children.
This research contributed to the sex trafficking research literature by producing sound evidence that the qualitative nature of transience and the unavailability of assistance when attempting to break away from sex work a have statistically-significant and positive effect on sex trafficking outcomes. Engaging in survival sex while running away was the most influential direct risk factor for minor sex trafficking victimization; thus, available resources may be better allocated toward education and outreach for runaway and homeless children in Ohio. Failure to find adequate assistance when attempting to leave from sex work was shown to considerably increase the odds of adult victimization, indicating adult victims’ lack of social capital creates barriers to their escape, even when they have the will to try. Educating communities on how to identify and aid sex trafficking victims, while shielding all parties from retaliation by their trafficker, should be considered a viable outreach strategy for adult victims.
Limitations and Future Research
Although the size and representativeness of the survey was sufficient for an informative analysis, the dataset presented several limitations. Principally, the findings relate only to Ohio urban, street-based sex workers, and thus, cannot be applied to the state’s general population. However, this more narrow focus is also beneficial because urban, street-based sex workers are an especially high-risk population. In addition, sex trafficking victims might have been underrepresented in the survey because traffickers regulate victims’ schedules, obstructing their participation (Williamson & Cluse-Tolar, 2002).
This study’s cross-sectional survey design inhibited causal inferences. Ideally, trafficking victims would be compared with demographically similar individuals from proximate geographic locations in a case-control study, or longitudinally in a cohort design, such as that used by Reid (2010). However, such datasets are cost-prohibitive; therefore, the demand and willingness-to-pay for such surveys must exceed high cost of such surveys in order to generate more reliable and generalizable inferences in future sex trafficking research.
Proper RDS statistical sampling bias correction could not be implemented for the sample because the network violated the reciprocity assumption and featured incomplete referral data. More than one quarter (26%) of respondents listed their reference as a stranger, 22% listed as neighbors or someone they “kind of know,” and the remaining 52% listed their reference as friends or family members, the ideal relation for reciprocity, but not nearly enough to fulfill the reciprocity assumption of RDS. Inconsistent survey responses were removed from the final dataset, further fracturing the referral network, and therefore, violating RDS the assumption of complete interconnectedness among all respondents (Schonlau & Liebau, 2010). Nevertheless, this loss of functionality might be muted by the questionable effectiveness of RDS’ bias correction capability, which has been critiqued for producing inflated variance estimates and still lacks robust empirical validation (see Gile & Handcock, 2010; Goel & Salganik, 2010; McCreesh et al., 2010).
The dichotomous survey question format hindered the analytic potential of the dataset. Qualitative survey questions specifying the extent respondents experienced risk factors are preferable to the binary choice of having/not having experienced the event, condition, or trauma. Similarly, the amount of time individuals ran away from home or were homeless would be useful for assessing how the duration of runaway episodes affects the odds of victimization.
Questions assessing personality traits—such as levels of naïveté, self-esteem, and self-efficacy—would have been valuable to more accurately measure the role of the psyche and personality traits on contextual decision making. Such questions could be used to measure the role of individual differences and temperament on informal social control process, as suggested by Reid (2012a).
The author thanks the anonymous reviewers for providing very helpful feedback that drastically improved the manuscript. Special thanks to Samantha McNamara for patiently applying her editing skills to greatly improve the readability and conciseness of the article. Thank you, Dr. Celia Williamson of the University of Toledo, for the opportunity to conduct this research.
Notes
Michael L. Chohaney is doctoral candidate in Spatially-Integrated Social Science at The University of Toledo.
Correspondence regarding this article should be sent to Michael L. Chohaney, University of Toledo, Department of Geography and Planning, SM 3000, Mail Stop 140, 2801 W. Brantford Rd, Toledo, OH 43606.
References
Anselin, L., Ibnu S., & Youngihn K. (2006). GeoDa: An introduction to spatial data analysis. Geographical Analysis, 38, 5–22. http://dx.doi.org/10.1111/j.0016-7363.2005.00671.x Acharya, A. K. (2009). The dynamic of internal displacement, forced migration, and vulnerable to trafficking in Mexico. Journal of Human Ecology, 27(3), 161–170. Acharya, A.K. (2010). Feminization of migration and trafficking of women in Mexico. Review of Research and Social Intervention, 30, 19–38. Azaola, E. (2000), Boy and girl victims of sexual exploitation in Mexico. Mexico City: United Nations Children’s Fund. Retrieved from http://www.oas.org/atip/country%20specific/azaola%20mexico%20child%20sex%20exploitation.pdf Belser, P. (2005). Forced labour and human trafficking: Estimating the profits. (Working paper no. 42). Geneva, Switzerland: International Labour Organization. Retrieved from http://digitalcommons.ilr.cornell.edu/cgi/viewcontent.cgi?article=1016&context=forcedlabor Clawson, H. J., Dutch, N., Solomon, A., & Grace, L.G. (2009). Human trafficking into and within the United States: A review of the literature. Washington, DC: U.S. Department of Health and Human Services. Retrieved from http://aspe.hhs.gov/hsp/07/humantrafficking/litrev/ Cobbina, J. E., & Oselin, S. S. (2011). It’s not only for money: An analysis of adolescent versus adult entry into street prostitution. Sociological Inquiry, 81, 310–332. http://dx.doi.org/10.1111/j.1475-682X.2011.00375.x Curtis, R., Terry, K., Dank, M., Dombrowski, K., & Khan, B. (2008). Commercial sexual exploitation of children in New York City, Volume one: The CSEC population in New York City: Size, characteristics, and needs. New York, NY: John Jay College of Criminal Justice. Retrieved from http://www.courtinnovation.org/sites/default/files/CSEC_NYC_Volume1.pdf Deer, S. (2010). Relocation revisited: Sex trafficking of native women in the United States. William Mitchell Law Review, 36(2), 622–683. http://open.wmitchell.edu/facsch/157 Edwards, J. M., Iritani, B. J., & Hallfors, D. D. (2006). Prevalence and correlates of exchanging sex for drugs or money among adolescents in the United States. Sexually Transmitted Infections, 82, 354–358. http://dx.doi.org/10.1136/sti.2006.020693 Elder Jr., G. H., Johnson, M. K., & Crosnoe, R. (2003). The emergence and development of life course theory. In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of the life course. New York, NY: Springer. Estes, R. J., & Weiner, N. A. (2001). The commercial sexual exploitation of children in the U.S., Canada, and, Mexico [Report]. Philadelphia, PA: University of Pennsylvania. Retrieved from http://www.thenightministry.org/070_facts_figures/030_research_links/060_homeless_youth/CommercialSexualExploitationofChildren.pdf Flowers, R. B. (2001). Runaway kids and teenage prostitution: America’s lost, abandoned, and sexually exploited children. Westport, CT: Praeger. Gile, K. J., & Handcock, M. S. (2010). Respondent-driven sampling: An assessment of current methodology. Sociological Methodology, 40, 285–327. http://dx.doi.org/10.1111/j.1467-9531.2010.01223.x Goel, S., & Salganik, M. J. (2010). Assessing respondent-driven sampling. Proceedings of the National Academy of Sciences, 107, 6743–6747. http://dx.doi.org/10.1073/pnas.1000261107 Greene, J. M., Ennett, S. T., & Ringwalt, C. L. (1997). Substance use among runaway and homeless youth in three national samples. American Journal of Public Health, 87, 229–235. http://dx.doi.org/10.2105/AJPH.87.2.229 Ohio House Bill 262. (Human trafficking law with safe harbor provisions). H.B. 262, 129th General Assembly. (2012). Ohio Senate Bill 235. (Human trafficking as a specific criminal offense). S.B. 235, 128th General Assembly. (2010). Halcón, L. L., & Lifson, A. R. (2004). Prevalence and predictors of sexual risks among homeless youth. Journal of Youth and Adolescence, 33, 71–80. http://dx.doi.org/10.1023/A:1027338514930 Hanna, C. (2002). Somebody’s daughter: Domestic trafficking of girls for the commercial sex industry and the power of love. William and Mary Journal of Women and Law,29(1), 1–29. Retrieved from http://scholarship.law.wm.edu/wmjowl/vol9/iss1/2 Hawkins, J. D., & Weiss, J. G. (1985). The social development model: An integrated approach to delinquency prevention. Journal of Primary Prevention, 6, 73–97. http://dx.doi.org/10.1007/BF01325432 Holger-Ambrose, B., Langmade, C., Edinburgh, L. D., & Saewyc, E. (2013). The illusions and juxtapositions of commercial sexual exploitation among youth: Identifying effective street-outreach strategies. Journal of Child Sexual Abuse, 22, 326–340. http://dx.doi.org/10.1080/10538712.2013.737443 Jesson, J. (1993). Understanding adolescent female prostitution: A literature review. British Journal of Social Work, 23, 517–530. http://dx.doi.org/10.1093/bjsw/23.5.517 Johnson, J. J. (1992). Teen prostitution. Danbury, CT: Franklin Watts. Kaestle, C. E. (2012). Selling and buying sex: A longitudinal study of risk and protective factors in adolescence. Prevention Science, 13, 314–322. http://dx.doi.org/10.1007/s11121-011-0268-8 Kennedy, M. A., Klein, C., Bristowe, J. T. K., Cooper, B. S., & Yuille, J. C. (2007). Routes of recruitment: Pimps’ techniques and other circumstances that lead to street prostitution. Journal of Aggression, Maltreatment & Trauma, 15(2), 1–19. http://dx.doi.org/10.1300/J146v15n02_01 Kotrla, K., & Wommack, B.A. (2011). Sex trafficking of minors in the U.S.: Implications for policy, prevention and research. Journal of Applied Research on Children, 2, Article 5. Retrieved from http://digitalcommons.library.tmc.edu/childrenatrisk/vol2/iss1/5/ Krauss, L. (2014, April 19). Ohio a Major Hub for Interstate Trafficking. The Oberlin Review. Retrieved from http://oberlinreview.org/5567/news/ohio-a-major-hub-for-interstate-trafficking/ Logan, T. K., Walker, R., & Hunt, G. (2009). Understanding human trafficking in the United States. Trauma, Violence and Abuse, 10, 3–30. http://dx.doi.org/10.1177/1524838008327262 Loza, O., Strathdee, S. A., Staines, H., Ojeda, V. D., Martinez, G. A., Amaro, H., … Proyecto, M. S. (2010). Correlates of early versus later initiation into sex work in two Mexico–U.S. border cites. Journal of Adolescent Health, 46, 37–44. http://dx.doi.org/10.1016/j.jadohealth.2009.06.011 Magnani, R. A., Sabin, K. B.. Saidel, T. A., & Heckathorn, D. C. (2005). Review of sampling hard-to-reach and hidden populations for HIV surveillance. AIDS, 19, S67–S72. http://dx.doi.org/10.1097/01.aids.0000172879.20628.e1 Martin, L., Hearst, M. O., & Widome, R. (2010). Meaningful differences: Comparison of adult women who first traded sex as a juvenile versus as an adult. Violence Against Women, 16, 1252–1269. http://dx.doi.org/10.1177/1077801210386771 McCreesh, N., Frost, S.D.W., Seeley, J., Katongole, J., Tarsh, M. N., Ndunguse, R., … White, R. G. (2010). Evaluation of respondent-driven sampling. Epidemiology, 23, 138–147. http://dx.doi.org/10.1097/EDE.0b013e31823ac17c Meekins, J. O. (2013). Human trafficking in jails and correctional institutions. [PowerPoint]. Presented at the 10th Annual International Sex Trafficking and Human Trafficking Conferences, Toledo, OH. Nadon, S., Koverola, C., & Schludermann, E. (1998). Antecedents to prostitution. Journal of Interpersonal Violence, 13, 206–221. http://dx.doi.org/10.1177/088626098013002003 National Human Trafficking Resource Center. (n.d.). Fake massage businesses. Retrieved from http://www.traffickingresourcecenter.org/sex-trafficking-venuesindustries/fake-massage-businesses Nichols, A., & Schaffer, M. E. (2007, September 10). Clustered standard errors in Stata. Presented at the United Kingdom Stata Users’ Group Meeting, London, UK. Retrieved from http://EconPapers.repec.org/RePEc:boc:usug07:07 Peng, C., Lee, K., & Ingersoll, G. M. (2002). An introduction to logistic regression analysis and reporting. Journal of Educational Research, 96, 3–14. http://dx.doi.org/10.1080/00220670209598786 Perdue, T., Williamson, C., Billings, M., Schart, J., & Boston-Gromer, R. (2011). In the matter of human trafficking in Ohio: The pursuit for justice continues. Women’s Policy Journal of Harvard, 8, 3–15. Polaris Project. (n.d.). Street prostitution. Retrieved from http://www.polarisproject.org/human-trafficking/sex-trafficking-in-the-us/street-prostitution Polaris Project. (2011, April 15). Latino residential brothels and related sex trafficking networks at-a-glance. Retrieved from Content Pack of the National Human Trafficking Resource Center Newsletter available at: http://act.polarisproject.org/o/5417/t/0/blastContent.jsp?email_blast_KEY=1160584 Reid, J. A. (2010). A pathway to child sex trafficking in prostitution: The impact of strain and risk inflating responses (Doctoral dissertation, University of South Florida). Scholar Commons Graduate Theses and Dissertations. Retrieved from http://scholarcommons.usf.edu/etd/1747 Reid, J. A. (2011). An exploratory model of girl’s vulnerability to commercial sexual exploitation in prostitution. Child Maltreatment, 16, 146–157. http://dx.doi.org/10.1177/1077559511404700 Reid, J. A. (2012a). A girl’s pathway to prostitution: Linking caregiver adversity to child susceptibility. El Paso, TX: LFB Scholarly. Reid, J. A. (2012b). Exploratory review of route-specific, gendered, and age-graded dynamics of exploitation: Applying life course theory to victimization in sex trafficking in North America. Aggression & Violent Behavior, 17, 257–271. http://dx.doi.org/10.1016/j.avb.2012.02.005 Rocha, P. A. (2012). Our backyard slave trade: The result of Ohio’s failure to enact comprehensive state-level human sex-trafficking legislation. Journal of Law and Health, 25(2), 381–417. http://engagedscholarship.csuohio.edu/jlh/vol25/iss2/7/ Roe-Sepowitz, D. E. (2012). Juvenile entry into prostitution: The role of emotional abuse. Violence Against Women, 18, 562–579. http://dx.doi.org/10.1177/1077801212453140 Saewyc, E. M., MacKay, L .J., Anderson, J., & Drozda, C. (2008, May). It’s not what you think: Sexually exploited youth in British Columbia. Vancouver, BC: University of British Columbia School of Nursing. Retrieved from: http://www.nursing.ubc.ca/PDFs/ItsNotWhatYouThink.pdf Sampson, R. J., & Laub, J. H. (1993). Crime in the making: Pathways and turning points through life. Cambridge, MA: Harvard University Press. Sampson, R. J., & Laub, J. H. (2005). A life-course view of the development of crime. The Annals of the American Academy of Political and Social Science, 602, 12–45. Schonlau, M., & Liebau, E. (2010). Respondent-driven sampling. Stata Journal, 12, 72–93. Shih, M., & Sanchez, D. T. (2005). Perspectives and research on the positive and negative implications of having multiple racial identities. Psychological Bulletin, 131, 569–591. http://dx.doi.org/10.1037/0033-2909.131.4.569 StataCorp. (2011). Stata 12 base reference manual. College Station, TX: Stata Press. Spangenberg, M. (2001). Prostituted youth in New York City: An overview (NCJ 198146). End Child Prostitution, Child Pornography and Trafficking of Children for Sexual Purposes (ECPAT-USA. Retrieved from https://www.ncjrs.gov/App/publications/Abstract.aspx?id=198146 Twill, S. E., Green, D. M., & Traylor, A. (2010). A descriptive study on sexually exploited children in residential treatment. Child & Youth Care Forum, 39, 187–199. http://dx.doi.org/10.1007/s10566-010-9098-2 Tyler, K. A. (2009). Risk factors for trading sex among homeless young adults. Archives of Sexual Behavior, 38, 290–297. http://dx.doi.org/10.1007/s10508-007-9201-4 U. S. Census Bureau. (n.d.). Poverty thresholds. Retrieved from https://www.census.gov/hhes/www/poverty/data/threshld/ U.S. Department of Justice. (2004). Assessment of U.S. Government activities to combat trafficking in persons. Retrieved from http://www.justice.gov/archive/ag/annualreports/tr2004/us_assessment_2004.pdf U.S. Department of State. (2007). Trafficking in persons report2007. Retrieved from www.state.gov/documents/organization/82902.pdf U.S. Department of State. (2010). Trafficking in persons report2010. Retrieved from http://www.state.gov/j/tip/rls/tiprpt/2010/ Vittinghoof, E., & McCulloch, C. E. (2007). Relaxing the rule of ten events per variable in logistic and Cox regression. American Journal of Epidemiology, 165, 710–718. http://dx.doi.org/10.1093/aje/kwk052 Victims of Trafficking and Violence Protection Act of 2000, 22U. S.C., § 7101(Pub. L. 106–386). (2000). Walker, K. E. (2002). Exploitation of children and young people through prostitution. Journal of Child Health Care, 6, 182–188. http://dx.doi.org/10.1177/136749350200600304 Whitbeck, L. B., Hoyt, D. R., Johnson, K. D., Berdahl, T. A., & Whiteford, S. W. (2002). Midwest longitudinal study of homeless adolescents. Baseline report for all participating agencies. Lincoln, NE: University of Nebraska, Department of Sociology. Williamson, C., & Cluse-Tolar, T. (2002). Pimp-controlled prostitution: Still and integral part of street life. Violence Against Women, 8, 1074. http://dx.doi.org/10.1177/107780102401101746 Williamson, C., Karandikar, S., Barrows, J., Smouse, T., Kelly, G., Swartz, P., … Gezinski, L. (2010). Ohio Trafficking in Persons Study Commission Research and Analysis Subcommittee report on the prevalence of human trafficking to Attorney General Richard Cordray. (2010). Retrieved from www.centralohiorescueandrestore.org/uploads/Ohio_Trafficking_in_Persons_Research_Sub-Committee_Report_2010_2-8_Final.pdf Williamson, C., Perdue, T., Belton, L., & Burns, O. (2012). Domestic sex trafficking in Ohio. Retrieved from http://www.ohioattorneygeneral.gov/getattachment/1bc0e815-71b6-43f5-ba45-c667840d4a93/2012-Domestic-Sex-Trafficking-in-Ohio-Report.aspx Williamson, C., & Prior, M. (2009). Domestic minor sex trafficking: A network of underground players in the Midwest. Journal of Child & Adolescent Trauma, 2(1), 46–61. http://dx.doi.org/10.1080/19361520802702191 Willis, B., & Levy, B. (2002). Child prostitution: global health burden, research needs, and interventions. Lancet, 359(9315), 1417–1422. http://dx.doi.org/10.1016/S0140-6736(02)08355-1 Zimmerman, C., Yun, K., Shvab, I., Watts, C., Trappolin, L., Treppete, M., … Regan, L. (2003). The health risks and consequences of trafficking in women and adolescents: Findings from a European study. London School of Hygiene & Tropical Medicine. Retrieved from http://www.lshtm.ac.uk/php/ghd/docs/traffickingfinal.pdf
