Building the Evidence on Preventing Youth Commercial Sexual Exploitation: A Nonrandomized,Quasi-experimental Evaluation of the Not a Number Trafficking Prevention Program

Abstract

Few evaluations have been conducted of trafficking and educational programs aimed at reducing commercial sexual exploitation (CSE) and trafficking of youth, despite the growing implementation of these programs across the United States. The current study uses a quasi-experimental design to evaluate the Not a Number (NAN) trafficking prevention program in two high schools in Houston, TX (N = 449). In both high schools, changes were measured through preprogram and postprogram surveys (approximately 2 months postprogram delivery) for students receiving NAN during health class (implementation group) and students in physical education classes who did not receive NAN (control group). We hypothesized that students in the implementation groups who received the NAN program would see the following changes: (a) improved awareness and understanding of the problem of CSE; (b) increased help-seeking capacity; and (c) improved safety planning. The study found support for increased CSE awareness and increased help-seeking self-efficacy. There was mixed support for the program’s impact on safety planning. The current study is an advancement in knowledge about the efficacy of prevention education approaches to CSE of youth. While findings are promising, additional evaluations of this and other prevention education programs directed at CSE are needed to build the knowledge of what works to effectively prevent CSE among youths.

Keywords

prevention of child abuse child abuse sexual abuse adolescent victims sexual assault prevention

Prevention education has been used with success to provide children and youth with protective information about abuse and victimization risks, including sexual abuse, dating violence, and sexual assault (DeGue et al., 2021; Piolanti & Foran, 2021; Walsh et al., 2018). More recently, commercial sexual exploitation (CSE) of children and youth has been a focus of prevention efforts for youth, as the mental and physical health costs for this form of victimization have become better recognized. For youth, CSE refers to the exchange of sex or a sexual act for food, money, shelter, or other items of value by a person under the age of 18 (“Trafficking Victims Protection Act,” 2000). It is sometimes referred to as domestic minor sex trafficking or child sex trafficking victimization. Given the significant negative impact of CSE on youth (Cole et al., 2016), legislation has sought to increase education and prevention programming across the United States (Committee for Children, 2021). However, the optimal educational approaches for preventing CSE are still unknown, and evaluation research to date has been limited. The current study was aimed at improving this evidence base by examining trafficking/CSE prevention outcomes for high school students through a quasi-experimental evaluation of a widely adopted trafficking prevention program Not a Number.

Impact of CSE

Although the exact scope of CSE of children and youth remains unknown (Finkelhor et al., 2017; Franchino-Olsen, 2021), recent population-based estimates suggest rates between 1.4% and 7.4%, depending on the context (Head et al., 2021; Martin et al., 2021). For some populations of youth, rates are likely higher. Research suggests that children with complex histories of adversity including childhood physical and sexual abuse and family violence are at a higher risk for CSE (Franchino-Olsen, 2021). Sexual minority youth, indigenous youth, and youth who come from lower income communities have also been found to be at a higher risk of experiencing CSE (Martin et al., 2021; Middleton et al., 2018). Health professionals and advocacy organizations describe working actively with children and youth who have experienced CSE in their professional roles in both urban and rural communities (Cole & Sprang, 2015; Greenbaum, 2017). The negative impact of CSE on youth has been well-documented: CSE victims have higher rates of post-traumatic stress, depression, and suicidality than their nonexploited peers (Moore et al., 2021). The ability to develop and sustain healthy relationships can be affected for survivors of CSE, given disruptions in relationship skills and attachments with others as a result of the victimization (Franchino-Olsen, 2021; Preble, 2021). CSE is also highly correlated with concurrent victimization and adversities, including interpersonal violence and substance use (Reid, 2014). Survivors of CSE may experience heightened levels of sexually transmitted infections, and unwanted pregnancy and abortion (Barnert et al., 2017; Edinburgh et al., 2015; Greenbaum et al., 2018; Lederer & Wetzel, 2014). In the aftermath of CSE, survivors can also experience problems obtaining medical and mental health care (Garg et al., 2020), which further undermine their physical, psychological, and behavioral well-being.

Prevention Education on Trafficking and CSE

The Trafficking Victims Protection Act of 2000 directed federal attention to the importance of reducing CSE and emphasized the need for the “development of educational curricula regarding the dangers of trafficking” (“Trafficking Victims Protection Act,” 2000). In 2021, additional legislation was introduced to increase funding to “train students, teachers, and youth development personnel how to better understand, recognize, prevent, and respond to human trafficking and the exploitation of children and youth” (“Human Trafficking and Exploitation Prevention Training Act, S. 3333, 116th Cong.,” 2021). Many states have already implemented policies directing schools to provide trafficking education to youth. For example, Florida and North Carolina have made education on CSE for educators and students mandatory, while other states, such as Texas, have policies that encourage schools to implement educational programs (Committee for Children, 2021).

However, while CSE and trafficking prevention education programs are being increasingly implemented in schools and communities, there is limited evaluation research to support program efficacy or guide best practices. In a systematic review of the literature on CSE education, Rizo et al. found an absence of rigorous outcome research (Rizo et al., 2019). Most of the literature provided recommendations for educational programs or described the formation of new prevention programs. A process and outcome study of a program that aimed to reduce CSE victimization for American Indian and Alaskan Native adolescent girls reported success in improving conditions that were placing youth at risk for commercial sexual victimization (Pierce, 2012). Another study provided evaluation data on a social justice-focused curriculum and found that the program increased students’ awareness of human trafficking (Scott et al., 2019). Finally, Rothman et al. conducted an evaluation of My Life, My Choice, a 10-session course provided to youth identified at high risk for CSE. Participants showed increased awareness of the dynamics of CSE over time, less sexually explicit behavior, and were more likely to have given help to a friend from pre to posttest (Rothman et al., 2021). However, while encouraging, none of these evaluations included control groups, and many involved programs that were developed specifically for the community involved and would be difficult to scale nationwide.

Although CSE has some unique contexts and trajectories that distinguish it from other forms of victimization, there is overlap in risk and protective factors with dating violence and sexual assault (O’Brien et al., 2022). Prevention research in these other areas has a more developed research base. Meta-analyses of prevention directed at reducing dating violence among teens and college students found that these programs are able to successfully improve knowledge of and attitudes toward dating violence (De La Rue et al., 2017; Wong et al., 2021). Another recent meta-analysis found support for prevention education in reducing dating violence, but with unclear outcomes for sexual violence (Piolanti & Foran, 2021). Findings from these studies provide encouragement for a curriculum-based approach to CSE prevention education, as well as highlight important issues to consider and study as trafficking and CSE prevention education and research moves forward.

The Not a Number Trafficking Prevention Program

The Not a Number (NAN) curriculum is an interactive, five-module program developed by the Love146 nonprofit organization. Love146 is an international human rights organization founded in 2002, working to end child sex trafficking and exploitation both in the United States and abroad. They developed the NAN program in 2014. To date, the NAN program has been delivered to youth in over 30 U.S. states, primarily in schools. In its third iteration at the time of this evaluation, the curriculum is designed to help prevent child trafficking and exploitation by: (a) providing youth with information on human trafficking; (b) teaching youth to recognize exploiter recruitment tactics and understand personal vulnerability both in person and online; (c) practicing refusal skills with youth that can help them avoid or reduce risk both in person and online; (d) identifying healthy support systems; (e) teaching skills to navigate potential and existing exploitative situations; and (f) increasing youth awareness of helping resources within their communities as well as nationally (Love146, 2020).

The developers worked to build the program with effective prevention programs generally, including varied teaching methods, sufficient dosage, opportunities for positive relationships, and developmental and sociocultural relevance (Nation et al., 2003). The program also used theory to ground its development, and the curriculum was built by drawing from empowerment theory (Zimmerman, 2000), social cognitive theory (Bandura, 2002), health belief model (Sheeran & Abraham, 1996), and the precaution process model (Weinstein et al., 2020). Empowerment theory describes how important protective skills are gained when self-efficacy and awareness of power dynamics are increased. Social cognitive theory is often used in health psychology and provides a model for how learning and increased self-efficacy can lead to change. Finally, the precaution process model describes how people take action to prevent harm, outlining how change behaviors are improved through information targeted at different levels of awareness.

The Houston Independent School District (HISD) in Texas implements the program regularly in several of their high schools with Love146 staff delivering the program, and this city was selected as the location for the evaluation. While the current study represents the first evaluation of NAN using a control group, research has been conducted analyzing the extensive pre- and post-survey data collected by Love146 as a part of its ongoing implementation procedures (Jones et al., 2024).

Current Study

The current study aimed to build on prior research on the Not a Number program and begin building a more rigorous evidence base for CSE prevention education for youth. Using a quasi-experimental design, the evaluation was conducted in two high schools in Houston, TX (N = 449). In both high schools, changes were measured through preprogram and postprogram surveys (approximately 2 months postprogram delivery) for students receiving NAN during health class (implementation group) and students in physical education classes who did not receive NAN (control group). We hypothesized that students in the implementation groups who received the NAN program would see the following changes: (a) improved awareness of and attitudes toward the problem of CSE; (b) increased help-seeking capacity; and (c) increased safety planning behaviors.

Methods

Participants

Two Houston-area high schools agreed to participate in the study after the research team reached out to several schools in the district. Participation by both schools was approved by the Houston Independent School District (HISD). The first high school (HS1) serves over 1,500 9th to 12th grade students. Forty-nine percent of students are Black and 49% are Latino, and approximately 90% of students are classified as being from socioeconomically disadvantaged backgrounds. The second participating high school (HS2) also serves approximately 1,500 9th to 12th grade students. Ninety percent of students from this high school identify as Latino, with 96% coming from socioeconomically disadvantaged backgrounds.

Both schools agreed to facilitate pretest and posttest survey administrations for students in health and physical education classes. Students in health classes received the NAN program (3rd edition) following pretest survey administration and served as the implementation group; students in physical education classes did not receive the NAN program and served as the control group. For HS1, the study was conducted in spring 2022 and again in fall 2022. For HS2, the study was conducted in fall 2022. A total of 1,319 students were eligible to participate (implementation group n = 513; control group n = 806) (See CONSORT flow chart in Figure 1) across the three administrations. Pretest surveys were taken by a total of 597 students (45%). Signed caregiver consent was required for student participation, and this was achieved for a higher percentage of youth in implementation groups (73%, n = 375) versus control groups (24%, n = 222). Posttest surveys were completed and matched for a total of 449 students (74%). Retention rates were slightly higher for implementation groups (80%, n = 300) than for control groups (67%, n = 149).

Figure 1.

CONSORT flow chart.

Sample Demographic Information

Table 1 provides data on sample demographics across implementation and control groups for the final sample of 449 students. Across the full sample, the greatest percentage of students were in the 12th grade (44.7%), with 22.8% in 11th grade, 18.3% in 10th grade, and 14.1% in 9th grade. The sample was fairly evenly split between female (47.4%) and male students (51.7%), with a small percentage (<1%) identifying as transgender or another gender. The majority of the sample described their sexual orientation as heterosexual (78.8%), with 21.1% of youth describing other sexual orientations (e.g., homosexual, bisexual, asexual, or other minority sexual orientations).

Table 1.

Youth Demographic Information by Implementation Group.

Youth Characteristics	Full Sample N = 449% (n)	Control n = 149% (n)	Implement. n = 300% (n)	X²/t	p
Grade
9^th	14.1 (63)	35.1 (52)	3.7 (11)	206.55	<.001
10^th	18.3 (82)	39.2 (58)	8.0 (24)
11^th	22.8 (102)	22.3 (33)	23.1 (69)
12th	44.7 (200)	3.4 (5)	65.2 (195)
Gender
Female	47.4 (213)	53.7 (80)	44.3 (133)	7.34	.03
Male	51.7 (232)	44.3 (66)	55.3 (166)
Other gender	0.9 (4)	2.0 (3)	0.3 (1)
Sexual orientation
Heterosexual	78.8 (354)	72.5 (108)	82.0 (246)	5.40	.02
Other sexual orientation	21.1 (95)	27.5 (41)	18.0 (54)
Race/ethnicity
Asian	1.8 (8)	3.3 (5)	1.0 (3)	3.16	.07
Black/African American	35.0 (157)	32.9 (49)	36.0 (300)	0.42	.51
Hawaiian/Pacific Islander	0.4 (2)	0.7 (1)	0.3 (1)	0.26	.61
Latino/Hispanic	65.3 (293)	63.1 (94)	66.3 (199)	0.46	.50
Native American	1.3 (6)	1.3 (2)	1.3 (4)	0.01	.99
White	3.6 (16)	8.7 (13)	1.0 (3)	17.29	<.001
Other race/ethnicity	1.3 (6)	1.3 (4)	1.3 (2)	0.01	.99
Spent most nights (past 30 days)
Parent/guardian home	96.2 (432)	94.6 (141)	97.0 (291)	1.53	.22
Other housing	3.8 (17)	5.4 (8)	3.0 (9)
Adults living w/ student
Bio/adoptive mother	82.6 (371)	76.5 (114)	85.7 (257)	5.81	.02
Bio/adoptive father	49.7 (223)	45.0 (67)	52.0 (156)	1.97	.16
Grandparents	10.9 (49)	13.4 (20)	9.7 (29)	1.44	.23
Foster parents	1.1 (5)	2.0 (2)	1.0 (2)	1.64	.20
School connectedness (mean, SD)	3.14 (.84)	3.12 (.07)	3.15 (.05)	–0.28	.77

Note. Bold numbers highlight p-values significant at p < .05.

Most students in the sample described their race/ethnicity as Latino or Hispanic (65.3%) or Black/African American (35%), with less than 5% of the sample identifying with any of the following race/ethnic groups: Asian, Hawaiian/Pacific Islander, Native American, White, or “other” race/ethnicity. The majority of students spent most nights of the last month living in their parents’ or guardians’ home (96.2%). Adults living with the students included biological/adoptive mothers (82.6%), fathers (49.7%), grandparents (10.9%), and foster parents (1.1%).

Baseline Demographic Differences between Groups

Table 1 also provides baseline data on demographic differences between youth in implementation and control groups. There were significant differences in grade level for students in the two groups. Most students taking physical education (control group) were split between 9th (35.1%), 10th (39.2%), and 11th grades (22.3%), with only a small percentage (3.4%) of 12th graders. The majority of students in the implementation groups, on the other hand, were 12th graders (65.2%), with about a quarter in 11th grade (23.1%), and smaller percentages of students in 10th (8%) or 9th grade (3.7%). There were also small differences in gender distribution in groups with the control groups having a higher percentage of female students (53.7%) compared to the implementation groups (44.3%). Implementation groups also had slightly higher percentages of heterosexual students (82% vs. 72.5%), fewer White students (1% vs. 8.7%), and more students living with their biological or adoptive mothers (85.7% vs. 76.5%) compared to the control groups.

Procedures

Prior to pretest administration, caregivers of students in all schools were sent a letter that provided detailed information on the project and the option to complete either a paper or online consent form. The Houston school district requires active consent for student participation in research, so only youth with signed parental consent for participation were included in the study. Survey data were collected at two times: (a) before the intervention; and (b) three months after pretest administration (approximately 2 months after program administration for implementation groups). Data were collected via a secure online survey that is hosted by the University of New Hampshire. Students accessed the survey through a web address. Prior to pretest and posttest administrations, the on-site research assistants (who were members of the Love146 staff but not program facilitators) read an assent form verbally to the youth. Students who did not have parental consent to participate did other computer work. Online surveys were offered in both English and Spanish and completed during school hours. Surveys took the students 30 to 40 min to complete, on average. All procedures were approved by the Institutional Review Board (IRB) at the University of New Hampshire. Each participating school was provided with a $1,000 stipend for the help they provided with planning and data collection.

Facilitators taught the NAN program to students starting the week that the students completed the pretest survey. The timing of the delivery of the five NAN lessons varied across the three implementation settings with full implementation lasting from 5 to 8 weeks.

Measurement

The pre and posttest surveys collected information on student demographics, hypothesized program outcomes, and potential mediator and moderator variables. Below, we describe these measurement categories in more detail.

Hypothesized Program Outcomes

Attitudes toward CSE

The pre and postprogram surveys had five questions related to students’ attitudes toward CSE (see Table 2). Youth were asked to indicate their agreement with each item on a scale from 1 (strongly disagree) to 4 (strongly agree). For three of the items, the desired response was reversed (disagreement reflected greater awareness). Scores for these items were reversed for regression analyses. The estimated Cronbach’s alpha internal consistency reliability for the five CSE attitude items was very small (.04) for the pretest. Given the lack of correlations of items within the composite, subsequent analyses were conducted with individual items only.

Table 2.

Pre and Posttest CSE Mean Attitudes Scores by Group.

CSE Attitudes	Pretest			Posttest
CSE Attitudes	Control n = 149 mean (SD)	Implement. n = 300 mean (SD)	t	Control n = 149 mean (SD)	Implement. n = 300 mean (SD)	t
It can be hard for people who trade sex for money or other items of value to stop even if they want to.	2.68 (0.84)	2.72 (0.78)	–0.54	2.78 (0.82)	2.85 (0.78)	–0.88
Physical force is usually the way that someone gets a youth to trade sex for money or other items of value. (R)	2.61 (0.87)	2.78 (0.79)	–1.92	2.76 (0.86)	2.88 (0.76)	–1.54
Youth who choose to trade sex for money or other items of value are victims.	2.76 (0.85)	2.75 (0.87)	0.10	2.87 (0.81)	2.99 (0.81)	–1.52
If someone decides to trade sex for money or other items of value, it is their fault if bad things happen. (R)	2.67 (0.89)	2.73 (0.83)	–0.64	2.74 (0.85)	2.48 (0.88)	2.86**
Trading sex for things of value is a harmless way to make some money. (R)	1.73 (0.85)	1.84 (0.89)	–1.15	1.94 (0.94)	1.78 (0.91)	–1.73

Note: Scores range: 1 to 4; (R) identifies items in which disagreement (a lower score) is the desired response.

p < .01.

Help-seeking Capacity for CSE

Youth were asked six questions about what they would do if they were “feeling pressured by somebody to exchange sex for money or other items of value.” Questions were asked how likely they would be to: (a) get help from a parent or caregiver, (b) get help from another adult like a teacher or counselor, (c) get help from a friend, (d) look up advice on the internet, (e) get help from a crisis hotline or community agency, or (f) call 911/talk to police (see Table 3). A four-point scale was offered for response options (1 = very likely, 2 = somewhat likely, 3 = somewhat unlikely, and 4 = very unlikely). All items were reverse coded for analyses, so larger numbers represent a higher likelihood of getting help. The estimated Cronbach’s alpha internal consistency reliability for the six items was .79 at pretest and .83 at posttest, reflecting the moderate positive correlations between all items in the composite. Given the high reliability, subsequent analyses were conducted with the full composite.

Table 3.

Pre and Posttest Help-seeking Likelihood Mean Scores by Group.

	Pretest			Posttest
Help-seeking Likelihood	Control n = 149 mean (SD)	Implement. n = 300 mean (SD)	t	Control n = 149 mean (SD)	Implement. n = 300 mean (SD)	t
Likelihood of helping seeking for CSE from:
Parent/caregiver	3.45 (0.91)	3.42 (0.91)	0.34	3.29 (1.01)	3.33 (0.97)	–0.40
Another adult like teacher/counselor	2.87 (1.11)	2.80 (1.12)	0.59	2.82 (1.14)	2.94 (1.05)	–1.06
A friend	3.37 (0.88)	3.15 (0.99)	2.16*	3.22 (1.07)	3.26 (0.05)	–0.44
The internet	2.38 (1.14)	2.49 (1.09)	–1.00	2.29 (1.08)	2.72 (1.11)	–3.87***
A crisis hotline/community agency	2.48 (1.19)	2.63 (1.12)	–1.25	2.40 (1.20)	2.85 (1.11)	–3.80***
911/police	2.99 (1.19)	2.97 (1.14)	0.16	2.71 (1.25)	3.02 (1.07)	–2.64**
Help-seeking likelihood mean score	2.93 (0.71)	2.91 (0.77)	0.25	2.79 (0.80)	3.03 (0.78)	–2.97**

Note. Scores range: 1 to 4.

p < .05. **p < .01. ***p < .001.

Safety Planning for CSE Risk

Assessment of safety planning was covered with three questions measuring improved self-efficacy in responding to risks, and two questions about online safety behaviors. Specifically, three questions were asked to measure help-seeking safety planning including: (a) “I know of hotline, or specific place, like a crisis center, that I could contact if I was in a dangerous situation or relationship;” (b) “I would know how to help someone who was in a situation where they were exchanging sex for money or other items of value;” and (c) “I would know what to do if someone was pressuring me to exchange sex for money or other items of value” (see Table 3). A four-point scale was offered for response options (1 = strongly agree, 4 = strongly disagree). All three items were reverse scaled for analyses, so a higher number represents more knowledge of helping resources. The composite formed by averaging these items has an estimated alpha reliability of .54 at pretest and .64 at posttest. Given the relatively low reliability, analyses were conducted with the individual items. Additionally, two questions asked about online safety behaviors: “In the last 3 months, have you checked your social media privacy settings to make sure they were private?” and “In the last 3 months, have you blocked or stopped talking with someone online because they were making you feel uncomfortable” (see Table 4). Response options to all questions were yes/no. Analyses were conducted with summary variables indicating a positive response to at least one of the questions about online safety behaviors in the last 3 months.

Table 4.

Pre and Posttest Safety Planning Mean Scores by Group.

	Pretest			Posttest
Safety Planning Measures	Control n = 149	Implement. n = 300	t/Χ²	Control n = 149	Implement. n = 300	t/Χ²
Knowledge of helping resources mean (SD)
Know of a hotline or crisis center I could contact.	2.50 (0.89)	2.41 (0.95)	0.99	2.56 (0.95)	2.41 (0.95)	–3.63***
Know how to help someone who was exchanging sex for money or other items of value.	2.74 (0.84)	2.64 (0.81)	1.14	2.79 (0.77)	2.64 (0.81)	–0.68
Know what to do if someone was pressuring me to exchange sex.	3.16 (0.80)	3.19 (0.80)	–0.41	3.14 (0. 84)	3.19 (0.80)	–2.01*
Online safety behaviors (past 3 months) % (n)
Checked social media privacy settings to make sure they were private.	34.9% (n = 52)	27.0% (n = 81)	2.98	49.0% (n = 73)	27.0% (n = 81)	0.28
Blocked or stopped talking to someone online because they made you feel uncomfortable.	24.8% (n = 37)	25.0% (n = 75)	0.01	36.2% (n = 54)	25.0% (n = 75)	0.01

p < .05. ***p < .001.

Youth Characteristics

Demographic information was collected on students’ grade level (9th–12th grade). Racial/ethnic identity was collected across eight categories (Asian; Black/African American; Hawaiian/Pacific Islander; Latino/Hispanic; Native American; White; and other race/ethnicity). Youth also reported on gender identity (male, female, transgender, nonbinary/gender fluid, or other) and on sexual orientation (heterosexual/straight, gay/lesbian/homosexual, bisexual, asexual, pansexual, queer, not sure, or other). Information was also collected on the students’ living situation including where they spent most of their nights over the past 30 days (dichotomized to (a) parent or guardian’s house and (b) other living situations). Students were also asked which of the following adults lived with them: biological or adoptive mother, biological or adoptive father, foster parents, grandparents, other adults, or no adults. Finally, students were also asked items adapted from the School Connectedness Scale (Lohmeier & Lee, 2011) (alpha = .82). A mean score was calculated from a five-item scale (1 = strongly disagree to 5 = strongly agree) that asked questions like: “I feel happy to be at this school.”

Statistical Analysis

Analyses

Analyses for this paper were conducted using Stata statistical software, version 17.0 and SPSS version 29 (IBM Corp). Descriptive statistics were estimated for all demographic variables and outcome variables at both pretest and posttest. As noted above, given the low estimated reliabilities for composites formed from the items measuring CSE attitudes and safety planning, analyses were conducted using individual items. The composite was used for CSE help seeking. Next, difference scores between pretest and posttest were computed for the continuous outcomes, and differences between implementation and control groups were assessed. Finally, OLS regression models were fitted for continuous outcomes (CSE attitude items, likelihood of help-seeking composite, safety planning items measuring self-efficacy in responding to risks), and logistic regression models were fitted for dichotomous outcomes (online safety behaviors). These analyses estimated uncontrolled and controlled effects of the intervention on pre–post differences and whether program effects differ by student demographics.

Missing Data

Missing data for study outcome variables ranged from less than 1% to 4%. Given the low number of missing values, listwise deletion was used for all analyses.

Results

Group Differences at Pretest

Bivariate analyses were first conducted on questions about CSE attitudes for implementation and control groups (see Table 2). There were no statistically significant differences between groups on questions at pretest. There were significant differences at posttest for one question: “If someone decides to trade sex for money or other items of value, it is their fault if bad things happen” (t = 2.86, p = .004). While the implementation group improved in the desired direction on this question over time (decreased endorsement), the control groups moved away from the desired direction (increased endorsement) from pretest to posttest.

We also examined bivariate group differences in help-seeking likelihood for CSE and safety planning (see Table 3). At pretest, there were almost no differences between groups in the likelihood of seeking help from a variety of sources or of helping resources. The only difference was a slightly higher likelihood of seeking help from a friend by youth in control groups versus implementation groups (t = 2.16, p = .03). However, at posttest there were a number of significant differences between groups with implementation groups indicating a higher likelihood of seeking help from the internet (t = −3.87, p < .001), a crisis hotline/community agency (t = −3.80, p < .001), or 911/police (t = −2.97, p = .009) compared to the control groups. The overall help-seeking likelihood scale score was also significantly different for the groups, with higher scores for youth in implementation groups versus control groups (t = −2.97, p = .003).

There were no differences at pretest between groups in any of the safety planning measures (see Table 4). At posttest, more youth in implementation groups endorsed that they “know of a hotline or crisis center they could contact” (t = −3.63, p < .001) and “know what to do if someone was pressuring me to exchange sex” (t = −2.01, p = .045) than in control groups.

Regression Analyses: Program Impact

Using OLS regression and controlling for demographic variables, we found statistically significant effects of NAN implementation on responses to two of the CSE attitude questions. Statistically significant program effects were also found for the help-seeking likelihood composite and knowledge of hotline/crisis center helping resources. There were small average improvements in these outcomes for participants in the NAN program. No effects of NAN exposure were found for other safety planning measures. Effects are not systematically moderated or mediated by the covariates. More detailed information is presented below and in Table 5.

Table 5.

Parameter Estimates and Standard Errors From Fitted OLS Regression Models Estimating Effect of NAN Implementation on Each Outcome (Uncontrolled, Controlled).

Outcome	Effect of NAN Implementation		Effect of Control Variable
Outcome	Uncontrolled Est-β (SE) n	Controlled Est-β (SE) n	Control variable Est-β (SE)
CSE attitudes: fault (difference)	0.295** (0.102)438	0.272** (0.103)438	Dichotomous sexual orientation0.222~(0.09)
CSE attitudes: harmless (difference)	0.251* (0.113)439	0.279* (0.114)435	Dichotomous gender–0.213* (0.107)
CSE help-seeking composite (difference)	0.241** (0.083)430	0.237** (0.084)418	School connectedness–0.091* (0.046)
Safety planning: hotline awareness (difference)	0.386** (0.115)421	0.403** (0.116)418	Dichotomous gender–0.251* (0.107)

Note. OLS = ordinary least squares; NAN = not a number; CSE = commercial sexual exploitation.

p < .05. **p < .01.

CSE Attitudes

There was a statistically significant treatment effect on pre to posttest changes in response to the statement: If someone decides to trade sex for money or other items of value, it is their fault if bad things happen (reverse coded for analysis) (p < .01). The average difference for NAN participants was estimated to be 0.295 higher than the average difference for the control group (p < .05). Compared with average control group participants, youth in the implementation group tended to express less agreement with this statement. The only covariate that impacted responses about fault was sexual orientation (p = .059). The positive parameter estimate suggested that students identifying as heterosexual had a slightly larger change in their agreement with this statement than students who identified with nonheterosexual orientations, regardless of whether they were in the implementation or control group. The implementation effect decreased slightly when controlling for sexual orientation, but it was still statistically significant.

There was also a significant treatment effect on the pre to posttest changes in response to the statement: Trading sex for things of value is a harmless way to make some money. Since this item was also reverse coded, higher scores indicated less agreement with the statement. For the average NAN participant, the estimated pretest to posttest difference was 0.251 larger than for the average control group member (p < .05). The average control group participant had a decrease of about .2 in their agreement with this item, and the decrease was statistically significant (p < .05). For the implementation group, the average difference was 0.251 higher than for the average control group member (p < .05), but the average difference for the NAN group was very small (less than 0.06). This is consistent with the nonsignificant difference in the posttest means for this item presented in Table 2. There was also a statistically significant effect of gender (dichotomized), but no interaction between gender and group. The average difference in agreement with this item was 0.213 smaller for females than for males (regardless of group) (p < .05). The treatment effect increased slightly when controlling for gender.

CSE help-seeking composite

There was a statistically significant effect of program exposure on questions about the likelihood of help seeking between pretest and posttest (p < .01). The average difference for NAN participants was estimated to be 0.241 higher than the average difference for the control group. The positive parameter estimate suggested that the average NAN participant has about a 1/4 of a point increase in their reported likelihood of help seeking, compared with the average control participant. There was also a statistically significant impact of school connectedness on help seeking. Regardless of group, higher levels of school connectedness were associated with smaller pre–post changes in self-reported likelihood of seeking help. The treatment effect was slightly lower when controlling for school connectedness, but still statistically significant (p < .01).

Safety Planning

Of the three questions that aimed to capture increased safety planning, there was a significant effect of the program on the difference score for one item: I know of a hotline or specific place, like a crisis center, that I could contact if I was in a dangerous situation or relationship (reverse coded for analysis) (p < .01). On average, the implementation group had an average improvement of 0.487 in reported knowledge of hotline helping resource, compared with 0.101 for the control group (p < .01). The average pretest to posttest difference for the NAN group was 0.386 larger than for the control group. Controlling for program exposure, there was an effect of dichotomous gender on the average change in reported hotline knowledge (p < .05). The average change for females was approximately 0.25 lower than for males, regardless of group. In addition, controlling for gender, the effect of program exposure was slightly larger (0.403, p <.01).

Finally, no effects of program exposure were found on the probability that a participant reported either of the online safety behaviors over the prior 3 months. Both simple and multiple logistic regression models were fitted, with no significant effect of treatment or covariates.

Discussion

CSE education programs like the Not a Number (NAN) curriculum approach prevention by focusing on improving awareness of CSE dynamics, improving attitudes toward youth who experience CSE (e.g., reduced victim blaming), and improving help-seeking self-efficacy and safety planning. As programs to educate youth about CSE continue to develop, it is critical for evaluation to accompany program growth and help direct educational practices. The evaluation results of the current study found important initial support for several hypotheses including evidence of changed attitudes toward CSE, increased knowledge of helping resources like hotlines, and increased self-efficacy for help-seeking across a number of important resources. The research presented here was conducted as one part of a series of rigorous evaluations being conducted on the NAN program, and as is often typical with quasi-experimental designs, had limitations in power and measurement noted in more detail below. Nonetheless, the findings represent the first study of a prevention education program directed at CSE using a comparison group that we are aware of and provide important initial indications that CSE education programs like NAN can improve youths’ awareness of CSE and their help-seeking capacity. More evaluation research is needed to improve measurement, examine additional CSE program outcomes, extend follow-up periods, and examine the long-term outcomes on risk experiences for youth who participate in CSE prevention programs.

CSE Attitudes

Study results indicated that youth lack important information about areas relevant to CSE and have a number of misperceptions about the dynamics of this type of victimization risk. The findings from the study found that the NAN program provided them with improved information about trafficking that was maintained at least 2 months after the program had been delivered. Youth who received the NAN program were more likely to reject the idea that CSE is harmless and that victims are at fault. Minimization of CSE risks and victim blaming are arguably two of the most important barriers to improving problem recognition and help-seeking for victims of CSE (Menaker & Franklin, 2017; Prior et al.,2023). Given the historical treatment of trafficking victims and mislabeling youth CSE victims as offenders, victim blaming creates barriers to holding actual offenders accountable and directing victims to needed services (Wiener et al., 2021) and can discourage victims from seeking help. Lack of awareness of the harm caused by CSE has been identified as a barrier to proper identification and help-seeking among professionals and survivors (Peck et al., 2021). Increasing awareness about CSE harms and reducing victim blaming is important for the goal of improving CSE social norms and increasing peer support for youth exposed to CSE (Buller et al., 2020). These two issues have been a focus of sexual assault and dating violence prevention programs for decades (Jago & Christenfeld, 2018).

The study did not find program effects for some items measuring CSE awareness. It was concerning that youth continued to misperceive trafficking as typically involving physical violence. Research has found that youth endorse similar myths related to gender-based violence (Rollero & De Piccoli, 2020). Baseline responses to other items also suggest some ambivalence around victim blaming. Youth attitudes about CSE may be complicated by stereotypical portrayals in popular media: media narratives generally portray CSE as a female youth who is kidnapped and physically forced into unwanted sexual activities to provide money to a pimp (Bromfield, 2016). Prevention work should continue to incorporate program content that counters misleading stereotypes about trafficking that might increase youth risk. It is also important to note that our efforts to measure a general CSE awareness construct with the five items were not successful. The absence of shared variance across the items developed to measure CSE attitudes for this study suggests that youth’s understanding of CSE dynamics is complex and sometimes contradictory and varies in ways that needs to be better understood. There is a critical need for the field to develop measures with good psychometric strength that can effectively capture youth attitudes toward and awareness of CSE.

Help-seeking Self-Efficacy

Awareness of the dynamics of CSE is a critical foundation for improved help seeking. The Children’s Help-Seeking and Service Utilization Model (Srebnik et al., 1996), which is used widely in health research to consider trajectories of help seeking among youth (Park & Kwon, 2018), notes that problem identification is a necessary first step in the decision to seek help. Once a problem is recognized (step 1), the help-seeking model suggests that youth consider whether they should seek help (step 2) before ultimately contacting a helping resource (step 3). Application of the help-seeking model to youth impacted by CSE appears promising (O’Brien et al., 2024). An analysis of preprogram survey data collected by the NAN program suggests that prior to program delivery, a better understanding of CSE dynamics is one of the strongest predictors of the likelihood of seeking help (Jones et al., 2024).

Results from the current study suggest that for youth receiving the NAN program, one of the strongest program outcomes was improvements in students’ self-reported help-seeking self-efficacy. Youth exposed to the NAN program reported a significantly greater likelihood of seeking help from a number of different sources for CSE than nonexposed youth. Help-seeking self-efficacy and intention are likely to improve the chance that youth who are exposed to CSE or CSE risk will get help more quickly. Research on help-seeking self-efficacy and intention finds a significant relationship between intention to seek help and actual help-seeking behavior (Padilla-Medina et al., 2022; Webb & Sheeran, 2006). While it was outside of the scope of the current study to long-term help-seeking by youth who experienced CSE or other exploitation, previous research has identified help-seeking efficacy as an important aspect of primary prevention and recidivism reduction (De La Rue et al., 2017; Wong et al., 2021).

Safety Planning

One of the objectives of the NAN program is not only to increase youth’s understanding of the dynamics of CSE and their intention to seek help in cases of risk or victimization, but also to provide them with specific options for accessing help. This is done in a number of ways such as giving youth information about resources available to help them in the face of CSE victimization or risk, such as hotlines and crisis centers; helping youth identify trusted supports in their life who could help them if needed; and providing youth with ways to reduce their online risks. One of the most powerful evaluation outcomes for the program was an increase in the youths’ awareness of the hotline and crisis centers available to them. Suicide prevention education efforts have found similar success for postprogram impact on hotline awareness and use (Freedenthal, 2010), and research on hotlines and helplines generally shows that youth find them helpful (Mathieu et al., 2021). Not only are youth more likely to add these resources to their help-seeking options, but they may also share this knowledge with peers, parents, or other community members.

Other findings for the impact of NAN on youth safety planning were less conclusive. We did not find a significant impact of the program on two questions that were designed to capture whether youth would know what to do if they or someone else was experiencing CSE. However, it is possible that the questions did not successfully capture the degree to which youth felt they had options for handling CSE issues or risks. Additionally, there were no differences in program impact for online safety behaviors. Unfortunately, online safety education in general has been underevaluated (Jones & Mitchell, 2016), and the studies that have been conducted suggest that education programs have a difficult time changing youth online behavior (Jones et al., 2023). More research is needed to understand how violence prevention programs generally, and trafficking prevention specifically, can effectively incorporate information about online risks and protections.

Limitations

Findings from the current study should be considered in light of study limitations. Quasi-experimental designs are limited in power, and the implementation and control groups in the current study had significant preexisting demographic differences including age. We adjusted for these differences by statistically controlling for demographic and other variables in regression models. Nonetheless, there may have been group differences related to age that still impacted findings in ways we were not able to measure, such as literacy. Additionally, the implementation and control groups were from the same high schools, and there may also have been some dilution of program’s impact due to information sharing between the groups. Given the absence of rigorous outcome data for CSE prevention education (e.g., Rizo et al., 2019), the measurement tools used were created for the current study, and psychometric data suggested that our efforts to measure some of our outcomes may not have been optimal. In particular, as noted above, efforts to capture a general overall “CSE awareness” composite scale through a five-item scale was not successful. Additional psychometric work and research will be needed to improve these efforts.

Other factors may have further impacted study outcomes. There were some issues with program implementation being disrupted by school events (e.g., fire alarms, canceled classes). Postprogram follow-up was limited in time, and some critical long-term outcomes such as help-seeking behaviors and reductions in victimization were outside the purview of the current study. Finally, while the study has important hypothesis-building implications for understanding the impact of CSE and trafficking prevention education, the study was focused on one program, at two high schools in one U.S. city, and results may not be generalizable to other programs or other youth populations. However, it is worth noting the large percentage of Latino and Black students participating in this pilot. Given that minoritized and marginalized youth are disproportionately represented among CSE victims, evaluation research that includes these youth is an important priority.

Implications for Research and Practice

Findings from the current study offer encouraging support for the impact of the NAN program and similar approaches to youth education about CSE. Providing youth with information about CSE is a critical goal, and educational efforts are likely to continue to expand. The increase in development of CSE prevention is an important response to concerns about sexual exploitation of youth that are building in communities (Palmer et al., 2021). Practitioners, advocacy organizations, and service providers who are developing trafficking and CSE educational content can draw from this study’s findings as they create and revise curricula. Results emphasize the value of incorporating a focus on help-seeking into prevention programs and the importance of countering misperceptions that youth have about CSE. It also suggests both the challenge and the importance of incorporating online safety information into youth victim prevention education. Technology has become a tool for traffickers and exploiters (Finkelhor et al., 2022; O’Brien & Li, 2020), and communities are eager to find ways to help youth protect themselves, but research identifying effective messaging or strategies has been limited.

It is particularly critical that more evaluation research accompanies program development and education for youth on CSE and guides best practices. The current study improves on existing evaluation efforts by including a control group and analyzing changes in short-term preprogram versus postprogram outcomes. Accordingly, the current study represents an important advance in building this evidence base. However, findings need to be interpreted as one part of a much larger body of research and evaluation on CSE prevention education. This study itself builds on prior research conducted with the NAN program (Jones et al., 2024) and is being used to inform a more rigorous evaluation (Martin et al., 2025). Bullying prevention evaluation research began in earnest in the 1980s (Lane, 1989), and four decades later, we now have enough rigorous evaluation research to draw conclusions in general about effectiveness (Gaffney et al., 2021; Hensums et al., 2023). It is important to begin building toward a similar understanding for CSE prevention education, with more evidence on the impact of a variety of educational approaches, with diverse populations of youth, in both school and nonschool settings.

Additionally, in preparation for more evaluation research, survey instruments must be validated for measuring CSE-related outcomes such as awareness, knowledge, risk, victimization, needs, and well-being. Based on the data collected in this pilot, the research team has been working on extensive revisions of the measurement model used here, improving on outcome measurement efforts using community-based participatory research models and extensive pilot testing to develop broader measurement tools with strong psychometric structure (Martin et al., 2025).

Finally, as evaluation research on trafficking and CSE-focused prevention education builds, it is important that it place in the context of other violence prevention efforts. CSE often co-occurs with other victimization such as sexual abuse, sexual assault, and dating violence, and overlaps with other significant adversity risks for youth including running away and drug use (Gibbs et al., 2018; Newcomb, 1995; Sanchez et al., 2006; Wang & Fredricks, 2014). Structural inequality, economic disparities, sexism, and racism all contribute as causal factors as well. Successful prevention of CSE will require multiple and coordinated approaches accompanying evidence-based CSE prevention education.

Conclusions

The current study contributes important evaluation evidence to what we hope will be a growing effort to identify and document the efficacy of prevention education on commercial sexual exploitation (CSE) of youth. This quasi-experimental evaluation of the Not a Number (NAN) trafficking prevention program suggests that, as a result of NAN program exposure, youth learn important new knowledge and increase their help-seeking options and self-efficacy. While findings are promising for CSE prevention education approaches, additional evaluation research of other CSE prevention education programs is needed to ensure that efforts aimed at preventing youth from experiencing CSE are effective.

Footnotes

Acknowledgements

The authors would like to thank Aria Flood, Sarah Mouser, and other members of the Love146 team for supporting the research team during this study.

Research materials can be accessed for review by contacting the lead author

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by a research grant from Love146 to the University of New Hampshire. The funders had no role in the preparation of the manuscript.

ORCID iDs

Lisa M. Jones

Jennifer E. O’Brien

Author Biographies

Lisa M. Jones, PhD, is a research associate professor of psychology at the Crimes against Children Research Center (CCRC), University of New Hampshire (UNH). She has over 20 years of experience researching child victimization and adversity exposure, with a focus on commercial sexual exploitation of children (CSEC), sexual abuse, bias and hate crime victimization, and technology-based victimization of children.

Jennifer O’Brien, PhD, is an associate professor at the University of Texas at Arlington School of Social Work and a research affiliate at the Crimes against Children Research Center (CCRC) at the University of New Hampshire. Her areas of research include childhood sexual abuse, the commercialized sexual exploitation of children (CSEC), and Internet victimization. A licensed clinician, Dr. O’Brien’s research interests stem from her practice experiences both domestically and abroad with survivors of sexual abuse and trauma.

Suzanne E. Graham, EdD, is an associate professor in the Department of Education at the University of New Hampshire. Her research interests include longitudinal data analysis, multilevel modeling, causal inference, and mathematics achievement.

References

Bandura

(2002). Social cognitive theory in cultural context. Applied Psychology, 51(2), 269–290.

Barnert

Iqbal

Bruce

Anoshiravani

Kolhatkar

Greenbaum

(2017). Commercial sexual exploitation and sex trafficking of children and adolescents: A narrative review. Academic Pediatrics, 17(8), 825–829.

Bromfield

N. F.

(2016). Sex slavery and sex trafficking of women in the United States: Historical and contemporary parallels, policies, and perspectives in social work. Affilia, 31(1), 129–139.

Buller

A. M.

Pichon

McAlpine

Cislaghi

Heise

Meiksin

(2020). Systematic review of social norms, attitudes, and factual beliefs linked to the sexual exploitation of children and adolescents. Child Abuse & Neglect, 104, 104471.

Cole

Sprang

(2015). Sex trafficking of minors in metropolitan, micropolitan, and rural communities. Child Abuse & Neglect, 40, 113–123.

Cole

Sprang

Lee

Cohen

(2016). The trauma of commercial sexual exploitation of youth: a comparison of CSE victims to sexual abuse victims in a clinical sample. Journal of Interpersonal Violence, 31(1), 122–146.

Committee for Children. (2021). CSEC Education: What is CSEC? Who is at risk? How can education help prevent it? https://www.cfchildren.org/policy-papers/csec-education/

De La Rue

Polanin

J. R.

Espelage

D. L.

Pigott

T. D.

(2017). A meta-analysis of school-based interventions aimed to prevent or reduce violence in teen dating relationships. Review of Educational Research, 87(1), 7–34.

DeGue

Niolon

P. H.

Estefan

L. F.

Tracy

A. J.

V. D.

Vivolo-Kantor

A. M.

Little

T. D.

Latzman

N. E.

Tharp

Lang

K. M.

(2021). Effects of Dating Matters^® on sexual violence and sexual harassment outcomes among middle school youth: A cluster-randomized controlled trial. Prevention Science, 22, 175–185.

10.

Edinburgh

Pape-Blabolil

Harpin

S. B.

Saewyc

(2015). Assessing exploitation experiences of girls and boys seen at a child advocacy center. Child Abuse & Neglect, 46, 47–59.

11.

Finkelhor

Turner

Colburn

(2022). Prevalence of online sexual offenses against children in the US. JAMA Network Open, 5(10), e2234471.

12.

Finkelhor

Vaquerano

Stranski

(2017). Sex trafficking of minors: How many juveniles are being prostituted in the US? https://www.unh.edu/ccrc/sites/default/files/media/2022-03/sex-trafficking-of-minors-how-many-juveniles-are-being-prostituted-in-the-us.pdf

13.

Franchino-Olsen

(2021). Vulnerabilities relevant for commercial sexual exploitation of children/domestic minor sex trafficking: A systematic review of risk factors. Trauma, Violence, & Abuse, 22(1), 99–111.

14.

Freedenthal

(2010). Adolescent help-seeking and the yellow ribbon suicide prevention program: An evaluation. Suicide and Life-Threatening Behavior, 40(6), 628–639.

15.

Gaffney

Ttofi

M. M.

Farrington

D. P.

(2021). Effectiveness of school-based programs to reduce bullying perpetration and victimization: An updated systematic review and meta-analysis. Campbell Systematic Reviews, 17(2), e1143.

16.

Garg

Panda

Neudecker

Lee

(2020). Barriers to the access and utilization of healthcare for trafficked youth: A systematic review. Child Abuse & Neglect, 100, 104137.

17.

Gibbs

D. A.

Henninger

A. M.

Tueller

S. J.

Kluckman

M. N.

(2018). Human trafficking and the child welfare population in Florida. Children and Youth Services Review, 88, 1–10.

18.

Greenbaum

V. J.

(2017). Child sex trafficking in the United States: Challenges for the healthcare provider. PLoS Medicine, 14(11), e1002439.

19.

Greenbaum

V. J.

Dodd

McCracken

(2018). A short screening tool to identify victims of child sex trafficking in the health care setting. Pediatric Emergency Care, 34(1), 33–37.

20.

Head

S. K.

Eaton

Lloyd

P. C.

McLaughlin

Davies-Cole

(2021). Exchange sex among high school students—Washington, DC, 2017. Journal of Adolescent Health, 68(2), 350–356.

21.

Hensums

De Mooij

Kuijper

S. C.

Fekkes

Overbeek

(2023). What works for whom in school-based anti-bullying interventions? An individual participant data meta-analysis. Prevention Science, 24(8), 1435–1446.

22.

Human Trafficking and Exploitation Prevention Training Act, S. 3333,116th Cong. (2021). https://www.murkowski.senate.gov/imo/media/doc/Human%20Trafficking%20and%20Exploitation%20Prevention%20Training%20Act%20of%202021.pdf

23.

Jago

C. P.

Christenfeld

(2018). The impact of sexual assault prevention information on perceived control and victim blaming. https://osf.io/preprints/psyarxiv/m8qwc

24.

Jones

L. M.

Mitchell

K. J.

(2016). Defining and measuring youth digital citizenship. New Media & Society, 18(9), 2063–2079.

25.

Jones

L. M.

Mitchell

K. J.

Beseler

C. L.

(2023). The impact of youth digital citizenship education: Insights from a cluster randomized controlled trial outcome evaluation of the be internet awesome (BIA) curriculum. Contemporary School Psychology, 28, 1–15.

26.

Jones

L.M.

O’Brien

J.E.

Colburn

McKinney

Martin

(2024). Youth attitudes toward commercial sexual exploitation and help-seeking: A secondary analysis of data collected by the Not a Number trafficking prevention program. Research in Human Development, 21(2–3), 159–178. https://doi.org/10.1080/15427609.2024.2412427

27.

Lane

D. A.

(1989). Bullying in school: The need for an integrated approach. School Psychology International, 10(3), 211–215.

28.

Lederer

L. J.

Wetzel

C. A.

(2014). The health consequences of sex trafficking and their implications for identifying victims in healthcare facilities. Annals Health L., 23, 61.

29.

Lohmeier

J. H.

Lee

S. W.

(2011). A school connectedness scale for use with adolescents. Educational Research and Evaluation, 17(2), 85–95.

30.

Love146. (2020). Not A Number: Love 146’s Child Trafficking and Exploitation Prevention Curriculum (3rd ed.). Love 146.

31.

Mathieu

S. L.

Uddin

Brady

Batchelor

Ross

Spence

S. H.

Watling

Kõlves

(2021). Systematic review: The state of research into youth helplines. Journal of the American Academy of Child & Adolescent Psychiatry, 60(10), 1190–1233.

32.

Martin

Davis

Jones

L.M.

Voller

O’Brien

(2025). A mixed methods and community-based study to develop the logic model for a random controlled trial evaluation of a curriculum to prevent commercial sexual exploitation of youth. [Manuscript in preparation]. University of Minnesota, Minneapolis, MN.

33.

Martin

Rider

G. N.

Johnston-Goodstar

Menanteau

Palmer

McMorris

B. J.

(2021). Prevalence of trading sex among high school students in Minnesota: Demographics, relevant adverse experiences, and health-related statuses. Journal of Adolescent Health, 68(5), 1011–1013.

34.

Menaker

T. A.

Franklin

C. A.

(2017). Gendered violence and victim blame: subject perceptions of blame and the appropriateness of services for survivors of domestic sex trafficking, sexual assault, and intimate partner violence. In Human trafficking (pp. 117–135). Routledge.

35.

Middleton

J. S.

Gattis

M. N.

Frey

L. M.

Roe-Sepowitz

(2018). Youth experiences survey (YES): Exploring the scope and complexity of sex trafficking in a sample of youth experiencing homelessness. Journal of Social Service Research, 44(2), 141–157.

36.

Moore

J. L.

Goldberg

A. P.

Barron

(2021). Substance use in a domestic minor sex trafficking patient population. Pediatric Emergency Care, 37(4), e159–e162.

37.

Nation

Crusto

Wandersman

Kumpfer

K. L.

Seybolt

Morrissey-Kane

Davino

(2003). What works in prevention: Principles of effective prevention programs. The American Psychologist, 58(6–7), 449–456.

38.

Newcomb

M. D.

(1995). Identifying high-risk youth: Prevalence and patterns of adolescent drug abuse. NIDA Research Monograph, 156, 7–38.

39.

O'Brien

Finkelhor

Jones

(2022). Improving services for youth survivors of commercial sexual exploitation: Insights from interventions with other high-risk youth. Children and Youth Services Review, 132, 106313.

40.

O’Brien

J. E.

(2020). The role of the internet in the grooming, exploitation, and exit of United States domestic minor sex trafficking victims. Journal of Children and Media, 14(2), 187–203.

41.

O’Brien

J.E.

McKinney

Mart

Jones

L.M.

(2024). Help-seeking among children impacted by commercial sexual exploitation: A scoping review. Trauma, Violence, & Abuse, 25(5), 3571–3584. https://doi.org/10.1177/15248380241253045

42.

Padilla-Medina

D. M.

Williams

J. R.

Ravi

Ombayo

Black

B. M.

(2022). Teen dating violence help-seeking intentions and behaviors among ethnically and racially diverse youth: A systematic review. Trauma, Violence, & Abuse, 23(4), 1063–1078.

43.

Palmer

Menanteau

Glaccum

Flood

Miller

Schaefer

Cook

Orley

(2021). Partnering for human trafficking prevention: implementing Love146’s Not a Number Curriculum through Minnesota’s Safe Harbor Program. https://hphr.org/58-article-palmer/

44.

Park

Kwon

(2018). Health-related internet use by children and adolescents: systematic review. Journal of Medical Internet Research, 20(4), e120.

45.

Peck

J. L.

Meadows-Oliver

Hays

S. M.

Maaks

D. G.

(2021). White paper: recognizing child trafficking as a critical emerging health threat. Journal of Pediatric Health Care, 35(3), 260–269.

46.

Pierce

(2012). American Indian adolescent girls: vulnerability to sex trafficking, intervention strategies. American Indian and Alaska Native Mental Health Research: The Journal of the National Center, 19(1), 37–56.

47.

Piolanti

Foran

H. M.

(2021). Efficacy of interventions to prevent physical and sexual dating violence among adolescents: A systematic review and meta-analysis. JAMA Pediatrics, 176(2), 142–149.

48.

Preble

K. M.

(2021). Influence of survivors’ entrapment factors and traffickers’ characteristics on perceptions of interpersonal social power while being recruited to be trafficked. Violence Against Women, 27(8), 1091–1114.

49.

Prior

Shilo

Peled

(2023). Help-seeking and help-related experiences of commercially sexually exploited youth: A qualitative meta-synthesis. Trauma, Violence, & Abuse, 24(3), 1693–1711.

50.

Reid

J. A.

(2014). Risk and resiliency factors influencing onset and adolescence-limited commercial sexual exploitation of disadvantaged girls. Criminal Behaviour and Mental Health, 24(5), 332–344.

51.

Rizo

C. F.

Klein

L. B.

Chesworth

B. R.

O'Brien

J. E.

Macy

R. J.

Martin

S. L.

Crews

M. E.

Love

B. L.

(2019). Educating youth about commercial sexual exploitation of children: A Systematic Review. Global Social Welfare, 6(1), 29.

52.

Rollero

De Piccoli

(2020). Myths about intimate partner violence and moral disengagement: An analysis of sociocultural dimensions sustaining violence against women. International Journal of Environmental Research and Public Health, 17(21), 8139.

53.

Rothman

E. F.

Farrell

Paruk

Bright

Bair-Merritt

Preis

S. R.

(2021). Evaluation of a multi-session group designed to prevent commercial sexual exploitation of minors: The “My life my choice” curriculum. Journal of Interpersonal Violence, 36(19–20), 9143–9166.

54.

Sanchez

R. P.

Waller

M. W.

Greene

J. M.

(2006). Who Runs? A Demographic Profile of Runaway Youth in the United States. Journal of Adolescent Health, 39(5), 778–781.

55.

Scott

L. M.

Crenshaw

Venegas

E. M.

(2019). Bodies are not commodities: Examining A21 curriculum’s impact on students’ knowledge, attitudes, and advocacy about anti-human trafficking rights and issues. Journal of Human Trafficking, 5(1), 25–42.

56.

Sheeran

Abraham

(1996). The health belief model. Predicting Health Behaviour, 2, 29–80.

57.

Srebnik

Cauce

A. M.

Baydar

(1996). Help-seeking pathways for children and adolescents. Journal of Emotional and Behavioral Disorders, 4(4), 210–220.

58.

Trafficking Victims Protection Act, 22 U.S.C. 7101 (2000). https://www.govinfo.gov/content/pkg/BILLS-106hr3244enr/pdf/BILLS-106hr3244enr.pdf

59.

Walsh

Zwi

Woolfenden

Shlonsky

(2018). School-based education programs for the prevention of child sexual abuse: A Cochrane systematic review and meta-analysis. Research on Social Work Practice, 28(1), 33–55.

60.

Wang

M.-T.

Fredricks

J. A.

(2014). The reciprocal links between school engagement, youth problem behaviors, and school dropout during adolescence. Child Development, 85(2), 722–737.

61.

Webb

T. L.

Sheeran

(2006). Does changing behavioral intentions engender behavior change? A meta-analysis of the experimental evidence. Psychological Bulletin, 132(2), 249.

62.

Weinstein

N. D.

Sandman

P. M.

Blalock

S. J.

(2020). The precaution adoption process model. In The Wiley encyclopedia of health psychology (pp. 495–506).

63.

Wiener

R. L.

Berry

M. C.

Wertheimer

Petty

Martinez

(2021). The public’s judgment of sex trafficked women: Blaming the victim? Journal of Experimental Psychology: Applied, 27(3), 529.

64.

Wong

J. S.

Bouchard

Lee

(2021). The effectiveness of college dating violence prevention programs: a meta-analysis. Trauma, Violence, & Abuse, 24, 684–701.

65.

Zimmerman

M. A.

(2000). Empowerment theory. In Handbook of community psychology (pp. 43–63). Springer.