Despite a growing literature on commercial sexual exploitation and sex trafficking of minors in the United States, reliable estimates elude the field. When estimates are presented, they are generally accompanied by qualifiers and caveats. For example, Stransky and Finkelhor (2008) reviewed extant estimates of the number of juvenile prostitutes in the United States and found that the estimates ranged from 1,400 to 2.4 million. Indeed, following the recitation of these estimates, Stransky and Finkelhor implore readers, “PLEASE DO NOT CITE THESE NUMBERS [emphasis in original]” (Stransky and Finkelhor, 2008, p. 1). This chapter explains why such a caveat may be appropriate for all extant estimates of commercial sexual exploitation and sex trafficking of minors in the United States and why obtaining such estimates is so challenging. The chapter describes the current state of research focused on estimating commercial sexual exploitation and sex trafficking of minors in the United States. It includes examples of data collection efforts at the local, state, and federal levels and their relative strengths and limitations. It describes the difficulties inherent in counting crime in general and in counting commercial sexual exploitation and sex trafficking of minors in particular. The information presented illuminates why the estimates available today do not offer a full accounting of the incidence or prevalence of commercial sexual exploitation and sex trafficking of minors in the United States and why a full accounting of these problems may never be accomplished. The committee’s review offers insights into the complexities of different research strategies, and also underscores the challenges posed by current estimates for policy makers, practitioners, advocates, and researchers. At the same time, the committee explains that, while
perfect estimates of commercial sexual exploitation and sex trafficking of minors in the United States may not be attainable, improving the estimates is a worthy and attainable goal.
EXISTING ESTIMATES OF COMMERCIAL SEXUAL
EXPLOITATION AND SEX TRAFFICKING OF
MINORS IN THE UNITED STATES
Scholarly interest in commercial sexual exploitation and sex trafficking of minors has been increasing, especially since the Trafficking Victims Protection Act was passed in 2000.1 Still, this area of research is underdeveloped and uneven. Estimates of the incidence and prevalence of commercial sexual exploitation and sex trafficking of minors in the United States are particularly scarce. Further, there is little to no consensus on the value of existing estimates. This lack of consensus is not unusual and indeed is the case for estimates of other crimes as well (e.g., rape and intimate partner violence). As Best (2012) notes, no estimate is perfect, but some are more perfect than others. How perfect are extant estimates of commercial sexual exploitation and sex trafficking of minors in the United States?
Incidence and Prevalence
One of the most widely cited estimates of the commercial sexual exploitation of children comes from the research of Estes and Weiner (2002). That research, using data gathered from January 1, 1999, to March 31, 2001, indicates that between 244,000 and 325,000 children are at risk of commercial sexual exploitation in the United States, in addition to the estimated “105,000 children who are substantiated or indicated to be victims of other types of child sexual abuse annually in the United States” (Estes and Weiner, 2002, p. 46). Estes and Weiner’s 27-month study gathered data from a variety of sources in 17 U.S. cities, as well as cities in Canada and Mexico. These sources included, but were not limited to, interviews with key stakeholders, commercial sexual exploitation customers, law enforcement representatives, and human service representatives. In addition, the authors gathered information from 124 sexually exploited runaway and thrown-away street children and 86 sexually exploited children under the care of human service and law enforcement agencies.
The limitations of Estes and Weiner’s (2002) estimates are well documented (see, e.g., Stransky and Finkelhor, 2008) and are clearly acknowledged by the authors. First, these estimates focus on youth “at risk” for
1Victims of Trafficking and Violence Protection Act of 2000, Public Law 106-386, Division A, 114 Stat. 1464.
commercial sexual exploitation as opposed to actual victims; a distinction between actual victims and those at risk for exploitation is not made. A second issue is that the at-risk numbers do not take into account individuals who may belong in multiple risk categories (Stransky and Finkelhor, 2008). For instance, if a child were homeless, lesbian, gay, bisexual, transgender, or questioning, and the victim of child sexual abuse, he/she would belong in three of the risk categories and might have been counted three (or more) times in the estimates. While Estes and Weiner (2002) describe how their research corrects for possible duplicate counts in some categories, the extent of duplicate counting, if any, is unknown. The authors acknowledge that with greater resources, a different methodological approach might be taken that could yield improved estimates. Specifically, they note that “a different type of study from ours—one that used a different methodology and a higher investment of resources—is needed to carry out a national prevalence and incidence survey that could produce an actual headcount of the number of identifiable commercially sexually exploited children in the United States” (Estes and Weiner, 2002, p. 143). Even with these limitations, however, Estes and Weiner’s estimates are the most widely cited national-level estimates of commercial sexual exploitation of minors.
Subsets and Subpopulations of Victims
In an effort to understand the nature and extent of commercial sexual exploitation and sex trafficking of minors in the United States, some have examined specific subpopulations or subsets of victims. For example, Greene and colleagues (1999) focus on the prevalence of survival sex among a shelter sample and a sample of street youth. The shelter youth participants were selected using a multistage sample methodology to derive nationally representative data on youth in shelters. The street youth participants were selected as a convenience sample from 10 U.S. cities. Aggregated, these data, gathered in November and December 1992 and focusing on youth aged 12 to 21, revealed that 9.5 percent of the shelter youth and 27.5 percent of the street youth had engaged in survival sex during that year (Greene et al., 1999). While this work did not yield an estimate of the prevalence of minors who are victims or survivors of commercial sexual exploitation and sex trafficking, it demonstrates that shelter and street youth may be at high risk for being subject to these crimes given their involvement in survival sex. Nonetheless, a limitation of this research for purposes of understanding the extent of commercial sexual exploitation and sex trafficking of minors is that the samples included only youth aged 12 to 21. Unfortunately, the study does not disaggregate minors (under age 18) from others in the samples, and it does not offer information on minors under age 12. This
limitation is not uncommon as much research in this area combines a range of ages encompassing children, adolescents, and young adults.
Yet another approach used to understand the extent of commercial sexual exploitation and sex trafficking of minors in the United States is to examine arrest records. Using 2009 data from the Federal Bureau of Investigation’s (FBI’s) Uniform Crime Reporting (UCR) program, Puzzanchera and Adams (2011) estimate that 1,400 youth were arrested for prostitution and commercialized vice in the United States during that year. In this work, the authors define juveniles as persons younger than age 182 and detail the characteristics of those juveniles arrested. Fully 78 percent of the 1,400 arrests involved a female, 12 percent involved a juvenile under age 15, and 40 percent involved a white juvenile (Puzzanchera and Adams, 2011). Arrest statistics, however, while informative, are no substitute for prevalence rates. One cannot know, for example, whether the 1,400 arrests were of unique individuals or included individuals who were arrested on multiple occasions for the same offense. Furthermore, the number of arrests is not equivalent to the number of incidents of commercial sexual exploitation and sex trafficking of minors; one incident may have involved multiple juveniles who were arrested for prostitution and commercialized vice. Further, this estimate includes only arrests for prostitution of minors and does not offer insight into the number of juveniles engaging in other related activities. And finally, arrest numbers cannot convey the number of minors being prostituted who escaped detection from police; were arrested for a different offense (e.g., drug related); or were handled in ways not involving arrest at the discretion of the officer or other criminal justice official, including those who were recognized as victims and treated accordingly. Therefore, the use of prostitution arrest numbers offers a highly conservative and incomplete estimate of commercial sexual exploitation and sex trafficking of minors in the United States.
Another approach used to estimate commercial sexual exploitation and sex trafficking of minors in the United States is to quantify known victims by groups such as service providers, health care providers, school personnel, and law enforcement, among others. An example is estimates derived from
2The authors note that their definition of juvenile (i.e., under age 18) conflicts with the legal definition of the term in 13 states. In 10 states, juveniles are defined as persons aged 16 and younger and in 3 states as those under age 16.
the FBI’s Innocence Lost National Initiative, indicating that the program has led to the “recovery”3 of more than 2,100 children being prostituted since its inception in 2003 (FBI, 2012). This number, while indicative of the FBI’s success in helping victims of commercial sexual exploitation and sex trafficking of minors, must be viewed as a highly conservative estimate as well. First, the number of children recovered from prostitution includes only those victims of commercial sexual exploitation and sex trafficking who come to the attention of law enforcement. Further, as with arrest records, one cannot know whether this estimate includes some minors who have been recovered multiple times. And finally, the estimate is limited to those minors who have been prostituted; it does not include minors involved in survival sex or other forms of sexual exploitation.
Human Trafficking Task Force Data
The Bureau of Justice Assistance requires its human trafficking task force grantees to collect data and enter them into the Human Trafficking Reporting System. This system was developed and is maintained by researchers at Northeastern University and is designed to track the performance of federally funded human trafficking task forces. It is currently the only source of information from state and local law enforcement agencies involved in human trafficking investigations. Investigations are guided by the definition of human trafficking found in the Trafficking Victims Protection Act of 2000.
The data reported to the Human Trafficking Reporting System include incidents of labor and sex trafficking of minors and adults investigated by Bureau of Justice Assistance human trafficking task forces. Data in this system have been examined by analysts at the Bureau of Justice Statistics. That research indicates that from January 2008 through June 2010, federally funded human trafficking task forces identified 1,016 alleged incidents of prostitution or sexual exploitation of children (Banks and Kyckelhahn, 2011). While this number encompasses both prostitution and sexual exploitation of children, it is limited. First, it reflects only those cases that came to the attention of the 45 federally funded task forces, which are not geographically representative of the nation (Banks and Kyckelhahn, 2011). Second, data were entered into the system only if an investigator spent more than 1 hour on the alleged case. Third, the data represent alleged cases only and thus include cases that may subsequently have been determined to be unfounded. Fourth, an analysis of the data indicated that data entry was incomplete as a result of variation in reporting by individual task forces
3The FBI offers no definition of “recovered” children.
(Banks and Kyckelhahn, 2011).4 While these data offer some insight into the commercial sexual exploitation and sex trafficking of minors in the United States, then, they are imperfect and incomplete.
A number of states have conducted prevalence studies of commercial sexual exploitation and sex trafficking of minors within their borders (Gragg et al., 2007; Ohio Trafficking in Persons Study Commission Research and Analysis Sub-Committee, 2012; Quin et al., 2011). These studies were conducted at the request of different offices and agencies (e.g., the state attorney general, state legislatures). The studies used different research methods and different definitions and produced—understandably—notably different findings.
In New York, a study that included as one of its purposes estimating the prevalence of commercial sexual exploitation was conducted for the state’s Office of Children and Family Services at the request of the New York State Legislature. This study defined sexually exploited children as those younger than 18 who “have engaged or agreed or offered to engage in sexual conduct with another person in return for a fee, traded sex for food, clothing or a place to stay, stripped, been filmed or photographed performing or engaging in sexual acts, or loitered for the purpose of engaging in a prostitution offense” (Gragg et al., 2007, p. i). The authors gathered data for July 15 to September 15, 2006, from a broad range of stakeholders and agencies in a nonprobability sample of upstate counties and New York City boroughs. The study used a multimethods approach that included multiple mail surveys, qualitative interviews, and focus groups. The mail surveys were sent to 159 entities (e.g., law enforcement and probation agencies, detention facilities, child advocacy centers, shelters, rape crisis centers) in areas known to have high rates of prostitution arrests and reports of child sexual abuse. Qualitative interviews were conducted with numerous stakeholders. Focus groups involved survivors/victims of commercial sexual exploitation in New York City. Survivors were asked about their “perception of their home environment when they entered ‘the life’ and (in a few cases) when they returned to it; how they became involved in sexual exploitation; their experience on the street, particularly regarding violence and involvement with exploiters and solicitors; leaving the life; and supports needed to succeed” (Gragg et al., 2007, p. 43). The researchers also used arrest records and information from Office of Children and Family Services intake records to provide an annual
prevalence estimate of commercial sexual exploitation. Prevalence data were gathered using mail surveys sent to children known to be survivors of commercial sexual exploitation.
In total, this study identified 2,652 victims of commercial sexual exploitation on an annual basis in New York State (Gragg et al., 2007). The authors, however, note limitations associated with this research, including that the estimates are based on identified victims only and cannot account for those who have been victimized but not identified. In addition, the research focuses on a relatively short time frame (2 months in 2006). Also, two large law enforcement agencies failed to participate, and the authors believe that “participation from these agencies would have increased the prevalence counts” (Gragg et al., 2007, p. 16). Germane to present purposes, this research has additional limitations. First, it utilizes a broader definition of commercial sexual exploitation than that used by the committee, so its estimates will be higher than those derived from work using a narrower definition. Second, the findings pertain to parts of New York State only; they cannot be generalized statewide (given the methodology used), and they cannot provide nation-level information.
While some research offers estimates based on state- and city-level data (Curtis et al., 2008), at least one example of research focused on commercial sexual exploitation uses a nationally representative sample. In an examination of the national incidence and utilization of social networking-facilitated commercial sexual exploitation, Mitchell and colleagues (2010) estimate that law enforcement made between 440 and 669 arrests for this crime in 2006 in the United States. That this estimate is so low is not surprising given the study’s narrow focus on Internet-facilitated commercial sexual exploitation and resulting arrests. One cannot know from this work how many Internet-facilitated commercial sexual exploitation incidents did not come to the attention of law enforcement or the extent to which those that did come to attention resulted in an arrest.
ISSUES IN DEVELOPING ESTIMATES OF COMMERCIAL
SEXUAL EXPLOITATION AND SEX TRAFFICKING
OF MINORS IN THE UNITED STATES
Efforts to measure commercial sexual exploitation and sex trafficking of minors in the United States confront virtually every challenge associated with measuring crime. This section identifies elements of “good” estimates and points out challenges associated with generating “good” crime estimates.
Best (2012) identifies five necessary elements of a valid estimate. First,
the estimate must be based on more than guessing. As Best notes, sometimes a guesstimate or the inclusion of some guessing is the best option available for a social problem. Nonetheless, while a good guess may be a starting point, it rarely produces a good or valid estimate. Second, the estimate must be based on a clear and reasonable definition. As discussed in this report, definitions of commercial sexual exploitation and sex trafficking of minors in the United States are inconsistent (see Chapter 1 and Appendix A). Without a clear and reasonable definition, what exactly is to be measured is unclear. Without clarity, measurement is muddled, and this directly influences the quality of the resulting estimate. Third, the estimate must come from clear and reasonable measurement. That is, the measurement method (e.g., survey questions, interviews with service providers, observations) must gather information based on the clearly outlined definition. Fourth, a good estimate is one based on data from a quality sample. Finally, a good estimate is more than a number; in addition to the number, information must be provided about the definition, measurement, and sample used to generate this estimate so the reader can understand the estimate’s limitations. A surprising number of estimates related to commercial sexual exploitation and sex trafficking of minors in the United States are accompanied by no information on the methods used to generate them (Stransky and Finkelhor, 2008). Even with these five elements present, there may be disagreement regarding the quality of an estimate; using these basic guidelines, however, interested parties can work toward producing improved estimates.
The following sections outline the measurement and sampling issues and other methodological challenges associated with measuring crime. These issues directly affect efforts to measure commercial sexual exploitation and sex trafficking of minors in the United States. Understanding these challenges provides insight into the difficulties involved in estimation of such crimes and points to strategies for addressing them.
Definitions drive measurement. As discussed above, without a clear and reasonable definition, clear and reasonable measurement (and estimation) of a crime is impossible. A clear definition makes clear the elements of the crime. For example, stalking is defined in part as being something that is “repeated.” Thus during measurement, one ascertains whether the behavior in question occurred two or more times. A reasonable definition is one that makes sense. For example, defining sex trafficking of minors as “children transported without their parents’ presence” is not reasonable because it means that when a child rides a school bus, is picked up by a nanny for an after-school event, or is taken by a friend or nonparent relative to a birthday party, sex trafficking has occurred. A reasonable definition
is not overly broad or unnecessarily narrow. While it may not be possible to achieve total agreement on what a reasonable definition is, a definition should be understood by most interested parties as reasonable.
Efforts to define commercial sexual exploitation and sex trafficking of minors in the United States are characterized by disagreement. For example, agreement is lacking on whether a third-party exploiter is necessary for a situation to be considered commercial sexual exploitation or sex trafficking. (See Chapter 1 for the committee’s discussion of key terms and additional definitions.)
The lack of consensus on clear definitions of commercial sexual exploitation and sex trafficking of minors in the United States contributes directly to the difficulty of measuring and estimating the extent and nature of the problems. In addition to technical difficulties, the lack of consensus creates measurement challenges because it fosters a lack of understanding among a variety of actors that has real consequences. For example, if commercial sexual exploitation and sex trafficking of minors were clearly defined and if those definitions were widely accepted and publicized in a way that could reach minors, a victim might be better able to self-identify. At present, victims may instead view themselves as criminal offenders and fail to cooperate with entities seeking to assist them. Similarly, without clear and widely accepted definitions of commercial sexual exploitation and sex trafficking of minors, individuals in a position to provide assistance to victims (e.g., law enforcement agents, health care providers, teachers, social workers, foster parents) may not do so because they may see these minors as offenders, promiscuous adolescents, or difficult juveniles rather than the victims of a crime. Until agreement is reached on clear and reasonable definitions of both commercial sexual exploitation and sex trafficking of minors, measurement will remain difficult, and, as a result, it will not be possible to derive valid estimates.
Measuring crime is not easy, and in some cases, such as commercial sexual exploitation and sex trafficking of minors in the United States, it is very difficult. This section describes discusses several measurement issues associated with measuring crime.
Unit of Analysis
Before any measurement or the calculation of any estimate is conducted, one first must determine who or what is being studied, or the “unit of analysis.” Numerous units of analysis may be used in measuring crime. One may be interested in understanding the incident (which may have
multiple victims and even more victimizations), the victim (the people offended against), or the victimization (the crime committed during an incident). Different national data sources utilize different units of analysis. For example, the unit of analysis that forms the basis of official FBI estimates depends on the crime of interest: robbery and burglary counts focus on incidents, whereas rape counts focus on victimizations. In contrast, official estimates from the National Crime Victimization Survey (NCVS) focus on victimizations.
Lack of clarity about the unit of analysis can contribute to misreporting and distortion of research findings and estimates. In 2010, for example, the Bureau of Justice Statistics estimated that there were 15 violent victimizations per 1,000 people in the United States. It would not be surprising to find this statistic repeated as 15 victims of violent victimizations per 1,000 people in the United States. Though similar, these statements convey two very different findings: the first counts the number of victimizations, while the second counts the number of victims. It is theoretically possible, as conveyed by the latter finding, that each victim experienced one violent victimization during the year, making the two statements equivalent. This is unlikely, however, given the well-documented phenomenon whereby individuals who are victimized tend to be victimized repeatedly over time (Dodge and Balog, 1987; Pease, 1998). Attention to the unit of analysis is therefore critical.
In terms of commercial sexual exploitation and sex trafficking of minors in the United States, it is important to determine what one wishes to estimate (e.g., current victims, individuals at heightened risk of being commercially exploited and trafficked, victimizations, former victims, incidents, exploiters, abettors). And when reading published estimates, it is important to note what exactly has been estimated.
The Hierarchy Rule
Determining how to count crime is complex even when the unit of analysis is clearly identified. Imagine that an offender walks into a crowd of 20 people; takes everyone’s wallet; and hits one individual with a pipe on the head, leaving a gash that requires hospitalization. How many crimes were committed? There are multiple ways to count the number of crimes associated with this event, one of which involves the “hierarchy rule.” The hierarchy rule counts only the most serious type of victimization committed during an incident. Focusing on violence only, the UCR program5 ranks
5The newer-generation National Incident-Based Reporting System (NIBRS) uses the hierarchy rule in few cases, so it would count all of these crimes in this single incident. The NCVS could be used to count all of these crimes as well. Rarely, however, do researchers use the
homicide as the most serious type of crime, followed by rape, robbery, and aggravated assault. The NCVS ranks rape/sexual assault as the most serious type of violent crime, followed by robbery, aggravated assault, and simple assault.6 Returning to the above example, both the UCR and the NCVS would count the robberies as the most serious victimization in this incident; the aggravated assault (i.e., being hit on the head) would not be counted. Given the differences in the unit of analysis used by each data collection effort, the UCR would record a single robbery incident, while the NCVS would record 20 robbery victimizations.
Many crimes (e.g., burglary and robbery) are discrete events with a clear beginning and end. Others (e.g., intimate partner violence and bullying) are continuous in nature and have no clear beginning or end. The latter are referred to as series victimizations, and they pose additional challenges to counting crime. How one handles series victimizations has an enormous impact on resulting estimates. Commercial sexual exploitation and sex trafficking of minors appear to be a form of series victimization, which further complicates attempts to estimate these crimes in the United States.7
There is no agreement on how series victimizations should be counted. One possibility is to count such a victimization as a single (ongoing) crime, carrying the same weight in crime estimates as a single 1-minute discrete crime. Another option is to count a series victimization according to the number of days on which it has occurred. Thus, a boy bullied every day for a year would be included in official statistics as 365 victimizations. Some argue that this approach fails to capture the essence of the violence and artificially inflates resulting estimates. A third option is to ask the victim (if possible) how many times he/she thinks the violence occurred and include that number in the estimate. This approach leads to rounding and imprecision in estimates. Thus, series victimizations can be handled in a number of ways, none of which is perfect.
Asking Questions to Gather the Data
After determining the unit of analysis and making decisions about the hierarchy rule and series victimizations, investigators must ascertain how
data without the hierarchy rule. For one exception, in which researchers examined rape and its co-occurring crimes, see Addington and Rennison (2008).
6Because the NCVS interviews victims, it does not collect data on homicide.
7For additional information on approaches to counting series victimization and how the choice of approach influences estimates, see Lauritsen et al. (2012) and Rand and Rennison (2005).
to measure the phenomenon at hand. One might observe individuals in public to measure the presence of commercial sexual exploitation and sex trafficking of minors in the United States. One might choose to count ads on Backpage.com. Or one might interview service providers. While identifying the many possibilities is beyond the scope of this study, it is valuable to point out useful approaches for asking questions to elicit data with which to estimate commercial sexual exploitation and sex trafficking of minors in the United States.
Returning to the point made earlier that definitions guide measurement, a successful practice is to ask the respondent about elements of the definition of commercial sexual exploitation and sex trafficking of minors.8 An understanding of those elements can guide the types of questions used to measure the problems. Extant research indicates that asking victims whether they were subjected to commercial sexual exploitation or sex trafficking (or some other crime or event) is a poor substitute. An example illustrates this point.
The designers of the National Crime Survey, the predecessor of the NCVS, devised a methodology focused on asking about elements of an event. In the NCVS, for example, an “aggravated assault is defined as an attack or attempted attack with a weapon, regardless of whether or not an injury occurred and attack without a weapon when serious injury results” (Child Trends, 2012, p. 7). This definition indicates that when an offender brandishes a weapon, an aggravated assault has occurred. In addition, the definition makes clear that if there was no weapon but the victim sustained a serious injury (broken bones, lost teeth, loss of consciousness, stab wounds, gunshot wounds, internal injuries, or any unspecified injury requiring 2 or more days in the hospital), an aggravated assault took place (Bureau of Justice Statistics, 2013). This definition indicates exactly what needs to be asked of respondents during an interview. They are asked whether an offender threatened them or attacked them with any weapon. They are asked if they sustained any of the following injuries: broken bones, lost teeth, loss of consciousness, stab wounds, gunshot wounds, internal injuries, or any unspecified injury requiring 2 or more days in the hospital. If respondents answer affirmatively, an aggravated assault is recorded. An option is to ask respondents whether they were a victim of aggravated assault. This approach is problematic as individuals have different understandings of what an aggravated assault is. If respondents have no idea or differing ideas of what an aggravated assault is, the data gathered will be of little value. Asking clear questions about elements of a crime avoids this problem.
In addition, asking respondents about elements of a crime minimizes
8See Fisher et al. (2010) on the importance of screening questions in estimation.
misreporting by victims who may not view themselves as such or do not understand what constitutes a particular victimization. If questions used to measure a particular type of crime are carefully constructed, respondents are more likely to reveal information indicating that they are a victim of that crime without experiencing fear of retaliation or shame—a significant concern for crimes such as commercial sexual exploitation and sex trafficking. Such careful measurement, based on clear and reasonable definitions, can lead to more accurate revelations and as a result, better estimates.
A crucial aspect of generating valid statistics such as estimates is the sample used. Even if researchers have a clear and reasonable definition, well-designed methods for identifying victims (or victimizations or incidents), well-trained interviewers, and access to adequate resources, the use of a poor sample will produce poor data and unreliable estimates.
One of the major goals of generating an estimate of a crime such as commercial sexual exploitation or sex trafficking of minors is understanding how much of this activity (or how many victims or incidents or victimizations) occurs over a particular time period. To best accomplish this goal, researchers gather a sample they hope will inform them about the larger population of interest. Because time and financial constraints make obtaining information from every member of the population of interest unlikely, a sample is drawn, and the information gathered from that sample is used to draw inferences about the larger population. Sampling is simply the process of selecting observations from that larger population. There are two major forms of sampling—probability and nonprobability sampling.
Data collected using probability sampling allow one to generalize findings back to the population of interest. Generating a probability sample requires a sampling frame, which is a list of all elements in the population of interest. Thus to generate a probability sample for estimating the number of minors involved in commercial sexual exploitation and sex trafficking of minors in the United States, one would need a list of all minors in the United States. Using that list and one of numerous probability sampling approaches, a sample would be selected. Each element in the sample would have to have a known nonzero chance of being drawn into the sample. Then, all of the selected sample elements would be approached and interviewed (e.g., by phone, in person, by computer, or by mail) about their experiences to ascertain whether they are at risk of or are or were victims of commercial sexual exploitation or sex trafficking.
While a probability sampling approach offers many advantages, it has several limitations. First, it can be extremely expensive (depending on the population of interest). Drawing a probability sample of all minors in the United States—assuming that a sampling frame of all minors exists—would be extraordinarily expensive. For instance, the NCVS surveys a nationally representative sample of households in the United States. Fielding this survey on crime victimization costs approximately $30 million annually. The survey is only this “affordable” because it does not include individuals under age 12, homeless people, and persons in institutions, whose inclusion would greatly increase the cost and complexity of the sampling.
Even if cost were of no concern and a sampling frame of minors in the United States were available, this approach would pose the difficulty of locating and interviewing all the minors drawn into the sample. This would include finding those who had run away, who had been asked to leave home, and who were hiding or being hidden. Further, because the sample would comprise minors, the need to obtain parental permission to participate might be an issue under current federal research regulations.9
A second option is the use of nonprobability sampling. Nonprobability sampling entails gathering a subset of the population of interest; however, there is no known probability associated with the selection of each element into the sample. There are many ways to draw a nonprobability sample.10 One can find a person with the characteristic of interest (e.g., a minor known to have been commercially sexually exploited or trafficked for sex) and ask him/her to direct the researcher to other minors with that characteristic. Or one can go to an area where one believes these individuals congregate and gather information there. Or one can select respondents because they are convenient and available. These sorts of approaches are used widely and are responsible for an abundance of valuable information on a wide range of actors, such as burglars, runaways, robbers, drug dealers, and drug users.
Like all aspects of research, nonprobability sampling has its limitations. One limitation is that findings based on a nonprobability sample should not be generalized to the larger population of interest. Thus, even if one identified 10,000 or 100,000 commercially sexually exploited or trafficked minors using a nonprobability sampling approach, one could not be sure that the findings accurately described the population of all minors. Even if
945 C.F.R. 46.
10Again, a review of these methods is beyond the scope of this report; a multitude of books on the subject are available.
a researcher believed that his/her nonprobability sample was representative of the population, this could not be known for sure, limiting the value of findings derived from the research.
Still, nonprobability sampling can be useful, especially for issues on which little is known. And methods are available for enhancing the usefulness of nonprobability samples, which could lead to better estimates of commercial sexual exploitation and sex trafficking of minors in the United States. One such approach is the capture-recapture method. In general, this approach has two stages. As applied to human subjects, the first, the “capture stage,” entails gathering information from as many subjects as possible in a given area. The second stage involves attempting to “recapture” as many of the same subjects as possible for follow-up interviews, examinations, and so on. The data from the nonprobability sample are adjusted to derive a more precise estimate of the total population of interest, using judgments regarding how many of the same subjects were “recaptured” relative to the initial “capture.”
Another approach used to adjust to address issues encountered with traditional nonprobability samples is respondent-driven sampling (Gozdziak, 2008; Heckathorn, 1997; Tyldum and Brunovskis, 2005). With this approach, modeling techniques are used to weight data from a nonprobability sample to arrive at more precise estimates. It is argued that this approach minimizes some of the problems associated with traditional nonprobability sampling (Tyldum and Brunovskis, 2005). For example, Curtis and colleagues (2008) used this approach to conduct research on commercial sexual exploitation and sex trafficking of minors in New York City. The authors gathered their sample by beginning with initial research subjects (i.e., “seeds”), each of whom was paid $20 and given three coupons to pass to others who participated in the commercial sexual exploitation market.
Lastly, representative sampling can be used to attempt to extrapolate information from a small sample to a larger population. Study results can then be generalized to the larger population. For example, Svedin and Priebe (2007) used this approach to estimate the number of Swedish adolescents who were selling sex as a form of payment, in addition to identifying demographic and psychosocial variables associated with selling sex. The authors ensured that a proportional sample was obtained by applying a sampling procedure design.
While these and similar approaches may represent improvements over traditional nonprobability sampling, they are not substitutes for probability sampling, and the generalizability of the findings of such research remains questionable. Still, these approaches may offer the best opportunity for generating estimates of commercial sexual exploitation and sex trafficking of minors in the United States.
Additional Methodological Challenges
As is evident from the committee’s review of the commercial sexual exploitation and sex trafficking literature, as well as knowledge gained from general victimization research, challenges beyond definition, measurement, and sampling are encountered in measuring crime. One such challenge is dealing with what is known as the “dark figure of crime” (Biderman and Reiss, 1967), a phrase referring to the amount of unreported or undiscovered crime. The term was developed in response to the limitation of crime estimates based on official police reports that much crime is not discovered, and much that is discovered is not reported to the police. Researchers cannot say for certain how much crime this represents, just that it exists. The dark figure has many contributors, including individuals’ not reporting to the police, police officers’ not documenting crimes, crimes that are committed and never discovered, bodies that are never found, and failure to recognize that a crime has occurred.
Prior to 1972, the United States’ only national estimates of crime were calculated using data collected directly from law enforcement agencies. Researchers and policy makers understood that these estimates omitted crime that was not reported to the police, raising the question of just how much crime was being missed. This was a primary—if not the primary—reason that the NCVS was created. The fielding of the NCVS has demonstrated that the dark figure can be very large. For example, about 50 percent of all violent crime is not reported to the police or reflected in police estimates of crime (Truman, 2011). This rate of nonreporting is even higher for some crimes, such as rape (Hart and Rennison, 2003; Rennison, 2002).
Estimating commercial sexual exploitation and sex trafficking of minors is complicated by the dark figure of crime. Perhaps the greatest challenge of estimating these crimes is their hidden nature. Finkelhor and Ormrod (2004b) found, for example, that juvenile prostitution was more likely than adult prostitution to occur indoors, where it is less likely to be identified by police. Therefore, gaining a better understanding of the dark figure can be especially difficult for commercial sexual exploitation and sex trafficking of minors. Not only is commercial sexual exploitation and sex trafficking of minors generally hidden, but it affects hidden populations—populations whose size and boundaries are unknown. Because hidden populations have no sampling frames, the generation of estimates involving these populations is further complicated.
Uncovering the dark figure of commercial sexual exploitation and sex trafficking of minors in the United States is made even more difficult because their victims, even when located, may not cooperate with officials, service providers, and others who can assist them (Bales and Lize, 2007; Moossy, 2009). They may provide false information about themselves,
their ages, and their activities (Moossy, 2009). They themselves often fail to recognize that they are victims (Bales and Lize, 2007; Moossy, 2009). Their lack of cooperation can be due to a myriad of reasons, such as fear of retaliation, abuse, or force; language barriers; coercion; and fear of arrest, deportation, or entrance into the criminal justice system as an offender (Moossy, 2009; Rieger, 2007).
The hidden nature of commercial sexual exploitation and sex trafficking of minors is an issue with which the national crime data collection systems cannot fully contend. The UCR and the NCVS measure crimes selected because they are deemed to be serious, common, and relatively easy to measure. Commercial sexual exploitation and sex trafficking of minors in the United States, while serious, are not easy to measure. Thus both of these data collection systems as currently implemented are limited in their ability to offer estimates of the commercial sexual exploitation and sex trafficking of minors in the United States, and the degree to which these crimes occur remains largely unknown. More on these and other national crime measurement programs follows.
EXISTING CRIME MEASUREMENT PROGRAMS
This section describes the UCR program, the National Incident-Based Reporting System (NIBRS), the Supplemental Homicide Reports, and the NCVS, and explains how these programs can and cannot assist in providing estimates of commercial sexual exploitation and sex trafficking of minors in the United States.
The UCR program began in the 1920s and continues today, making it the nation’s oldest unified national crime data collection effort. Currently, the UCR program gathers crime reports from approximately 17,000 (of the more than 18,000) law enforcement agencies in the United States and its territories (FBI, 2013b). The purpose of the program was originally and continues to be to serve the needs of law enforcement agencies. Initially, this was done by gathering information on a broad range of nonfatal personal and property criminal offenses occurring to a person of any age or to any business. With the introduction of the NIBRS (see page 59), the UCR began being referred to as the Summary Reporting System (UCR/SRS) because it offers only counts for each category of crime; with very few exceptions, it does not gather information on characteristics of the crimes. This lack of detail and the use of the hierarchy rule limit what can be learned from the UCR/SRS data. For example, one could not determine the victim and of-
fender relationship in an assault; whether a weapon had been used during a rape; or a myriad of other characteristics of the event, victim, or offender.
Beginning in January 2013, the UCR/SRS started collecting data on human trafficking. In response to the William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008,11 the FBI is required to gather offense and arrest data related to human trafficking. A change of this magnitude is highly unusual in the UCR program. The last time a crime was added was in 1982, when arson became a permanent Part I crime in the system. The William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008 mandates that:
(a) The Director of the Federal Bureau of Investigation shall—(1) classify the offense of human trafficking as a Part I crime in the Uniform Crime Reports; (2) to the extent feasible, establish subcategories for State sex crimes that involve—(A) a person who is younger than 18 years of age; (B) the use of force, fraud, or coercion; or (C) neither of the elements described in subparagraphs (A) and (B); and (3) classify the offense of human trafficking as a Group A offense for the purpose of the National Incident-Based Reporting System.
(b) ADDITIONAL INFORMATION.—The Director of the Federal Bureau of Investigation shall revise the Uniform Crime Reporting System and the National Incident-Based Reporting System to distinguish between reports of—(1) incidents of assisting or promoting prostitution, which shall include crimes committed by persons who—(A) do not directly engage in commercial sex acts; and (B) direct, manage, or profit from such acts, such as State pimping and pandering crimes; (2) incidents of purchasing prostitution, which shall include crimes committed by persons who purchase or attempt to purchase or trade anything of value for commercial sex acts; and (3) incidents of prostitution, which shall include crimes committed by persons providing or attempting to provide commercial sex acts.
In accordance with the William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008, the UCR/SRS added two Part I crimes beginning on January 1, 2013: human trafficking/commercial sex acts and human trafficking/involuntary servitude. The newly included crime of human trafficking/commercial sex acts is defined as “inducing a person by force, fraud, or coercion to participate in commercial sex acts, or in which the person induced to perform such act(s) has not attained 18 years of age.”12 As a result, once enough data have accumulated, the UCR/SRS will offer estimates of the extent and nature of perpetrators who were arrested
11William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008, Public Law 110-457 237, 28 U.S.C. 534 note (2004).
12William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008, Public Law 110-457.
for engaging in these crimes. The act requires that data on age, sex, and race be gathered on persons arrested for engaging in (1) prostitution, (2) assisting or promoting prostitution, and (3) purchasing prostitution.
Supplemental Homicide Reports
The FBI has gathered Supplemental Homicide Reports annually since 1980. These incident-based reports include demographic information about the victim and offender in a homicide, as well as the victim and offender relationship, the type of weapon used, and the circumstances of the homicide (FBI, 2013a). Like all FBI crime data, these data are available at a variety of geographic levels. In addition to the changes in the UCR/SRS discussed above, the William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008 requires the inclusion of human trafficking as a circumstance code in the FBI’s Supplementary Homicide Reports.
Given the limitations of the UCR program and the success of the Supplemental Homicide Reports, calls were made for an enhanced data collection effort. The result was the creation of the NIBRS. Like the original UCR and the UCR/SRS, the NIBRS is a voluntary program and reflects only crimes known to the police. It augments the UCR/SRS by gathering detailed incident information about crimes, “including the nature and types of crimes in each incident, victim(s) and offender(s) characteristics, type and value of stolen and recovered property, and characteristics of arrested individuals” (Rennison, 2009, p. 385). There are a number of additional differences between the UCR/SRS and the NIBRS. The NIBRS measures crime using new definitions in some cases. Most notably, it redefines forcible rape13 to include male victims as well as incidents involving objects. The NIBRS also includes a new crime category—crimes against society—which comprises drug/narcotic offenses, pornography/obscene materials, prostitution, and gambling offenses (Rennison, 2009). Another important difference between the UCR/SRS and the NIBRS relates to the hierarchy rule. In the NIBRS, use of the hierarchy rule is seriously curtailed except for a limited number of crimes (e.g., motor vehicle theft, personal property theft, justifiable homicide). Still, the near elimination of the hierarchy rule in the NIBRS means that far more crime information is gathered in that system than in the UCR/SRS. The NIBRS and the UCR/SRS also differ in that the NIBRS distinguishes between an attempted and a completed crime. Unlike the UCR/SRS, the NIBRS allows one to link attributes of a crime.
13The UCR/SRS has since redefined rape in the same way.
For instance, “using the traditional system, with the exception of homicide, one could not link offender information, victim information, and incident information for a single incident. With NIBRS, for many crimes, one can link data on victims to offenders to offenses to arrestees” (Rennison, 2009, p. 386).
Like the UCR/SRS, the NIBRS is being changed in response to the William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008. Specifically, the NIBRS is now required to gather data on human trafficking/commercial sex acts and human trafficking/involuntary servitude, to be housed among Group A offenses. In addition, the NIBRS will provide data on incidents involving assisting or promoting prostitution, purchasing prostitution, and prostitution.14 While the NIBRS previously included categories for prostitution and assisting or promoting prostitution, purchasing prostitution is newly added. The modernization, enhancements, and improvements to the NIBRS produce data that better serve the needs of its primary constituency—law enforcement. In addition, the information gathered using the NIBRS has the potential to better inform the public and policy makers about victimization risk and related topics.
In 1965, the United States was experiencing high and rapidly increasing levels of violent and property crime. At that time, the only crime data available were those from the UCR program, and it was clear that these data were limited for answering pressing questions regarding crime in the United States. The consensus was that a new measure of crime in the United States was needed to compensate for the limitations of the UCR program. In July 1991, the National Crime Survey (later renamed the NCVS) was launched.
The NCVS is one of the nation’s primary sources of information about the frequency, characteristics, and consequences of victimization of individuals aged 12 and older and their households in the United States. NCVS crime data come from surveys administered by field representatives to a representative sample of households in the United States throughout the year, in person and over the phone (Rennison, 2009). Households are selected through a stratified, multistage, cluster sampling process (a type of probability sampling). The NCVS uses a rotating panel design in which people are interviewed every 6 months for a total of seven interviews. Households in the sample are designed to be representative of all households as well as noninstitutionalized individuals aged 12 or older in the United States (Rennison, 2009). The sample size is very large. In recent years, approxi-
14William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008, Public Law 110-457.
mately 70,000 persons in 40,000 households have been interviewed twice a year as part of the NCVS (Truman, 2011).
Using the UCR/SRS, Supplemental Homicide Reports, NIBRS,
and NCVS for Estimating Commercial Sexual Exploitation
and Sex Trafficking of Minors in the United States
While each of the data collection efforts described above performs well for the purposes for which it was designed, each is limited in its ability to offer estimates of commercial sexual exploitation and sex trafficking of minors in the United States.
Historically, the UCR/SRS could not inform estimates of commercial sexual exploitation and sex trafficking of minors as it included no categories for these offenses. It included a category for prostitution; given the nature of the UCR/SRS, however, it provided no detail on the victim (even assuming that the prostituted minor and not the purchaser is the victim), including the victim’s age.
A shift in the way UCR/SRS data are being collected, begun in January 2013, will better enable identification of commercial sexual exploitation and sex trafficking of minors in the United States. Even with these improvements, however, an important limitation remains in that FBI crime data reflect only those crimes of which the police became aware, that they recorded as commercial sexual exploitation or sex trafficking of minors, and that they submitted voluntarily to the FBI. If the police are not notified or the crime is not recognized, it will not appear in the FBI crime data.
Similarly, the Supplemental Homicide Reports historically could offer no information regarding the homicide of minors resulting from commercial sexual exploitation and sex trafficking. The changes made to the Supplemental Homicide Reports in accordance with the William Wilberforce Trafficking Victims Protection Reauthorization Act of 2008 do not allow the estimation of commercial sexual exploitation and sex trafficking of minors resulting in homicide. The act calls for the identification of homicides resulting from human trafficking and does not allow one to disaggregate sex trafficking from the broader trafficking category.
The NIBRS has historically shared some of the same limitations of the UCR/SRS with respect to estimating the commercial sexual exploitation and sex trafficking of minors in the United States. In the past, it included no categories for these crimes. It did include categories for those arrested for prostitution and pornography, which provided some useful information. Using these data, however, one could not disaggregate prostitution from pornography. These are the data used by Finkelhor and Ormrod (2004a,b) to study pornography and prostitution involving juveniles. As these authors note, while useful, the NIBRS has limitations that preclude its use for mak-
ing valid or reliable estimates (Finkelhor and Ormrod, 2004a,b). First, the NIBRS treats prostitution offenses as “crimes against society” rather than crimes against individuals, meaning one cannot obtain information on the victims of prostitution. Second, information about solicitors and purchasers also is masked as it is aggregated into the category “all other offenses.”
Much of this criticism is subject to change as the NIBRS was also required, beginning in January 2013, to include information pertinent to the estimation and understanding of commercial sexual exploitation and sex trafficking of minors in the United States. Like most NIBRS data, these new categories are not subject to the hierarchy rule. The greatest limitation of these new data is that they are subject to police reporting. Like all official data sources (including the UCR/SRS), the NIBRS is dependent on accurate recording of an event to and by a police officer. If an officer encounters a commercially sexually exploited 12-year-old who has committed another crime (e.g., shoplifting), the incident may not be recorded in a way that identifies commercial sexual exploitation or sex trafficking. And because the NIBRS (and UCR/SRS) are voluntary, estimates depend upon an agency’s submitting the data to the FBI, and not all agencies do so. Currently, NIBRS data are not collected from a representative set of law enforcement agencies in the nation; only a fraction of the nation’s agencies submit data to the system, and participating agencies tend to be smaller in size (Lynch and Addington, 2007). No jurisdiction with more than 1 million citizens reports data to the NIBRS; thus, commercial sexual exploitation and sex trafficking that occurs in large metropolitan areas will not be identified by the system.
Unlike the UCR/SRS and NIBRS, the NCVS is not limited by police practices or policies. However, it is characterized by limitations germane to the estimation of commercial sexual exploitation and sex trafficking of minors in the United States. First, the NCVS does not collect data on the crimes of sexual exploitation, sex trafficking, survival sex, kidnapping, prostitution, pornography, or a myriad of others; it collects information on a relatively restricted list of street and property crimes (e.g., rape, sexual assault, robbery, aggravated assault, simple assault, purse snatching, pocket picking, burglary, motor vehicle theft, property theft). Even if a category for commercial sexual exploitation and sex trafficking of minors were added to the NCVS, problems related to estimation would remain. Recall that the NCVS interviews only persons aged 12 or older. This means that commercial sexual exploitation and sex trafficking of persons aged 11 or younger would not be counted in the NCVS. Another limitation is that the NCVS is based on a household sample; that is, one must be living in a household to be eligible to participate. This methodology means that homeless persons are not reflected in the NCVS data. The literature strongly suggests that gathering information from the homeless population is vitally important in
understanding commercial sexual exploitation and sex trafficking of minors in the United States.
EXISTING NON-CRIMINAL JUSTICE MEASUREMENT PROGRAMS
Numerous data collection efforts outside of the criminal justice system have the potential to advance understanding of the commercial sexual exploitation and sex trafficking of minors in the United States. This section includes brief descriptions of selected efforts that gather data relevant to commercial sexual exploitation and sex trafficking (e.g., child sexual abuse) or vulnerable populations (e.g., homeless youth) and risk behaviors (e.g., substance use/abuse). While data of this nature may not contribute to a more accurate accounting of victims of commercial sexual exploitation and sex trafficking, they may be a crucial source of information regarding risk and opportunities for prevention and intervention.
Youth Risk Behavior Survey (YRBS)
The YRBS is a national school-based survey that examines heath-risk behaviors among students in grades 9-12. The YRBS, developed by the Centers for Disease Control and Prevention (CDC) in 1990, is conducted biennially to monitor behaviors that “contribute to the leading causes of morbidity and mortality among youth and adults, often are established during childhood and adolescence, extend into adulthood, are interrelated, and are preventable” (CDC, 2012b, p. 1). State and local education and health agencies are permitted to supplement the national survey to meet their individual needs. The large sample size and relatively long history of the YRBS allow for trend analysis of health-risk behaviors, as well as comparisons among states. However, the data captured by the national YRBS are limited to voluntary self-reporting by students enrolled in grades 9-12 of public or private schools. The survey collects information on sexual behaviors (e.g., having four or more sexual partners and first sexual intercourse prior to age 13) that increase the risk for unintended pregnancies, HIV, and other sexually transmitted infections.
The YRBS may provide the ability to identify risk for commercial sexual exploitation and sex trafficking of minors, as well as potential prevention and intervention points. Like all the data sources discussed here, however, it is limited in its ability to support comprehensive estimates of these crimes in the United States. While the survey offers access to minors at risk of or currently victimized by commercial sexual exploitation and sex trafficking, the data have important limitations for better understanding these crimes. First, YRBS data reflect a limited population. The survey does not gather information from some populations that are at increased risk for
commercial sexual exploitation and sex trafficking of minors, such as youth who are not enrolled in school for any reason, those housed in a juvenile detention facility, those enrolled in alternative schools, and those who are home-schooled. Second, while the YRBS gathers information from adolescents approximately aged 14 to 17, it offers no information on younger children and adolescents. Yet extant findings indicate that commercial sexual exploitation and sex trafficking can and do affect younger persons. Third, like all surveys, the YRBS depends on respondents’ recognizing commercial sexual exploitation and sex trafficking and revealing sensitive information associated with this victimization. As noted earlier, research indicates that victims of commercial sexual exploitation and sex trafficking often are unable to recognize themselves as victims of these crimes, and that victims often refuse to share this information when they know it. The ability of the YRBS to elicit responses related to commercial sexual exploitation and sex trafficking is further compromised by the absence of direct questions regarding these crimes. While the YRBS offers some advantages for understanding commercial sexual exploitation and sex trafficking of minors in the United States, these limitations are significant.
National Longitudinal Study of Adolescent Health (Add Health)
The National Longitudinal Study of Adolescent Health, referred to as Add Health, currently is the most comprehensive longitudinal study of the factors influencing adolescent health and the associations between adolescent experiences and social, behavioral, and health outcomes in adulthood (Harris, 2011). Data collection for Waves I and II of Add Health began in 1994 and included in-school questionnaires administered to adolescents, peers, and administrators; in-home interviews of adolescents, parents, siblings, and other members of the household; and community data sources such as the U.S. Census and the National Center for Health Statistics (Harris et al., 2009). Questionnaires and interviews elicited information regarding the adolescents’ individual traits, personal relationships (family, peer, and romantic), sexual experiences, delinquency, exposure to violence, physical health, and mental health. In 2001-2002, when participants were between the ages of 18 and 26, Wave III examined how experiences and behaviors in adolescence influenced the transition to adulthood. For example, data were collected on participants’ income, spirituality, involvement with mentoring, and family relationships. In 2007-2008, 15,701 original Add Health participants between the ages of 24 and 32 were interviewed for Wave IV “to study developmental and health trajectories across the life course of adolescence into young adulthood using an integrative approach that combines social, behavioral, and biomedical sciences in its research objectives, design, data collection, and analysis” (Harris, 2011, p. 9). The Add
Health data sets have resulted in a proliferation of research across multiple disciplines. Researchers have studied how adolescent experiences and outcomes in young adulthood are associated with multiple factors, including the impact of mentoring relationships, the influence of peer networks, the roles of family members, sexual experiences, substance use, mental health, education, and religion.
While Add Health data are widely used and extremely valuable, the data are limited in their ability to support estimates of commercial sexual exploitation and sex trafficking of minors in the United States. The clearest limitation in this regard is that Add Health is a panel study, and in-sample respondents are currently 28 to 36 years of age. Therefore, obtaining information about commercial sexual exploitation and sex trafficking from these respondents would require adding to the next wave of interviews questions about possible experiences with these crimes that occurred in the mid-1990s. Add Health currently includes no questions focused specifically on commercial sexual exploitation and sex trafficking of minors. The Wave III, Section 16, portion of the questionnaire does focus on sexual experiences. Among the questions asked are whether the respondent paid to have sex with someone else or was paid to have sex with another. Each of these questions is asked with an “ever” reference period and with an “in the last 12 months” period (Harris et al., 2009). While responses to these questions may include instances of commercial sexual exploitation and sex trafficking of minors, there is no way to parse these cases. Moreover, no follow-up questions are asked about the age at which these events happened. Given that these questions were asked of people aged 18 to 26, even the shorter reference period of 12 months cannot assist in identifying commercial sexual exploitation and sex trafficking of minors in the United States.
Wave IV asked questions of respondents aged 24 to 32 regarding sexual activity. Included was a question asking, “In the past 12 months, how many times have you paid someone to have sex with you or has someone paid you to have sex with them?” (Harris et al., 2009). Unfortunately for purposes of estimating commercial sexual exploitation and sex trafficking of minors, this question conflates those who paid for sex with those who were paid for sex in the last 12 months. Because of the 12-month retrospective nature of the question, moreover, this item captures behavior of adults aged 23 to 31. While questions explicitly asking whether the respondent was paid to have sex while a minor or paid to have sex with a minor could theoretically be added in the future, there currently is no indication that this will occur. And if questions were asked about commercial sexual exploitation and sex trafficking that may have occurred to respondents when they were minors, it is important to recognize that responses would be subject to memory errors. Finally, even if the Add Health respondents were used to
gather retrospective information on commercial sexual exploitation and sex trafficking that occurred when they were children and adolescents, it is important to reiterate that those participating in Add Health were listed on school rosters in the mid-1990s; children and adolescents who were not in school for any reason, including being housed in a detention facility or being home-schooled at that time, were ineligible for inclusion in the survey, and their experiences are not reflected in the data.
Runaway and Homeless Youth Management
Information System (RHYMIS)
The Runaway and Homeless Youth Management Information System (RHYMIS) is an automated data collection tool that collects information from grantees in three federal programs that provide support to homeless and runaway youth in the United States: the Basic Center Program, the Transitional Living Program, and the Street Outreach Program (ACF, 2012). The Family and Youth Services Bureau in the Administration on Children, Youth and Families awards and administers federal grants in these programs. Collectively, these programs support homeless and runaway youth in need of emergency services (e.g., referrals to emergency shelters, hygiene products, crisis intervention services, counseling) and longer-term services (e.g., stable living arrangements [up to 21 months], General Educational Development preparation, vocational training, parenting skills, assistance with substance abuse, routine physical health care). The Reconnecting Homeless Youth Act of 200815 (the most recent amendment to the original authorizing legislation, the Runaway Youth Act of 1974) directs grantees to collect, maintain, and provide to the Department of Health and Human Services, to the maximum extent possible, statistical data on the youth served by each of the three programs. The information submitted to RHYMIS includes demographic characteristics of youth served, types of services provided, living arrangements prior to entering the program, and aftercare plans upon exit from the program. A child or adolescent may enter a program, receive services, exit the program, and then subsequently reenter the program. Each entry into a program is recorded in RHYMIS by grantees in all three programs. During fiscal year (FY) 2011, the Basic Center Program recorded 36,859 entrances, the Transitional Living Program recorded 3,880 entrances, and the Street Outreach Program recorded 772,732 outreach contacts to youth (ACF, 2012). In addition to youth served, RHYMIS records the number of youth turned away; in FY 2011, this number was 2,484 for the Basic Center Program and 4,619 for the Transitional Living Program (ACF, 2012).
15Public Law 110-378.
An accurate, current national estimate of homeless and runaway youth does not exist. Available data sources, such as RHYMIS, likely underrepresent the actual number of children and adolescents who are homeless or living in unstable housing. This population of youth is highly vulnerable to significant health risks, including survival sex, sexually transmitted infections, substance abuse, depression, and suicide (Kaestle, 2012).
RHYMIS has the advantage of offering access to a portion of a difficult-to-reach population: homeless and runaway youth. Access to this population is crucial for improved estimation of commercial sexual exploitation and sex trafficking of minors in the United States. However, RHYMIS also is characterized by limitations for estimation purposes. First, as noted above, RHYMIS does not offer a representative accounting of all homeless and runaway youth, which limits its utility for estimating commercial sexual exploitation and sex trafficking of minors in the United States. To the extent that minors who are victims of these crimes do not obtain services that report to RHYMIS, those minors will not be represented in these data. Second, RHYMIS data offer counts of services provided, not prevalence or incidence information. The same individual may obtain services multiple times for multiple incidents of commercial sexual exploitation or sex trafficking. Third, RHYMIS includes only children and adolescents who are runaways or are homeless. While extant research documents the vulnerability and victimization of runaway, thrown-away, and homeless children and adolescents, findings also indicate that many who are at risk and many who are being victimized have not run away and are not homeless.
Adverse Childhood Experiences (ACE) Study
The ACE Study was a longitudinal study that examined the impact of early life experiences on future health risk behaviors. It was developed by CDC and conducted at Kaiser Permanente’s Health Appraisal Clinic in San Diego, California. Between 1995 and 1997, more than 17,000 individuals who underwent a comprehensive physical examination voluntarily completed anonymous surveys regarding childhood experiences, if any, with abuse (emotional, sexual, and physical), household dysfunction (e.g., mental illness of a caregiver or incarceration of a household member), and neglect (physical and emotional) (CDC, 2012a). The total number of adverse childhood experiences reported by a study participant represents the individual’s ACE score. The major findings of the ACE Study suggest that the higher the ACE score, the greater is the risk for developing behaviors that contribute to the leading causes of illness, death, and poor social outcomes in adulthood, including alcoholism, drug abuse, and depression (Anda et al., 2006). The study’s data sets are available to researchers and have resulted in numerous scientific publications regarding the association
of adverse childhood experiences with various poor health and social outcomes in adulthood. For example, victims of childhood sexual abuse are more than twice as likely to attempt suicide and are at a significantly higher risk for interpersonal problems than adults who did not experience such abuse (Dube et al., 2005). CDC and Kaiser Permanente continue to track the health and social outcomes of the study participants in order to assess the associations with earlier adverse childhood experiences.
Although the study sample was large (n = 17,337), it was not demographically representative. For example, the sample comprised 74.8 percent Caucasians, and more than 90 percent of those surveyed had graduated from high school. In addition, all participants were over the age of 18, and nearly half were over age 60 (CDC, 2012a). The demographic characteristics of the sample raise questions about the generalizability of the study’s findings to broader populations, as well as the accuracy of recollections of childhood experiences.
The limitations of the ACE Study mirror those of Add Health: The study is not accepting new participants, and there is no evidence of plans to contact original participants for future waves of questions. A second issue is that the study did not gather data specifically on the commercial sexual exploitation or sex trafficking of minors. Since these crimes are forms of child sexual abuse, however, the study findings regarding the association between child sexual abuse and increased risk for poor health and social outcomes in adulthood likely are applicable to these victims of exploitation. As discussed in Chapter 3 of this report, the adverse childhood experiences tracked by the ACE Study are believed to be risk factors for commercial sexual exploitation and sex trafficking of minors; therefore, child and adolescent victims of these crimes are likely at increased risk for the poor health and social outcomes identified by the ACE Study and related research.
National Child Abuse and Neglect Data System (NCANDS)
The NCANDS is an electronic data collection and analysis system that examines child abuse and neglect reports made to state child welfare agencies in all 50 states, the District of Columbia, and the Commonwealth of Puerto Rico (Children’s Bureau, 2011). The Child Abuse Prevention and Treatment Act,16 which authorized this national data collection and analysis system, defines the minimum threshold of child abuse as “any recent act or failure to act on the part of a parent or caretaker which results in death, serious physical or emotional harm, sexual abuse or exploitation; or an act or failure to act, which presents an imminent risk of serious harm” (Children’s Bureau, 2011, p. vii). The information collected and analyzed
1642 U.S.C. 5101.
through NCANDS includes the number of child abuse and neglect reports made to the state child welfare agency; the source of the report (e.g., service professional or neighbor); the type of maltreatment reported (e.g., neglect, physical or sexual abuse); the relationship between the perpetrator and the victim; demographic characteristics of the victim (e.g., age, race/ethnicity); demographic characteristics of the perpetrator; risk factors identified for the child and the family (e.g., substance abuse, physical disability, domestic violence); services, if any, that were rendered as a result of the report; and the size and responsibilities of the workforce responsible for managing reports of child abuse and neglect (Children’s Bureau, 2011). All jurisdictions voluntarily submit agency-level data to NCANDS, and nearly all states submit case-level data, referred to as the Child File, which tracks information for a specific child. Federal agencies use the NCANDS data to inform policies relating to child abuse and neglect, and the data also are available to independent researchers. In FY 2011, approximately 3.4 million reports of child abuse and neglect were made to state child welfare agencies, involving an estimated 6.4 million children (Children’s Bureau, 2011). Of those referrals, 1.7 million were screened in, and an investigation was initiated.
Victims of child abuse and neglect are at increased risk for commercial sexual exploitation and sex trafficking. NCANDS therefore provides an opportunity to identify risk, as well as potential for prevention and intervention.
Like all the data sources discussed here, NCANDS has advantages and disadvantages for purposes of measuring commercial sexual exploitation and sex trafficking of minors in the United States. An important advantage is that the data are based on child abuse and neglect reports and capture experiences of all children and adolescents, regardless of whether they are in school, living in a housing unit, or homeless. A limitation of the data pertains to dependence on the perception of a reporter. The data are focused on child abuse and neglect, so to the extent that a person witnessing the maltreatment mistakes a child or adolescent for an adult, the incident may go unreported. This issue is problematic for commercial sexual exploitation and sex trafficking of minors as research shows that many of the young people being exploited and trafficked appear older than 18. A second limitation is that the system gathers data only on acts committed by a parent or caretaker. While there is evidence that parents and caretakers commit acts of commercial sexual exploitation and sex trafficking involving their own children, evidence also suggests that these crimes often are committed by persons other than parents and caretakers. The latter instances would not be reflected in the NCANDS data.
CHALLENGES AND OPPORTUNITIES
Estimates of a social problem are important. They inform interested parties about the extent of the problem, they lead to attention to the need to address the problem and the funding required to do so, and they offer a baseline for evaluating policies. In addition, estimates provide information for policy makers, identify the need for education, point to where assistance is needed, and help direct assistance toward victimized populations. As this chapter has explained, however, generating crime estimates is not an easy task and in some cases is extraordinarily difficult. Commercial sexual exploitation and sex trafficking of minors in the United States are social problems that are characterized by numerous challenges to estimation. The population of interest is difficult to identify and locate and at times uncooperative. The sampling strategies available for estimating these problems are limited. Consensus is lacking on the definitions of the problems. Measurement of the problems is costly, time-consuming, and extremely difficult. It is not surprising, then, that extant estimates are imperfect.
Best (2012) notes that all statistics are imperfect, but some are less perfect than others. This is an important point. First, it highlights the costs of basing decisions on poor estimates. If an estimate is wildly inaccurate and grossly inflates the extent of a problem, scarce resources (time and money) will be misallocated, and other, more prevalent problems will go unaddressed. Best’s comment also indicates that a perfect estimate is unlikely. While identifying the limitations of an estimate is easy, identifying what is needed to improve it is more difficult. Certainly, the use of a probability sample of all minors in the United States would be useful, but it may not be feasible for many reasons. At some point, one must ask whether the available national estimates are good enough. Do extant estimates suggest that the problems are large enough to warrant attention and resources? Would slightly higher estimates than those already obtained change the approach to the problems or alter their importance? Is there value in seeking more perfect estimates in an environment of scarce resources?
The committee identified two potential opportunities to enhance measurement and understanding of commercial sexual exploitation and sex trafficking of minors in the United States: (1) leverage existing measurement efforts that collect data on related issues or populations, and (2) shift focus and resources from exclusively national-level counting to more targeted counting.
As noted in this chapter, two existing crime measurement efforts (NIBRS and UCR/SRS) have begun to collect data related to all forms of human trafficking, including sex trafficking of minors (as of January 2103). Although existing non-criminal justice measurement efforts do not currently collect data on commercial sexual exploitation and sex trafficking, a
number of these efforts that collect information on related issues and potential risk factors and from relevant populations could be adapted to do so. For example, the YRBS and Add Health collect data on related issues and potential risk factors, such as dating violence, sexually transmitted infections, substance use/abuse, and mental health status from national samples of adolescents. In addition, the RHYMIS collects data from a national sample of runaway, “thrown-away,” and homeless youth, populations that are at risk for commercial sexual exploitation and sex trafficking. Like the NIBRS and UCR/SRS, these measurement efforts could add specific questions and data elements related to commercial sexual exploitation and sex trafficking to enhance measurement and understanding of these problems. Additional research would be needed to determine specific questions to add to the various data collection efforts.
While changes to current data collection efforts could enhance measurement and understanding of commercial sexual exploitation and sex trafficking of minors in the United States, as discussed earlier, these efforts would still yield insufficient estimates of these problems. As noted in this chapter, efforts to estimate the overall occurrence of commercial sexual exploitation and sex trafficking of minors in the United States have been largely unsuccessful given the difficulties inherent in measuring these crimes. As a result, insufficient attention, research, and resources have been devoted to resolving these problems. It is the position of the committee that a continued focus on obtaining better national incidence and prevalence estimates of commercial sexual exploitation and sex trafficking of minors in the United States may detract from progress on other important aspects of these problems.
Other fields of research and practice demonstrate that it is possible to make progress on issues even in the absence of strong evidence on their nature and extent. For example, the scope and severity of sexual violence, intimate partner violence, and child abuse and neglect were viewed with skepticism and were characterized by poor estimates during the early stages of work on these problems. (For a more detailed discussion of estimates of child abuse and neglect, see the recent Institute of Medicine [IOM]/National Research Council [NRC] publication Child Maltreatment Research, Policy, and Practice for the Next Decade: Workshop Summary [IOM/NRC, 2012].) Although debates regarding estimates of sexual violence, intimate partner violence, and child abuse and neglect continue, these issues are now accepted as legitimate problems that have benefited from greater public attention, improved funding, and research. Ideally, work on commercial sexual exploitation and sex trafficking of minors in the United States will follow suit. Therefore, the committee does not suggest that efforts to obtain better estimates of commercial sexual exploitation and sex trafficking of minors in the United States be abandoned. Rather, the committee urges
that efforts to measure the scope and severity of these problems not inhibit other important work.
Focusing on better prevalence and incidence estimates is challenging and expensive. Devoting additional resources exclusively to further national-level counting efforts may not be the best strategy to advance work on commercial sexual exploitation and sex trafficking of minors in the United States. An alternative strategy is to shift focus and resources from national-level counting to more targeted counting (e.g., regional or subpopulation estimation). In this scenario, there are a number of possible benchmark measures that can be used to better understand commercial sexual exploitation and sex trafficking of minors in the United States, including counting
• survivors from a specified region or subpopulation receiving services,
• charges brought forth by prosecutors, and
• successful convictions of exploiters and traffickers.
Finally, valuable lessons regarding difficult-to-measure problems can be drawn from the above-mentioned IOM/NRC report Child Maltreatment Research, Policy, and Practice for the Next Decade: Workshop Summary (IOM/NRC, 2012) and the forthcoming NRC report Estimating the Incidence of Rape and Sexual Assault.
In sum, based on its review of the available evidence, the committee maintains that, despite the current imperfect estimates, commercial sexual exploitation and sex trafficking of minors in the United States clearly are problems of great concern and worthy of attention. Therefore, this report and the committee’s recommendations go beyond counting commercial sexual exploitation and sex trafficking of minors in the United States to emphasize that, unless additional resources become available, existing resources should be focused on what can be done to assist the victims of these crimes.
FINDINGS AND CONCLUSIONS
This chapter has emphasized that efforts to generate estimates of commercial sexual exploitation and sex trafficking of minors in the United States are characterized by many difficulties. Given the inherent challenges, no approach will result in perfect—or perhaps even nearly perfect—estimates. Still, attempts to provide better, more targeted estimates are warranted. The committee finds that
|2-1||No reliable national estimate exists of the incidence or prevalence of commercial sexual exploitation and sex trafficking of minors in the United States.
|2-2||Definitional issues inhibit the measurement of these problems.
|2-3||Clear and reasonable definitions of commercial sexual exploitation and sex trafficking of minors are needed to generate better estimates.
|2-4||Reliable estimates of numbers of exploiters and traffickers and solicitors and purchasers do not exist for commercial sexual exploitation and sex trafficking of minors in the United States.
|2-5||Estimates that rely solely on interactions with law enforcement, crime victim surveys, and arrest records are inherently underinclusive.
|2-6||Specific data fields related to commercial sexual exploitation and sex trafficking of minors have been added to existing crime measurement efforts (e.g., UCR/SRS and NIBRS) to enhance estimates of these crimes; similar changes could be made to existing non-criminal justice measurement efforts (e.g., YRBS and Add Health) to enhance estimates of commercial sexual exploitation and sex trafficking of minors that may not involve law enforcement (e.g., instances in which police are not notified or the crime is not recognized).
|2-7||Increased awareness and understanding of commercial sexual exploitation and sex trafficking of minors in the United States can improve data collection efforts and estimates.
|2-8||While some research methods (e.g., capture/recapture and respondent-driven sampling) look beyond simple reporting of known incidents to generate estimates, the methods currently used to gather national-level data (i.e., UCR/SRS, NIBRS, NCVS) cannot capture the full extent of commercial sexual exploitation and sex trafficking of minors in the United States. The limited information available from the UCR/SRS, NIBRS, and NCVS needs to be combined with data from other sources or circumstances in which victims are identified.
|2-9||Devoting additional attention and limited resources to the extremely difficult task of counting victims of commercial sexual exploitation and sex trafficking of minors is unnecessary.
The next several chapters describe an array of settings and stakeholders that—knowingly or not—interact with underage victims of commercial sexual exploitation and sex trafficking and individuals who are at risk for these forms of victimization. Chapter 11 includes the committee’s recommendations for strategies for future measurement efforts.
ACF (Administration for Children and Families). 2012. The Runaway and Homeless Youth Management System (RHYMIS). https://extranet.acf.hhs.gov/rhymis/custom_reports.jsp (accessed December 21, 2012).
Addington, L. A., and C. M. Rennison. 2008. Rape co-occurrence: Do additional crimes affect victim reporting and police clearance of rape? Journal of Quantitative Criminology 24(2):205-226.
Anda, R. F., V. J. Felitti, J. D. Bremner, J. D. Walker, C. Whitfield, B. D. Perry, S. R. Dube, and W. H. Giles. 2006. The enduring effects of abuse and related adverse experiences in childhood. A convergence of evidence from neurobiology and epidemiology. European Archives of Psychiatry and Clinical Neuroscience 256(3):174-186.
Bales, K., and S. Lize. 2007. Investigating human trafficking: Challenges, lessons learned, and best practices. FBI Law Enforcement Bulletin 76(4):24-32.
Banks, D., and T. Kyckelhahn. 2011. Characteristics of suspected human trafficking incidents, 2008-2010. Washington, DC: Bureau of Justice Statistics.
Best, J. 2012. Damned lies and statistics: Untangling numbers from the media, politicians, and activists. Berkeley, CA: University of California Press.
Biderman, A. D., and A. J. Reiss. 1967. On exploring the “dark figure” of crime. The ANNALS of the American Academy of Political and Social Science 374(1):1-15.
Bureau of Justice Statistics. 2013. The nation’s two crime measures. http://bjs.gov/index.cfm?ty=tp&tid=3 (accessed April 24, 2013).
CDC (Centers for Disease Control and Prevention). 2012a. Adverse Childhood Experiences (ACE) Study participant demographics. http://www.cdc.gov/ace/demographics.htm (accessed December 21, 2012).
CDC. 2012b. Youth risk behavior surveillance—United States, 2011. Morbidity and Mortality Weekly Report Surveillance Summaries 61(4):1-168.
Child Trends. 2012. Violent crime victimization indicators on children and youth. http://childtrendsdatabank.org/sites/default/files/71_Violent_Crime_Victimization.pdf (accessed April 24, 2013).
Children’s Bureau. 2011. Child maltreatment 2011. http://www.acf.hhs.gov/sites/default/files/cb/cm11.pdf (accessed December 20, 2012).
Curtis, R., K. Terry, M. Dank, K. Dombrowski, and B. Khan. 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: Center for Court Innovation.
Dodge, R. W., and F. D. Balog. 1987. Series crimes: Report of a field test. Technical report, NCJ 104615. Washington, DC: Bureau of Justice Statistics.
Dube, S. R., R. F. Anda, C. L. Whitfield, D. W. Brown, V. Felitti, M. X. Dong, and W. H. Giles. 2005. Long-term consequences of childhood sexual abuse by gender of victim. American Journal of Preventive Medicine 28(5):430-438.
Estes, R. J., and N. A. Weiner. 2002. The commercial sexual exploitation of children in the U.S., Canada, and Mexico. Philadelphia, PA: Center for the Study of Youth Policy.
FBI (Federal Bureau of Investigation). 2012. Innocence lost national initiative. http://www.fbi.gov/about-us/investigate/vc_majorthefts/cac/innocencelost (accessed September 17, 2012).
FBI. 2013a. Supplementary homicide report (OMB form no. 1110-0002). http://www.fbi.gov/about-us/cjis/ucr/nibrs/addendum-for-submitting-cargo-theft-data/shr (accessed April 24, 2013).
FBI. 2013b. Uniform crime reporting statistics. http://bjs.gov/ucrdata/abouttheucr.cfm (accessed April 24, 2013).
Finkelhor, D., and R. Ormrod. 2004a. Child pornography: Patterns from NIBRS. Washington, DC: Office of Juvenile Justice and Delinquency Prevention.
Finkelhor, D., and R. Ormrod. 2004b. Prostitution of juveniles: Patterns from NIBRS. Washington, DC: Office of Juvenile Justice and Delinquency Prevention.
Fisher, B. S., L. E. Daigle, and F. T. Cullen. 2010. What distinguishes single from recurrent sexual victims? The role of lifestyle-routine activities and first-incident characteristics. Justice Quarterly 27(1):102-129.
Gozdziak, E. M. 2008. Human trafficking in the United States: Knowledge gaps and research priorities. Geneva, Switzerland: International Organization for Migration.
Gragg, F., I. Petta, H. Bernstein, K. Eisen, and L. Quinn. 2007. New York prevalence study of commercially sexually exploited children. Rensselaer: New York State Office of Children and Family Services.
Greene, J. M., S. T. Ennett, and C. L. Ringwalt. 1999. Prevalence and correlates of survival sex among runaway and homeless youth. American Journal of Public Health 89(9):1406-1409.
Harris, K. M. 2011. Design features of add health. Chapel Hill, NC: Carolina Population Center, University of North Carolina at Chapel Hill.
Harris, K. M., C. T. Halpern, E. Whitsel, J. Hussey, J. Tabor, P. Entzel, and J. R. Udry. 2009. The national longitudinal study of adolescent health: Research design. http://www.cpc.unc.edu/projects/addhealth/design (accessed April 24, 2013).
Hart, T. C., and C. M. Rennison. 2003. Reporting crime to the police, 1992-2000. Washington, DC: Bureau of Justice Statistics.
Heckathorn, D. D. 1997. Respondent-driven sampling: A new approach to the study of hidden populations. Social Problems 44(2):174-199.
IOM/NRC (Institute of Medicine/National Research Council). 2012. Child maltreatment research, policy, and practice for the next decade: Workshop Summary. Washington, DC: The National Academies Press.
Kaestle, C. E. 2012. Selling and buying sex: A longitudinal study of risk and protective factors in adolescence. Prevention Science 13(3):314-322.
Lauritsen, J., J. Owens, M. Planty, M. Rand, and J. Truman. 2012. Methods for counting high-frequency repeat victimization in the national crime victimization survey. Washington, DC: Bureau of Justice Statistics.
Lynch, J., and L. A. Addington. 2007. Understanding crime statistics: Revisiting the divergence of the NCVS and the UCR. New York: Cambridge University Press.
Mitchell, K. J., D. Finkelhor, L. M. Jones, and J. Wolak. 2010. Use of social networking sites in online sex crimes against minors: An examination of national incidence and means of utilization. Journal of Adolescent Health 47(2):183-190.
Moossy, R. 2009. Sex trafficking: Identifying cases and victims. National Institute of Justice Journal 262:2-11.
Ohio Trafficking in Persons Study Commission Research and Analysis Sub-Committee. 2012. Report on the prevalence of human trafficking in Ohio to Attorney General Richard Cordray. Toledo: Ohio Human Trafficking Study Commission.
Pease, K. 1998. Repeat victimisation: Taking stock. Crime prevention and detection series paper 90. London: Home Office.
Puzzanchera, C., and B. Adams. 2011. Juvenile arrests 2009. Washington, DC: U.S. Department of Justice Office of Juvenile Justice and Delinquency Prevention.
Quin, M., P. Merritt, M. Garner, J. Robinson, A. Tedford, T. Allen, A. Allen, C. Purviance, M. Prickett, A. Sanders, P. Beck, A. Marcotte, and N. Garcia. 2011. Tennessee human sex trafficking and its impact on children and youth. Nashville: Tennessee Bureau of Investigation and Vanderbilt Center for Community Studies.
Rand, M. R., and C. M. Rennison. 2005. Bigger is not necessarily better: An analysis of violence against women estimates from the national crime victimization survey and the national violence against women survey. Journal of Quantitative Criminology 21(3): 267-291.
Rennison, C. M. 2002. Rape and sexual assault: Reporting to police and medical attention, 1992-2000. Washington, DC: Bureau of Justice Statistics.
Rennison, C. M. 2009. Crime reports and statistics. In 21st century criminology: A reference handbook, edited by J. M. Miller. Thousand Oaks, CA: SAGE Publications, Inc.
Rieger, A. 2007. Missing the mark: Why the trafficking victims protection act fails to protect sex trafficking victims in the United States. Harvard Journal of Law and Gender 30(1):231-256.
Stransky, M., and D. Finkelhor. 2008. How many juveniles are involved in prostitution in the U.S.? Durham, NH: Crimes Against Children Research Center.
Svedin, C., and G. Priebe. 2007. Selling sex in a population-based study of high school seniors in Sweden: Demographic and psychosocial correlates. Archives of Sexual Behavior 36(1):21-32.
Truman, J. 2011. Criminal victimization, 2010. Washington, DC: Bureau of Justice Statistics.
Tyldum, G., and A. Brunovskis. 2005. Describing the unobserved: Methodological challenges in empirical studies on human trafficking. International Migration 43(1-2):17-34.