A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM

101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010

A REPORT ON THE USE

OF TECHNOLOGY TO

RECRUIT, GROOM AND

SELL DOMESTIC MINOR

SEX TRAFFICKING VICTIMS

JANUARY 2015

In collaboration with Dr. Vanessa Bouché, Assistant Professor,

Department of Political Science, Texas Christian University

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 2

TABLE OF CONTENTS

INTRODUCTION 5

OVERVIEW OF SURVEY & PARTICIPANTS 6

MEETING & PSYCHOLOGICAL COERCION 10

RECRUITMENT & FRAUD 14

COMMUNICATION, ADVERTISING & AUTONOMY 17

HELP AND ASSISTANCE 25

TEXT ANALYSIS 34

RECOMMENDATIONS 38

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 3

Dr. Vanessa Bouché is Assistant Professor of Political Science at Texas Christian

University. Dr. Bouché is currently a co-principal investigator on three federally-funded

grants on human tracking, two from the National Institute of Justice and one from the

United States Agency on International Development. Her research has been published in

the Journal of Politics, Journal of Public Policy, Politics & Gender, among other outlets.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 4

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 5

Thorn’s mission is to drive technology innovation to combat the sexual exploitation of

children. In an eort to do this most eectively, Thorn commenced a survivor survey in

2012 to understand the ways in which controllers leverage various technology platforms

to recruit, groom and sell domestic minors. The purpose of this report is to disseminate

the results of the survivor survey and to inform key anti-tracking stakeholders in

the private, non-proﬁt and public sectors about general trends at the intersection of

technology and domestic minor sex tracking.

The questions we sought to answer with the survivor survey include:

IS THE INTERNET BEING USED TO RECRUIT DOMESTIC MINORS FOR THE PURPOSE OF

COMMERCIAL SEXUAL EXPLOITATION? IF SO, TO WHAT DEGREE? WHAT SITES ARE

USED?

IS THE INTERNET BEING USED TO GROOM DOMESTIC MINORS FOR THE PURPOSE OF

COMMERCIAL SEXUAL EXPLOITATION? IF SO, HOW?

HOW OFTEN IS THE INTERNET USED TO SELL DOMESTIC MINORS FOR THE

PURPOSE OF COMMERCIAL SEXUAL EXPLOITATION? WHO POSTS THE ADS? WHO

COMMUNICATES WITH THE BUYERS? HOW DO THEY COMMUNICATE WITH THE

BUYERS?

WHAT LEVEL OF ACCESS DO TRAFFICKING VICTIMS HAVE TO THE INTERNET AND

OTHER TELECOMMUNICATIONS DEVICES? TO WHAT EXTENT IS THEIR ACCESS

MONITORED?

HOW CAN THE INTERNET AND OTHER TELECOMMUNICATIONS DEVICES BE

LEVERAGED TO ASSIST VICTIMS AND HELP THEM ESCAPE THEIR SITUATION?

In order to create technology-based programs to address domestic minor sex tracking,

it is important to understand how controllers and survivors utilize technology. This

knowledge will inform our approach to disrupt extant methods used by controllers and

create new and innovative interventions. We believe this report is a critical ﬁrst step in

that direction. However, this is only the beginning of an ongoing eort aimed to track

changes overtime in the ways in which technology is used to exploit children. Thorn

will continue to ﬁeld the survivor survey and produce periodic reports on trends at the

intersection of technology and domestic minor sex tracking.

INTRODUCTION

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 6

SURVEY PROCEDURE

Thorn made initial email contact with 73 anti-tracking organizations across the country

asking them if they would like to participate in the survivor survey. Of the 73 organizations

contacted, 26 declined to participate, 28 did not respond to multiple follow-up emails,

and ﬁve agreed to participate but still had not returned surveys as of the writing of this

report. Thus, 14 organizations from across the country participated in the survivor survey,

and this generated a total of 115 respondents.

1

Thorn compensated the organizations/

participants for their time in the form of a check or gift card, depending on the

organization’s stated preference.

TABLE 1: PARTICIPATING ORGANIZATIONS

OVERVIEW OF SURVEY

& PARTICIPANTS

1 There were four respondents that the researchers deemed were not human tracking cases, and therefore dropped from the sample. The respondents

were older than 18 years of age when they entered prostitution, they reported that they did not have a pimp, and there was no indication of any force,

fraud and coercion in the interview transcript.

ORGANIZATION

NUMBER OF

RESPONDENTS

CITY

GEMS 33 New York

There Is Hope For Me 25 Miami

My Life, My Choice 9 Boston

SAGE 9 San Francisco

Breaking Free 8 St. Paul

Fair Girls 5 Washington, DC

Polaris Project 5 Newark

Safe House San Francisco 5 San Francisco

Covenant House 4 New York

New Friends, New Life 7 Dallas

LA Probation 2 Los Angeles

Bilateral Safety Corridor

Coalition

1

San Diego

Generate Hope 1 San Diego

GraceHaven 1 Columbus

14 ORGANIZATIONS 115 RESPONDENTS 11 CITIES

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 7

Organizations that agreed to participate were emailed a copy of the survey instrument

and the informed consent form (see Appendix A for copy of survey and informed

consent). Survivors understood that participating in the survey was completely voluntary,

that their responses were anonymous and conﬁdential, and that they could stop at any

time or refuse to answer any question. After they agreed to participate, per the informed

consent procedure, survivors completed the paper surveys, sometimes on their own and

sometimes with the assistance of a case worker. Paper surveys were generally sent back

to Thorn in the mail, although some were scanned and emailed. Thorn then coded the

paper surveys into a dataset in preparation for analysis.

In addition to the paper surveys, we asked case workers to audio record follow up semi-

structured interviews with the victims. The purpose of the interviews was to obtain more

detailed information and to clarify points that may not have been made explicit on the

paper survey. Some case workers and organizations were willing, while others were not.

In total, we obtained audio recordings of follow-up interviews from 77 respondents.

The audio recordings were then transcribed verbatim in order to conduct additional

qualitative analysis.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 8

SURVEY PARTICIPANTS

The demographics of the respondents are found in Table 2 below. Among the most

important things to note is that, among the 111 respondents of this survey, the average age

when ﬁrst tracked was 16. This is important for both internal and external validity. First,

in terms of internal validity, the purpose of the survey is to understand how technology

is used in the recruitment, grooming and sale of domestic minors; therefore, we must be

sure we are surveying the correct population (i.e., domestic minors). Second, in order to

make generalizable inferences beyond the victims surveyed, we need a relatively large

sample population of domestic minor victims of sex tracking. Seventy-ﬁve percent of

our sample was 17 years old or younger when they were ﬁrst tracked.

TABLE 2: DEMOGRAPHICS OF RESPONDENTS

Although the vast majority of the respondents were minors when they were ﬁrst

tracked, at the time of the survey, the vast majority were adults. The average age when

interviewed for this research was 23, and 78 percent of the respondents were 18 years

of age or older at the time the survey was taken. Thus, the average number of years

since respondents were ﬁrst recruited is seven years ago (see the mean for “years since

recruitment” in Table 2). This is important to note because we expect technology to play a

greater role today in tracking minors than it did ﬁve to ten years ago.

2

For this reason, the forthcoming analysis will use “years since recruitment” to measure

correlations between technology and domestic minor sex tracking. This measure was

obtained by subtracting “age when ﬁrst tracked” from “age when interviewed.” It is the

most appropriate measure for at least two reasons: 1) it inherently captures, and is thus

a proxy for, “age when interviewed.” Indeed, the two variables are very highly correlated

(r=0.88***).

3

The longer it has been since an individual was recruited, the older they are;

and 2) it is a measure of how long ago they were ﬁrst tracked. The further in the past

they were ﬁrst tracked (even if they were a minor when they were ﬁrst tracked) the

less likely it is that technology was used in the recruitment and grooming process.

2 This is one reason why it will be extremely important to continue track changes overtime.

3 Throughout the remainder of the report, Pearson correlation coecients (r) are reported in parentheses with either one, two, or three asterisks. One

asterisk means the relationship between the two variables is signiﬁcant at the 0.10 level, or that we can be 90% conﬁdent that the relationship is not due

to random chance. Two asterisks means that we are 95% conﬁdent, and three asterisks is 99% conﬁdence.

VARIABLE MEAN

ST.

DEV.

MIN MAX

Age when ﬁrst tracked 16 4.49 5 36

Age when interviewed 23 9.42 14 56

Years since recruitment 7 8.53 0 41

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 9

In other words, using “age when ﬁrst tracked” in the analysis of the relationship

between tracking and technology is an inaccurate measure because, even if

respondents were ﬁrst tracked as a minor, that could have been over 10 years ago, in

which case we would not expect technology to play a large role. On the other hand, using

“age when interviewed” lacks validity because the majority of the sample was not minors

when they were interviewed, and there is no way to establish with that measure alone

whether they were once minor victims. Thus, the measure of “years since recruitment”

was generated as it allows us to test for dierences in technology between minor victims

that are still minors and adult victims that were ﬁrst tracked as minors. We hypothesize

that the use of technology in the recruitment, grooming and sale of the former group is

more pronounced than in the latter. Figure 1 below breaks down the sample based on

years since recruitment. Almost 40 percent of the respondents were recruited less than

three years ago.

FIGURE 1: NUMBER OF YEARS SINCE RECRUITMENT

4 TO 9

YEARS

40%

OVER

9 YEARS

21%

3 YEARS

OR LESS

39%

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 10

The vast majority of respondents (81%) reported meeting their controller in person. Only

14 percent met their controller online. It is important to note, however, that respondents

recruited more recently are signiﬁcantly more likely to have met their controller online

(r=0.22***). Of the 16 total respondents that met their controllers online, 77 percent of

them were 18 years old or younger when they met their controller online. Thus, although it

will be important to continue to track changes over time, our preliminary ﬁnding from the

survivor survey is that the majority of victims are still meeting their controllers in person,

but that younger victims who have been recruited more recently are signiﬁcantly more

likely to meet their controllers online.

4

FIGURE 2: METHOD OF MEETING CONTROLLER

MEETING &

PSYCHOLOGICAL

COERCION

INTERNET

14%

PHONE

5%

IN PERSON

81%

4 Respondents who met their controllers online were asked to indicate the website where they ﬁrst met. All of the following websites were mentioned

by at least one respondent: mofos.com, backpage.com, fabscout.com, facebook.com, usasexguide.com, craigslist.com, tagged.com, mocospace.com,

myspace.com.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 11

In addition to meeting online, we asked respondents about the method of communication

used to develop a relationship with their controllers, and even fewer use the Internet for

this purpose. Again, the vast majority developed a relationship with their controllers in

person. However, per Figure 3, younger victims recruited more recently are more likely to

use the Internet to develop a relationship with their controllers (r=0.16*).

5

Of those that

developed a relationship with their controller over the Internet (11 respondents total), only

3 did not meet their controller online. Of the respondents that met their controller online,

half of them also developed a close relationship with their controller online. Our tentative

ﬁnding, therefore, is that younger victims who have been recruited more recently are

more likely to use the Internet to meet and develop a relationship with their controller.

This ﬁnding is particularly troubling given that younger respondents express less

reticence to meet people online in general. In particular, the younger respondents in our

sample were signiﬁcantly more likely to believe that the Internet is a safe place to meet

people (r=0.20**).

Forty-eight percent of respondents report that the controller earned their trust within

the ﬁrst month of meeting, and a total of 82 percent report that the controller earned

their trust within the ﬁrst four months. There appears to be three main reasons that the

controllers are able to gain the victims’ trust so quickly. First relates to the victims’ desire

for family and belonging. Although the semi-structured interviews did not explicitly ask

about family dynamics, 73 percent of those interviewed mentioned lacking love in their

lives and/or having a bad family situation, which made them desire a close family.

6

One

survivor who ran away from home stated, “I just wanted love from my mom. From him.”

Another survivor, after being orphaned at age 12 when both of her parents passed away,

said she “wanted a family.” Even while her mother was alive she “wasn’t there because she

had mental – she was really addicted to cocaine and other drugs.” A third victim, when

asked what kind of help she wanted, stated, “just help as in being a family, just being a

regular kid, with none of the issues, and just, you know, a family setting, I guess.”

5 Percentage total is greater than 100 because respondents could select more than one answer for how they developed a close relationship with their

controller. However, most respondents used only one mode of communication to develop the relationship.

6 It is important to reiterate that these comments about bad family relationships were unsolicited. The semi-structured interview protocol did not ask

about family dynamics. Thus, we believe that this is a conservative ﬁgure, and that it is likely that more than 73 percent of the 77 interviewed had a

dicult family life.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 12

FIGURE 3: COMMUNICATION METHODS USED TO DEVELOP A RELATIONSHIP WITH

THE CONTROLLER BASED ON THE NUMBER OF YEARS SINCE THEY WERE RECRUITED

In addition to belonging, controllers earned victims’ trust quickly because controllers

made them feel loved and cared for, as part of a family. Seventy-seven percent of those

interviewed mentioned in some manner that the controller gave them the love for which

they were searching. One survivor said he made her feel like she was “joining a part

of a family, that I didn’t have at home.”

7

Another said that the controller served as the

father ﬁgure she never had: “I never had like fatherly love.” In many cases, the controllers

promised that they would take care of them: “He said that I’d never starve and that as

long as I’m with him, I don’t have to worry about no money, and that was it.”

The third reason the controllers earned their trust so quickly is the way they woo the

victims with promises of romance. One survivor said, “Yeah, like he loved me, he wanted

me to be his girlfriend, he wanted to take me out of town so that we could see the world

together, he would never leave me for another female. When he told me those things he

made me feel loved. Like someone actually cared about me.”

It follows, therefore, that 61 percent of respondents report having a “very romantic,”

“romantic,” or “somewhat romantic” relationship with their controller. The controllers are

extremely adept at using psychologically coercive language and actions that engender

trust and feelings of love and romance. One survivor noted, “He kind of played on my

emotions to get me to stay. I don’t want you to leave, you’re a part of the household.”

Another survivor describes the manipulation of her controller this way:

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

4 OR MORE

3 OR LESS

IN PERSONINTERNETPHONE

7 Among the 77 survivors interviewed, all of them had male controllers. It is important to note, however, that controllers are not always male.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 13

In the beginning he made me feel like I was walking on clouds. And then he started

beating on me, and started making me feel bad, like low. Then when I started selling my

body, I was just like, wait. This is not – I felt really dirty and stu like that, but I mean,

through all of that, I still felt like I loved him.

The psychologically coercive tactic of ﬂip-ﬂopping from nice and loving to emotionally

and physically abusive was common among those interviewed. One survivor said, “Yeah,

he was really nice at ﬁrst. And then after like a while, I don’t know, he wasn’t as nice. But I

thought he was nice. Then he’d always tell me he was sorry and he would make it better.”

In short, psychological coercion is very high, and there is no statistically signiﬁcant

dierence in terms of usage of these tactics based on age when ﬁrst tracked, age when

interviewed, or years since recruitment.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 14

Respondents were asked what the controllers recruited them to do and then what they

actually did. Figure 4 shows the total number or respondents that were told they would

perform a certain activity versus the total number of respondents that actually performed

that activity.

8

Escort services and street prostitution were the most common activities

that respondents were told they would do and that they actually did. This chart is limited,

however, in that it does not indicate the level of fraud for individual respondents. Were

the same respondents that were being told they would perform a certain activity actually

engaging in that same activity, or did the controller deceive them into thinking they would

be doing something else?

Figure 5 provides a more accurate picture of levels of fraud based on type of activity.

These percentages were calculated by dividing the total number of people that were told

they would be doing an activity by the number of those that were told they would be

doing it and actually ended up doing it. Thus, higher percentages indicate less fraud (i.e.,

one person was told they would work in a nail salon, and that person ended up working

in a nail salon, therefore the percentage is 100%). Figure 5 shows that only about half of

those that were told they would be working in strip clubs actually ended up working in

strip clubs and about 60 percent of those told they would work in massage parlors and

as waitresses ended up working in that capacity. On the other hand, almost all of those

that were told they would be doing street prostitution actually ended up doing street

prostitution.

RECRUITMENT

& FRAUD

8 Some of the “other” activities respondents were told they would be doing or actually did include: selling drugs, modeling, dancing/working at parties/

casinos/bars, assisting him/driving him around, managing the other prostituted women, swinger parties, and steal money from buyers.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 15

FIGURE 4: NUMBER OF RESPONDENTS TOLD THEY WOULD DO AN ACTIVITY VERSUS

NUMBER OF RESPONDENTS THAT DID THAT ACTIVITY

FIGURE 5: PERCENTAGE OF RESPONDENTS TOLD THEY WOULD BE DOING AN

ACTIVITY THAT THEY ACTUALLY ENDED UP DOING

0

10

20

30

40

50

60

70

OTHER

NAIL

SALON

WAITRESS

MASSAGE

PARLOR

STREET

PROSTITUTION

ESCORT

SERVICE

PORN

PRODUCTION

STRIP

CLUBS

WHAT YOU DID

WHAT YOU WERE TOLD

NUMBER OF RESPONDENTS

0

20

40

60

80

100

OTHER

NAIL

SALON

WAITRESS

MASSAGE

PARLOR

STREET

PROSTITUTION

ESCORT

SERVICE

PORN

PRODUCTION

STRIP

CLUBS

PERCENTAGE OF RESPONDENTS

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 16

It is important to note that about half of the respondents reported engaging in more

than one of the above activities, so it is also appropriate to measure fraud in terms of the

number of activities in which the victims were told they would be engaging versus the

number in which they actually engaged. Were they told they would do only one activity,

and then forced to do more? The mean number of activities victims were told they would

be doing is 1.58, and the mean number in which they actually engaged is 1.86. A dierence

in means test indicates that these are signiﬁcantly dierent from one another (p=0.01).

Therefore, our ﬁnding is that there was fraud in the recruitment process in terms of the

types and number of dierent activities in which respondents were told they would be

engaging.

The interview transcripts reveal no pattern regarding whether the victims thought they

would be doing commercial sex work or something totally dierent when they were

recruited. One thought she would be working in a business setting: “Well I mean, he

talked very business relations and like, you know, business hosting, business events, gonna

be like executive hosts and that kinda thing, I never thought it would be pimping.” Others

knew it would be some type of commercial sex work, but not prostitution. For example,

“Pretty much he told me that we were going to be in clubs giving lap dances for money.

He never told me I was going to have to sell my vagina.”

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 17

As reported in Figure 7, the majority of respondents had cell phones and 42 percent

of respondents report that they had access to the Internet while they were being

tracked.

9

Of those with Internet access, a little over half report having full access with

no monitoring on the part of the controller.

FIGURE 7: COMMUNICATION OPTIONS WHILE BEING TRAFFICKED

0 10 20 30 40 50 60 70

CELL PHONE

INTERNET

NO OPTIONS

LAND LINE

PERCENTAGE OF RESPONDENTS

COMMUNICATION,

ADVERTISING &

AUTONOMY

9 Some of the “other” activities respondents were told they would be doing or actually did include: selling drugs, modeling, dancing/working at parties/

casinos/bars, assisting him/driving him around, managing the other prostituted women, swinger parties, and steal money from buyers.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 18

However, a deeper look into these ﬁndings reveals that it is the younger victims recruited

more recently who are signiﬁcantly more likely to have access to cell phones (r=0.28***)

and the Internet (r=0.20***), while those recruited further in the past are signiﬁcantly

more likely to report having no communication options at all (r=0.23***). Figure 8

provides additional insight. Speciﬁcally, for those recruited three or less years ago, about

60 percent report having cell phone access, and about 40 percent report having Internet

access. Those recruited between four and nine years ago report the most communication

options, with close to 80 percent reporting access to cell phones and a little over half

reporting access to the Internet.

10

The older respondents recruited 10 or more years

ago are least likely to have access to cell phones and the Internet, and those that do

have Internet access are signiﬁcantly more likely to report that this access is monitored

(r=0.23***). These combined results indicate that younger respondents recruited more

recently are more likely to have cell phone and Internet access, and their Internet access is

less likely to be monitored.

FIGURE 8: CELL PHONE AND INTERNET ACCESS BY YEARS SINCE RECRUITMENT

Among those who were allowed to access the Internet, when asked what speciﬁc sites

they were allowed to access (either with or without being monitored), the major websites

listed are Craigslist, Facebook, Backpage, MySpace and email accounts, including Yahoo

Mail and Gmail. While these are sites where sex can be advertised, it is not clear from the

survey or transcripts whether they were going to these sites speciﬁcally to advertise. One

respondent reported she was allowed to access only “prostitution sites,” which may be a

trend among the respondents’ access to the Internet.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

INTERNET

CELL PHONE

OVER 9 YEARS4 TO 9 YEARSLESS THAN 3YRS

PERCENTAGE OF RESPONDENTS

YEARS IN THE LIFE

10 It is unclear why there is a bell shaped distribution based on years since recruitment. Future deployments of the survivor survey will seek to investigate

this pattern further.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 19

Regarding the ways in which the controllers advertised the victims, 63 percent report

being advertised online.

11

Those recruited more recently are signiﬁcantly more likely to be

advertised online (r=0.43***). It appears that the Internet is being leveraged increasingly

by controllers to meet, develop relationships with, and advertise for sale younger victims

for the purpose of commercial sex. Yet, these same victims have more and freer access to

various communication devices, including cell phones and the Internet. Therefore, there is

signiﬁcant opportunity to leverage the Internet to 1) prevent recruitment in the ﬁrst place,

and 2) inform young victims once they have been recruited about options to seek help.

The interview reports provide information on the process of posting the advertisements

online, including who posts the ads.

12

The ﬁndings suggest that the controllers post the

ads more often than the victims. Eighteen percent of victims report that only they posted

their own ads online, 56 percent of victims say only the controller posted the ads, and 17

percent report that both they and their controller posted the ads.

Many of those that posted their own ads were required to post a certain number of times

a day. For example, “I would have to copy and paste what he wrote and then repost.

Cause I had to post a certain amount of times a day.” Another respondent said she posted

“morning, afternoon, and night.” Several survivors said they were allowed online only for

as long as it took to post ads, and that time varied for dierent websites given dierent

pay structure for the sites. One said she could be online for “probably like 10, 10 or 15

[minutes]. Cause Craigslist you don’t have to go through all the putting your credit card

and number and all that. Backpage you have to put your credit card number and all that

stu. So depending on, I guess, we used to do both, so I guess it would take me probably

like half an hour.”

Among those that report being advertised online, Backpage

13

is the most common

website used with almost half reporting they were advertised on Backpage.

13

Craigslist

and Facebook were the second most popular websites for advertising.

14

Younger

respondents reported that advertisements sometimes used words and phrases that

signaled their age, including: fresh meat, young, virgin, prime, coochie (shaved), non-pro,

new, barely legal/18, college student/girl, lovely, daddy’s little girl, sweet, 1986 Firebird,

new in the life, liked girls, youthful, and fantasy. Others of those interviewed said that they

did not use key words to communicate their age because “most people who are going to

like pay for sex don’t care how old you are. Cause actually the younger you are, the better

it is for them. The younger you are, the more money it is. Because, innocence. And that’s

what everybody wants.”

11 This is in contrast to only 9 percent that were advertised in print/newspaper. However, a large number report being advertised “in person,” as well

(66%). In fact, the majority of those advertised in person were also advertised online (53%), and a majority of those advertised online were also advertised

in person (57%). Controllers appear to take a two-pronged approach to advertising as a way to maximize the customer base, as those buying online tend

to be a dierent demographic than those buying in person.

12 The survey did not explicitly ask who posted the advertisements online. We were able to obtain this information from unsolicited comments made

about the process of posting the advertisements in the interviews.

13 Although the interviewers did not regularly ask questions about how the trackers paid for the ads, one survivor noted that her pimp used prepaid

credit cards.

14 Other websites mentioned by at least one respondent for the purpose of advertising online include: sugardaddy.com, cityvibe.com, erosguide.com,

modelmayhem.com, bangbros.com, mofos.com, fabscout.com, friendﬁnder.com, myredbook.com, usasexguide.com, and tna.com.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 20

FIGURE 9: PERCENTAGE OF RESPONDENTS THAT WERE ADVERTISED ONLINE THAT

REPORT BEING ADVERTISED ON THESE WEBSITES

As shown in Figure 10, the majority of respondents report that pictures were almost

always used in the advertisements. However, those recruited more recently are

signiﬁcantly less likely to report photos being used to advertise them (r=-0.27***), and

signiﬁcantly more likely to report that the pictures used to advertise them were actually

pictures of someone else (r=0.23***). These results indicate that if there is no photograph

posted with the ad, there is a greater chance that this ad is for a minor. Moreover, pictures

that are used to advertise minors are very likely to be pictures of someone else in order to

hide visual allusions to the actual victims’ young age.

This comports with the anecdotal evidence from the interviews. One of the survivors

said, “When I hit 17, 18, I stopped hiding my face. I was hiding my face because he felt

like I was real young and had a baby face. So when I hit 17, 18, he started unblocking my

face.” Another survivor stated something similar: “I think it was hidden because people

could obviously see that I was young.” On the other hand, there were other respondents

that said that they either did not know why their face was hidden, or that their face was

hidden so that people could not recognize them. “I hid my face in all of them because I

didn’t want anyone else to be able to recognize me. Also I had fake ads as well. So I didn’t

want them to be like, this is this girl. It would mess up business.”

0 10 20 30 40 50 60

FACEBOOK

CRAIGSLIST

BACKPAGE

PERCENTAGE OF RESPONDENTS

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 21

FIGURE 10: HOW OFTEN PHOTOGRAPHS USED IN ADVERTISEMENTS

The devices used to take pictures of the victims varied, but, as represented in Figure 11,

the vast majority of those interviewed reported that cell phones were used to take their

photographs. Respondents recruited more recently are signiﬁcantly more likely to have

their picture taken with cell phones (r=0.55***), and signiﬁcantly less likely to have their

picture taken using disposable cameras (r=-0.26***). Of the 53 victims that were recruited

in the last 5 years, only two of them used disposable cameras, whereas 46 of them had

used cell phones to take their pictures.

MOST

20%

SOMETIMES

13%

ALWAYS

44%

NEVER

19%

RARELY

4%

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 22

FIGURE 11: DEVICE USED TO TAKE PICTURE

The survey also asked the victims what modes of communication technology were used

to communicate with the buyers. The vast majority of respondents reported that phone

calls were used and over half of the respondents reported that they used text messaging.

There is no signiﬁcant dierence in the mode of communication used to communicate

with buyers based on the number of years since recruitment except for text messaging.

Those recruited more recently are signiﬁcantly more likely to report using text messaging

to communicate with buyers (r=0.26***). Figure 12 provides a more detailed look at text

message use to communicate with buyers according to number of years since being

recruited. Almost 70 percent of those recruited three years ago or less report texting

to communicate with buyers, and the same is true for those recruited between four and

nine years ago; on the other hand, only 40 percent of those recruited over nine years

ago use text messaging to communicate with buyers. It is apparent, therefore, that

younger victims recruited more recently are signiﬁcantly more likely than older victims to

communicate with buyers via text messaging.

The survey asked the victims whether they or their controller mostly communicated with

the buyers. Fifty-three percent report that they communicated with the buyers, and 29

percent said the controllers did. Interestingly, there is a signiﬁcant correlation between

personally communicating with the buyers and the number of years since recruitment.

Respondents recruited more recently are more likely to personally communicate with the

buyers (r=0.22**). Those who personally communicated with the buyers are signiﬁcantly

likely to communicate with the buyers via escort sites (r=0.28***) and email (r=0.16*). Also

of note, those communicating with the buyers are signiﬁcantly more likely to have access

to a cell phone (r=0.38***). In short, younger victims appear to have more autonomy, are

more likely to personally communicate with the buyers than older victims, and are using

text messaging, escort sites and email to do so.

0 10 20 30 40 50 60 70 80

CELL

PHONE

DIGITAL

CAMERA

WEBCAM

DISPOSABLE

CAMERA

PERCENTAGE OF RESPONDENTS

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 23

FIGURE 12: PERCENTAGE OF RESPONDENTS COMMUNICATING VIA TEXT MESSAGE BY

YEARS SINCE RECRUITMENT

FIGURE 13: METHOD OF COMMUNICATION WITH THE BUYERS

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

OVER 9 YEARS4 TO 9 YEARSLESS THAN 3YRS

PERCENTAGE OF RESPONDENTS

YEARS SINCE RECRUITMENT

0 20 40 60 80 100

PHONE CALLS

TEXT MESSAGING

ESCORT/

CLASSIFIED SITES

EMAIL

FACEBOOK

PERCENTAGE OF RESPONDENTS

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 24

This is a meaningful ﬁnding for a variety of reasons. First, it indicates that the text short

code that Thorn developed in collaboration with Polaris Project and Twilio has promise to

reach domestic minor sex tracking victims. Second, it indicates that there is room for

additional technology interventions to assist domestic minor victims of sex tracking

beyond the text short code. Potential technology interventions can include such platforms

as email and escort sites that are being accessed regularly by victims.

The interviews revealed that, for some, the person and process of communicating with the

buyers depended on the mode of communication. One survivor said that she

communicated with the buyers, but “only if it was phone calls and in the street. My

tracker only if it was online, cause they didn’t know if it was him or me. So he would

write as if it were me.” Another survivor noted the same: “I mean he did the text. And I

just always answered the phone.”

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 25

Eighty percent of victims reported that they wanted help while they were being tracked

at some point. Forty-three percent said unequivocally that they wanted help, while an

additional 37 percent reported wanting help sometimes. Those who did not want help

typically had greater autonomy, including completely unmonitored access to the Internet

(r=0.23***). Those that report less autonomy are signiﬁcantly likely to report wanting

help. In particular, victims reporting that their controller communicated with buyers are

signiﬁcantly more likely to want help while in the life (r=0.22***). Women working in strip

clubs, pornography production and street prostitution showed no signiﬁcant patterns

regarding whether or not they wanted help; however, women working in escort services

were signiﬁcantly likely to report that they wanted help while they were in the life

(r=0.22**).

The interview data reveals an interesting pattern of the survivors desiring help after they

had been physically abused. The level of physical brutality that many of the survivors

endured is alarming. There are numerous stories of wanting to escape when they were

beat, hit, threatened at gunpoint, put in the basement without eating, and required to

stand outside “half-naked” in the cold. One described her situation in this way:

When he almost killed me, or whatever. One time he was beating me up with his hands,

with a bat or whatever, and this time not even no, it wasn’t even a bat, that was before,

he was hitting me with a [sic], from the thing, he kept hitting me, hitting me with it, and

I was like crying, I was like [sic] then kicking me, kicking me, kicking me, so he made me,

he put the water mad hot, and made me put my feet in it, you know like scalding hot

water, and told me to stay with my feet in it, and I kept saying ouch or whatever, I was

really, I was just really hurt.

Figure 15 reports the percentage of respondents that wanted help in four dierent

categories: help escaping, mental health assistance, basic needs and medical assistance.

The most important thing to note is that 64 percent of the victims interviewed said that

they wanted help escaping and 61 percent said that they wanted mental health assistance.

More respondents reported wanting help with escaping and mental health assistance

than basic needs or medical assistance. There is no signiﬁcant dierence in terms of who

wanted each type of help based on number of years since recruitment. Older and younger

victims are equally as likely to report wanting help in each of the four categories.

HELP AND ASSISTANCE

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 26

FIGURE 14: PERCENT RESPONDENTS REPORTING WHETHER THEY WANTED HELP

FIGURE 15: PERCENT RESPONDENTS WANTING DIFFERENT TYPES OF HELP

NO

20%

YES

43%

SOMETIMES

37%

0

10

20

30

40

50

60

70

MEDICALBASICS

(Food, Clothing, Shelter)

MENTAL

HEALTH

HELP

ESCAPING

PERCENTAGE OF RESPONDENTS

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 27

About one-third of the respondents reported wanting only one type of help. Of the

victims wanting only one type of help, 53 percent said that they only wanted help

escaping and 32 percent said that they only wanted help with mental health issues.

Similarly, among those wanting two or three types of help, the most common two types

of help requested were help escaping and help with mental health.

Generally, there were no distinct patterns regarding the total number of categories with

which respondents wanted help and the type of work they were doing, except for the

survivors who worked in pornography production and street prostitution. Those working

in pornography were signiﬁcantly likely to report that they wanted help across multiple

categories (r=0.18**), as were those in street prostitution (0.16*).

Given that the most common form of help that the respondents desired was help

escaping, there is clearly a strong need for programs that assist them in getting out.

These can include prostitution diversion or probation programs that assist with job

training, housing and counseling, or other specialized programs that meet the unique

needs of this particular demographic. Most noteworthy is the fact that the majority of

respondents did not want to be in the life and desired help getting out.

FIGURE 16: PERCENT RESPONDENTS WANTING HELP IN 0-4 DIFFERENT CATEGORIES

Unfortunately, although the majority of respondents wanted some type of help at some

point while they were in the life, 65 percent report that nobody ever or hardly ever

reached out to help them while they were in the life. Among those interviewed, there was

not a clear indication of the best ways people can reach out to help them. One survivor

stated that people should proactively reach out and help. “[T]he only way they could have

reached out to me was a phone call. Like to ﬁnd my ad posted on Backpage, like, ‘Hey,

you don’t have to do this no more, here’s help, call this number and I need you to go here.

0

5

10

15

20

25

30

35

HELP WITH

ALL ITEMS

321NO HELP

PERCENTAGE OF RESPONDENTS

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 28

Cause this is going to help you turn your life around.’” On the other hand, another survivor

said that the help should come in more passive forms, and that the proactive move should

be on the part of the victim. She put it this way:

I think that the person who is being tracked should contact you. Because if you try to

step in, they’re going to think that – they feel like you can’t help them even if you want

to. Even if they want the help they’ll want to be stubborn. So I would say like a billboard,

phone numbers, Internet sites. Let them come to you. If they want help they’re going to

be looking for like signs and things on the way.

Other survivors’ stories indicated that without the help of another, they would still be

in the same situation. Teachers, police ocers and males posing as buyers were all

mentioned as individuals that helped.

There were several poignant stories of teachers and school administrators reaching out

to assist victims. One survivor shared a powerful story of her teacher who persisted in

reaching out:

I had a teacher at school that saw the guy waiting for me when she got out of work.

This woman. If it wasn’t for her, I swear to god I would be dead right now. She’s an older

woman and she was quiet about it. But she saw everything. She saw the way he would

be there, dropping me o and be there at the end. And she said, “That guy looks kinda

old. Is he your boyfriend?” and I was like, “yeah, he’s my boyfriend.” And then she’s like,

“Interesting.” . . . And then she said, if you ever need help with anything just let me know.

And I was like, I don’t need help with shit. Stay out of my business. And she was trying to

get me to go see the school counselor, get me to stay with her extra hours after school.

At ﬁrst I saw what she was trying to do and I was oended. “How could you think that?”

But then I was like, that’s what it is, what am I getting upset for? . . . And she asked me if

I had a phone, and I said, “I don’t have one.” And she came to class two days later with a

cell phone, a simple little phone for me. And she put her number in it and said if you ever

need help, call me. And there was nothing I can say. I was like, I’m not going to be mean

to her cause I need one now. And I was like, I can’t believe she’s doing this. And like, I

think the next day he wanted me to go to Brooklyn and he wanted me to come back and

do the swinger party. And I said, I don’t feel like it. Because she kinda buzzed in my ear

a little bit. She told me something like, when someone loves you they wouldn’t hurt you.

They’ll listen to you.” And then I went to school the next day and he was there. And my

teacher called the cops.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 29

FIGURE 17: PERCENT REPORTING SOMEONE REACHED OUT TO HELP

Police ocers also played a role in helping the survivors by oering to “lock up bad

dates,” putting a restraining order on the controllers, setting up domestic violence cases

against controllers and working with schools to investigate the cases. A number of the

survivors noted that the easiest and best escape is to get arrested. However, among those

that stated that getting arrested was the best thing for them, it is critical to note that all

of them received services as part of that experience. One survivor describes why these

services are so important:

Like when you get arrested and they make you go to court, pay ﬁnes, and sit in classes?

Instead of sending someone to a class, who tells you that you’re an evil person and that

your soul will burn in hell because you’re a prostitute, they should oer you services. They

should oer hard skills. Don’t just tell me not to do it. I had to pay $300 for classes, and

$300 for ﬁnes. Plus rent and feed and clothe myself. How am I supposed to get the

money? If I have no skills to get a job, how do I get a job? But if I don’t get a job I can’t pay

the ﬁne and go back to jail? But if I go out and get some money I can pay the ﬁne, I can

pay the rent, and I might go back to jail.

Although police were helpful in a variety of ways for some survivors, there were

other survivors who described situations where contact with law enforcement made

their situation worse. Multiple survivors mention the lack of help, understanding and

compassion they received from law enforcement.

I just hate cops. They treated me like I was a disgusting hooker instead of a 15 year old

who got caught. . . They kept saying, ‘You’re over 18. Stop lying to us.’ And I was like, ‘No,

NO,

NEVER

39%

YES,

HARDLY EVER

26%

YES,

FREQUENTLY

13%

YES,

OCCASIONALLY

22%

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 30

you bastards, I’m 15. What do you want? A cookie? Want a trophy for being a good cop

and a horrible person?’ So cops—they need more sensitivity training.

[The police] would come because [the pimp] was beating me up or something, and I

would tell them I was underage, and they wouldn’t do anything. And they’d be like, ‘Oh,

take a walk.’ And I’d say, ‘Could you just bring me home?’ And they would say, ‘We’re not

a cab service. You wanna go home, go home.’ And I would be like, ‘It’s not that simple.

It’s not that easy to just get up and leave.’ So I guess if they would have actually brought

me home, or brought me to a hospital to see if I was ok.”

In addition to schools and police, some survivors indicated that the buyers themselves

could help. One woman told her story: “One of my tricks helped me escape from [the

controller]. If it wasn’t for him, I would still be in the life right now.”

15

Two others noted

that men could pretend to be “tricks” in order to help them escape and provide them with

phone numbers and resources to receive help.

However, one survivor expressed feeling annoyed by her experiences in which men

pretended to be buyers, and she felt they were wasting her time. She said, “I’ve had a few

people call me to do dates and then they come in and they preach to me. They’re just like,

‘There’s a way out.’ And I’m like, ‘Yeah, the front door. You’re wasting my time. I’m making

money. Get out of my face cause, like, I’m happy.’ Cause I think I’m making 200 dollars,

and then they come in with the BS.” Yet, this same survivor also noted that she found

herself wishing someone reached out to help her on “busy days, when I’m seeing over

15

people a day and I’m just disgusted with myself. I’d want a way out.”

When respondents were asked if they had ever seen a number for a hotline oering help,

72 percent reported never seeing any number at all. This is disconcerting, as 50 percent

said they would have wanted to receive a national hotline number or information about a

local agency that could help.

In terms of the best potential ways to receive information oering help, respondents

report that the three most eective methods include: hotels, billboards, and text

messages. In addition, respondents with cell phones (r=0.21**), landlines (r=0.32***) and

internet access (r=0.27***) were signiﬁcantly likely to say that online advertisements

would be the best way to receive help messages. Those that had zero communication

options while being tracked were signiﬁcantly less likely to state that online

advertisements were an eective method (r=-0.25***). In fact, respondents with no

options to communicate are not signiﬁcantly likely to view any method of receiving

information as being eective. While this ﬁnding makes sense intuitively, it also highlights

the diculty of reaching out to and ﬁnding those who are kept completely isolated by

their controllers.

15 More often, the survivors would describe having a “bad date” and getting hurt by the “tricks.” One said, “They treat you like you’re not even a human

anymore. You’re just a prop. A little box of cereal in Wal-Mart. They just pick you and pay for you.”

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 31

Older respondents who were recruited less recently are signiﬁcantly more likely to report

that the non-technology options to receive helpline information would be most eective.

Speciﬁcally, those recruited further in the past state that it would be helpful to receive

helpline information via print advertisements in the newspaper (r=0.24***), on a billboard

(r=0.40***), or in a hotel (0.24***).

Although the survey does not directly ask about speciﬁc outlets where it might be helpful

to advertise a helpline, some of the interviewees recommended posting ads on the very

escort sites where many of the women and controllers post their advertisements. One

survivor recommended the following:

Cause I think if it’s in the Village Voice – in the back of that they have all those ads for

escorts. If they had an ad for like, if you’re in an abusive situation, if you’re being forced

to do this against your will, here’s this number. Because people look at Backpage all the

time, and you never – somebody might see the number and say, hey I really do need

help. It’s not like I could have searched online, “help for people being tracked.” I didn’t

even know that I was being domestically tracked. You know what I’m saying? . . . Even

like shady areas where you know that have been tracks for years. They could have a

poster up there. They could do something.

Although technology is being used by tracking victims while they are in the life, by

and large, they are not using technology in order to help them escape the life. Indeed,

64 percent of respondents report that they did not use technology at all in their escape.

Of those that did use technology, cell phones were used the most, although they were

generally used to call someone, rather than used to text message for help.

It is important to note that several of the survivors interviewed indicated that their

controller often checked their text and phone history. Checking their phone and text

history was a form of psychological control that generally worked. For example, one

survivor said, “He would go through [my phone and text history]. I didn’t really do

anything because I didn’t want to make him mad.” Another survivor stated, “Yeah, he

would check my phone – my calls. And who I was calling, and what I said in my text

messages.” She went on to explain that she would get in trouble for calls and texts that

she was not allowed to make. “The phone was supposed to be strictly for business or to

speak to him. Or if I was like, talking to people and it wasn’t directly about an exchange of

sex for money – just talking to them I would get in trouble, because there’s no purpose in

just talking to people.”

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 32

FIGURE 18: MOST EFFECTIVE MEANS TO COMMUNICATE HELPLINE INFORMATION

FIGURE 19: USE OF TECHNOLOGY TO ESCAPE

Given various testimonies of other survivors, “getting in trouble” most likely means being

beaten, as this was the consequence other survivors faced. “When he told me not to call

my mom and I would I would get in trouble, and he would beat the shit out of me and I

would go downstairs in the basement and I wouldn’t be able to eat or drink for three days.

0 10 20 30 40 50

HOTEL

BILLBOARD

TEXT MESSAGE

ONLINE AD

PRINT AD

FACEBOOK

MESSAGE

EMAIL

PERCENTAGE OF RESPONDENTS

0

7

14

21

28

35

42

49

56

63

70

NO

TECHNOLOGY

CELL PHONETEXTINGLANDLINEFACEBOOKEMAIL

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 33

Just water basically, that’s it.” Another survivor said, “I used to get beat after I talked to

my mom or my friend. I couldn’t even talk to my mother.”

Based on the survey responses, it seems appropriate to design proactive programs and

attempt to reach out to the victims directly via text or phone. However, these survivor

testimonies demonstrate potential problems that this could create for victims. Any

proactive attempts to reach victims directly should be thoughtful and done with extreme

caution as it could lead to unintended consequences for the victim.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 34

In addition to analyzing the survey results and triangulating them with the anecdotal

evidence from the interview transcripts, the transcriptions of all 77 interviews were

analyzed using Linguistic Inquiry Word Count software (LIWC).

16

This software allows

us to evaluate the percentage of the transcripts that included various psychological

and biological constructs and personal concerns, including, but not limited to: positive

emotion, negative emotion, anxiety, anger, sadness, insight, sex/sexuality, work,

achievement, leisure, money and death.

Table 3 divides the content of the transcripts into ﬁve overarching categories: cognitive,

social, aective, perceptual, and biological processes. The cognitive category measures

words that indicate the ways in which survivors are thinking about their situations. The

social category measures relational content from the transcripts. The aective category

measures the positive and negative emotional content. The perceptual category measures

the physical senses such as seeing, hearing, and feeling. Finally, the biological category

is a measure of words relating to the physical body. Overall, LIWC is a unique analysis of

the content of the transcripts; it catches implicit meaning and subtle messages that may

otherwise be overlooked. Reported in Table 3 are the means and standard deviations for

each of the LIWC measures.

It is immediately notable that cognitive processes appear more in the transcribed

interviews than social, aective, perceptual, and biological processes. There are several

reasons that the transcripts might reveal this pattern of survivors thinking about their

situation rather than expressing feelings. First, the mode of communication was an

interview with prescribed questions. With another person asking questions, survivors may

feel more inhibited expressing emotions. For example, we might see a dierent trend if

the analysis was based on journal entries rather than interview transcripts. Second, the

interviews were semi-structured and followed a relatively deﬁned set of questions. There

was less space and opportunity for survivors to discuss their situations through a stream

of consciousness. Also, the interviews were structured around the survey instrument

as a means of asking clarifying questions, and the nature of the survey was to obtain

very speciﬁc information about the use of technology in the recruitment, grooming,

and sale of domestic minors for sex. For each of these reasons, it is not necessarily

surprising that cognitive processes overshadow the other categories in the LIWC analysis.

TEXT ANALYSIS

16 With LIWC, each transcript is run through the software, which generates a newly designated Excel ﬁle with approximately 80 output variables written

as one line of data. This data includes 4 general descriptor categories (total word count, words per sentence, percentage of words captured by the

dictionary and percent of words longer than six letters), 22 standard linguistic dimensions (e.g., percentage of words in the text that are pronouns, articles,

auxiliary verbs, etc.), 32 word categories tapping psychological constructs (e.g., aect, cognition, biological processes), 7 personal concern categories

(e.g., work, home, leisure activities), 3 paralinguistic dimensions (assents, ﬁllers, nonﬂuencies), and 12 punctuation categories (periods, commas, etc.). For

more information on LIWC, see liwc.net.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 35

Future research might explore ways to allow survivors to express their stories in a non-

structured, more open-ended format (e.g., journaling) to evaluate how this might change

the distribution of the various psychological processes at play.

Beyond examining the overall descriptive statistics, it is useful to investigate these results

for dierent groups of survivors. Speciﬁcally, how do older respondents recruited further

in the past dier in their cognitive, social, aective and perceptual processes from the

younger victims who were recruited more recently?

Those recruited further in the past are signiﬁcantly more likely to mention family

(r=0.36***). They are also signiﬁcantly more likely to register negative emotions (r=0.27**)

and anger (r=0.35**). Their cognitive processes are also much stronger (r=0.26**).

Interestingly, their level of insight is signiﬁcantly greater (r=0.37***), yet they tend to

speak in less certain terms (r=-0.22*). Those recruited further in the past register more

feeling (r=0.26**) and use more inclusive language (r=0.34***). Moreover, their openness,

and perhaps awareness, of their biological processes is also much greater (r=0.28**), with

signiﬁcantly more mentions of health (r=0.22*), sex/sexuality (r=0.32***), and ingestion

(r=0.23**). Finally, those recruited further in the past mention religion (r=0.19*) and death

(r=0.22**) signiﬁcantly more than those recruited more recently, and they are signiﬁcantly

less likely to use assenting language (r=-0.32***).

These combined results indicate that older respondents who were recruited further in

the past have a level of awareness and openness that the younger victims do not have,

even in a semi-structured interview format. Their increased expression of anger and other

negative emotions may be a function of their heightened insight into their thoughts

and emotions, rather than an indication that they are actually more angry and negative.

They are also more open and willing to discuss uncomfortable subject matter. In general,

the LIWC reveals a maturity among the older respondents compared with the younger

respondents.

Further insight is gained by investigating the LIWC variables as they relate to those

respondents with free access to communication options. Those with free access are

signiﬁcantly more likely than those without to register positive emotions (r=0.20*). Their

perceptual process of feeling is also signiﬁcantly more pronounced than those without

communication access (r=0.22*). These combined results indicate that free access to

communication options is a signiﬁcant predictor of a survivor’s emotional wellbeing while

in the life.

Also of note is that those with open communication options are signiﬁcantly more likely

to discuss money (r=0.28***). Although it is unclear why there is a signiﬁcant relationship

between open communication options and discussing money, future research should dig

more deeply into why this relationship exists. For example, is this an indication that those

with open communication options also face more or less autonomy over their earnings?

Interesting patterns also emerge among survivors who were involved in dierent parts of

the commercial sex industry. First, survivors tracked into pornography production are

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 36

signiﬁcantly less likely to use words that are social in nature (r=-0.20*), and signiﬁcantly

more likely to use exclusive language, such as but, without or exclude (r=0.23**). They

also are signiﬁcantly less likely to use motion words, such as arrive, car and go (r=-0.28**).

These combined results suggest that the survivors who reported being tracked into

pornography production face more social isolation and a lack of ability to move from one

place to another. One survivor who had been tracked into pornography production said,

“I would have been out sooner than I was if I had just somebody there, it would have been

totally dierent.”

Survivors who reported working in strip clubs were signiﬁcantly less likely than those who

did not work in strip clubs to mention friends (r=-0.22**) and to register any aect (r=-

0.20*), indicating that women in strip clubs might be lonely and emotionally numb. On

the other hand, they were signiﬁcantly more likely to discuss work (r=0.20*), achievement

(r=0.21*), and leisure (r=0.26**). These combined patterns tentatively suggest that

survivors who worked in strip clubs reached a level of resignation with their situation,

accepting it, to a degree, as their job or lot. However, reaching this resignation required

some level of emotional dissociation.

In addition, survivors who worked in strip clubs are signiﬁcantly more likely than survivors

who worked in other parts of the commercial sex industry to mention money (r=0.43***).

The substantive and statistical signiﬁcance of this relationship is much greater than the

relationship between survivors with open communication options who mentioned money.

The reasons behind this relationship are not immediately apparent. It could be that

women in strip clubs have more or less autonomy over their earnings, but these results

warrant further investigation.

Women working in escort services reveal yet another pattern. They are signiﬁcantly

more likely than those that did not work for an escort service to use words that indicate

sadness (r=0.28**) and exclusion (0.28***). Additionally, they are signiﬁcantly more likely

than survivors of other types of commercial sex work to mention work (r=0.23**). The

tentative ﬁndings here suggest that women working in escort services also feel socially

isolated and acknowledge the sorrow this isolation engenders. Similar to women working

in strip clubs, women working in closed-door prostitution via escort services likely

acknowledge this as “work” but, unlike those in strip clubs, those in escort services allow

themselves to feel.

These results are interesting in the context of the public nature of strip clubs versus

the private nature of closed-door prostitution. Strip clubs are social by nature—it is a

gathering of people in a public place—thus, rather than being physically isolated, women

in strip clubs are among a crowd. On the other hand, women working in escort services

are physically isolated—usually conﬁned to a private room with strangers. Where the

social setting appears to be correlated with emotional dissociation, the isolated setting is

correlated with an acknowledgement of sadness.

Finally, survivors involved in street prostitution are signiﬁcantly less likely than women

in other types of commercial sex work to mention family (r=-0.22*), friends (r=-0.23**),

and positive emotions (r=-0.24**). These ﬁndings might suggest that women in street

prostitution lack social and emotional bonds, but are not isolated in the same way that

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 37

those in porn production and escort services are. Survivors of street prostitution are

also more likely to discuss biological (r=0.39***) and sexual (r=0.38***) processes. While

each type of commercial sex work is dierent, they all involve some biological/sexual

component; therefore, it is not clear why survivors of street prostitution are more likely

to mention these processes. Finally, women in street prostitution are signiﬁcantly more

likely to use language that is relative (r=0.29***), including space, such as down, in, or thin

(r=0.33***). This is not surprising given that women in street prostitution discuss walking

“the tracks” and being on “the streets.”

Overall, the results of the LIWC reveal that there is a great deal of variance in the social,

aective, cognitive, perceptual, and biological processes of dierent women that have

been tracked into commercial sex. Variance is evident along a range of dimensions,

from years since recruitment to the type of sex work into which survivors were tracked.

These preliminary results indicate that victims of sex tracking in the United States are

very dierent from one another. Their age, the length of time they were tracked, the

type of work they performed, and their level of autonomy and freedom to communicate

with the outside world all must be taken into consideration when developing dierent

mechanisms of intervention, assistance and rehabilitation. Still, the results of the LIWC

consistently reveal that positive aect, inclusion, and belonging are rarely felt among

tracking victims. Therefore, we can reasonably conclude that social isolation is among

the most challenging barriers to eective outreach.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 38

Based on the preliminary ﬁndings in this report, there are a number of recommendations at

the intersection of research, policy and technology that can impact strategies which combat

domestic minor sex tracking, protect victims, and help prosecute controllers. These

recommendations are enumerated below.

1. REVISE AND CONTINUE TO FIELD THE SURVIVOR SURVEY

The survivor survey has assisted in the collection of unique data that has produced

invaluable insight into the use of technology for the recruitment, grooming, sale and escape

of domestic minor sex tracking victims. However, technology is a dynamic ﬁeld. With new

trends constantly emerging, we can expect controllers to adapt quickly. When one method is

compromised, their adjustments will likely align with new technology options. Thus, we

should expect sometimes dramatic changes in the role technology plays in domestic minor

tracking situations over relatively short increments of time. A sustained survivor survey can

help track and understand these changes.

Continued eorts to revise and ﬁeld the survey will provide updated data, including new

websites, new keywords, and new methods of recruitment and sale, such as technology and

communication strategies. This information will continue to inform evidence-based policy

and practice in combating domestic minor sex tracking.

Therefore, it is recommended that this survey be revised and adapted every two years. To

that end, simultaneous to the publishing of this report, Thorn has deployed the second wave

of the survivor survey. Thorn welcomes partnership with any organization that would like to

collaborate.

2. INTERNET STRANGER DANGER CAMPAIGNS

This research suggests that there is an increasing trend among domestic minor sex

tracking victims of online recruitment, sometimes on gaming websites and Facebook. One

strategy to combat this trend is to develop advertisements and PSAs that draw the attention

of young people and send a clear message that meeting new people on the Internet is not

always safe. These should be linked to resources individuals can access if they need help or

are in an uncomfortable situation.

3. INTERNET HELP ADS

The survey also revealed that many victims do have Internet access, especially younger

victims, and we believe this trend will only increase overtime. This provides an opportunity to

leverage the Internet to increase online messaging to get the National Human Tracking

RECOMMENDATIONS

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 39

Resource Center hotline and corresponding text short code into the hands of those who

need it most. The majority of victims surveyed said that they wanted help escaping, but

did not know how or where to receive that assistance. For victims with access to the

Internet, one suggestion is to display helpline information in bold ads on strategic

websites where victims post their advertisements. Thorn has started this process and

welcomes partnerships with any organizations that would like to collaborate.

4. EXPLORE OPPORTUNITIES TO EXPAND TEXT SHORT CODE

Continued research and exploration into proactive marketing of the text short code is

recommended. This would need to be done in a very cautious way, as multiple victims

report strict monitoring of their communications and violent reactions from controllers

when they communicate with unauthorized individuals. However, some victims report less

restricted communication options and could beneﬁt from more proactive measures. With

continued research into the dierent types of escort page postings, it may be possible to

predict with some degree of certainty which ads are linked to individuals under tighter

control.

5. COMPREHENSIVE SAFE HARBOR LEGISLATION

Minor victims of sex tracking suer severe forms of abuse; they are not criminals or

juvenile delinquents. Accordingly, they should not be prosecuted as criminals. They

deserve various social services, including prostitution diversion programs and child

protective services. Many states have recognized this as a matter of law, including New

York, Illinois, Massachusetts, Minnesota, New Jersey, Ohio, Vermont and Washington.

Moreover, in 2010, the Texas Supreme Court ruled that minors engaged in prostitution are

victims, not criminals, and should qualify for child protective services. All states should

move to adopt comprehensive safe harbor legislation for domestic minor sex tracking

victims that not only provide immunity or diversion for victims, but also establish holistic

services for the rehabilitation and reintegration of these child victims.

6. IMPROVED SERVICES FOR DOMESTIC MINOR SEX TRAFFICKING VICTIMS IN

JUVENILE DETENTION

The results of this survey indicate that minors arrested for prostitution and placed in

juvenile detention are not being given the services that they need. Juvenile detention

centers across the country must recognize the unique needs of children that have been

tracked and should partner with qualiﬁed service providers to increase counseling,

education and job training opportunities available for domestic minor sex tracking

victims while in juvenile detention. Rather than a lost opportunity, victims’ time in juvenile

detention could be life changing with the right combination of rehabilitation and

reintegration services that enable them to live healthy and productive adult lives.

A REPORT ON THE USE OF TECHNOLOGY TO RECRUIT, GROOM AND SELL DOMESTIC MINOR SEX TRAFFICKING VICTIMS | 40

TABLE 3: EXAMPLES AND DESCRIPTIVE STATISTICS FOR LIWC VARIABLES

CONSTRUCT EXAMPLES MEAN ST. DEVIATION

Social Processes

mate, talk, they, child

14.42% 2.90%

Family

daughter, husband, aunt

0.38% 0.49

Friends

buddy, friend, neighbor

0.41% 0.34

Aective Processes

happy, cried, abandoned

3.18% 0.95

Positive emotion

love, nice, sweet

2.30% 0.87

Negative emotion

hurt, ugly, nasty

0.85% 0.45

Anxiety

worried, fearful, nervous

0.12% 0.15

Anger

hate, kill, annoyed

0.28% 0.29

Sadness

crying, grief, sad

0.08% 0.11

Cognitive Processes

cause, know, ought

19.53% 3.06

Insight

think, know, consider

2.23% 1.07

Causation

because, eect, hence

1.58% 0.64

Discrepancy

should, would, could

2.97% 0.96

Tentative

maybe, perhaps, guess

3.39% 1.38

Certainty

always, never

1.46% 0.61

Inhibition

block, constrain, stop

0.30% 0.21

Inclusive

and, with, include

5.54% 1.59

Exclusive

but, without, exclude

3.61% 1.36

Perceptual Processes

observing, heard, feeling

1.45% 0.7

See

view, saw, seen

0.52% 0.36

Hear

listen, hearing

0.57% 0.42

Feel

feels, touch

0.33% 0.25

Biological Processes

eat, blood, pain

1.34% 0.62

Body

cheek, hands, spit

0.32% 0.32

Health

clinic, ﬂu, pill

0.48% 0.3

Sexual

horny, love, incest

0.42% 0.36

Ingestion

dish, eat, pizza

0.17% 0.25

Relativity

area, bend, exit, stop

13.03% 2.43

Motion

arrive, car, go

2.24% 0.85

Space

down, in, thin

5.40% 1.36

Time

end, until, season

6.42% 1.33

Personal Concerns

Work

job, majors, xerox

1.02% 0.71

Achievement

earn, hero, win

0.84% 0.35

Leisure

cook, chat, movie

0.94% 0.68

Home

apartment, kitchen, family

0.67% 0.49

Money

audit, cash, owe

0.64% 0.47

Religion

altar, church, mosque

0.07% 0.15

Death

bury, con, kill

0.02% 0.05

Other

Fillers

Blah, Imean, youknow

1.07% 0.8

Assent

agree, ok, yes

2.03% 1.66

101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010101010