Evaluation of Safe Harbor Laws and Their Impact on the Commercial Sexual Exploitation of Children, United States, 2015-2018 (ICPSR 37497)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Hide

Study Purpose View help for Study Purpose

The purpose of this study was to evaluate the impact of safe harbor laws on the commercial sexual exploitation of children. The goal of all safe harbor laws is to repeal the punitive approach to minors who are victims of commercial sexual exploitation and sex trafficking (i.e., prostituted minors). In 2014, the U.S. Department of Justice's Office of Juvenile Justice and Delinquency Prevention awarded Development Services Group (DSG) a grant to evaluate safe harbor laws and their impact on the Commercial Sexual Exploitation of Children (CSEC). DSG evaluation, was conducted in three phases, with the findings from each presented in overlapping and standalone reports.

Phase 1, a legal review, addressed the evolution of safe harbor laws in the United States. It provided an overview of CSEC, detailed the philosophy and conceptualization of safe harbor laws, and presented findings regarding state-level efforts to adopt safe harbor laws. The remaining 15 states did not have a safe harbor law (13 continued to treat these victims as criminals, and 2 states allowed for an affirmative defense). This phase was completed in March 2018.

Phase 2, a quantitative study, utilized elements of the legal review to conduct an assessment of safe harbor laws using a quasi-experimental, longitudinal design to compare counties that have and have not implemented safe harbor laws over an 11-year period (2005-15). Conversely, since safe harbor laws not only prohibit the arrest of these youth but also often redirect them to agencies that can provide specialized services, it is expected that sexual abuse maltreatment cases will increase in counties that have implemented safe harbor laws relative to counties that have not implemented such laws. This phase was completed in December 2018.

The purpose of Phase 3 is to understand the implementation processes within each state to identify themes, gaps, key processes, and best practices associated with effective outcomes for safe harbor laws. Phase 3 involves an in-depth qualitative assessment of two states: one that implemented safe harbor laws and demonstrated positive outcomes, and one that implemented safe harbor laws without demonstrating positive outcomes. The states were selected based on the outcomes identified in the Phase 2 analysis. An effective outcome was defined as a decrease in the number of prostitution arrests among individuals younger than 18 since the inception of safe harbor laws and policies. The researchers used the initial pool of state cases with effective and ineffective outcomes as identified in Phase 2 to develop the set of candidate states with effective and ineffective outcomes in Phase 3. To limit differences, they reviewed and matched these states on three criteria: region of the country, structure of safe harbor laws, and timing of policy adoption.

Study Design View help for Study Design

Phase 1: To assess the current scope, nature, and impact of safe harbor legislation, DSG implemented a three-phase design that integrates qualitative and quantitative components. The first phase (legal review) involved:

1. Conducting an environmental scan to develop a better understanding of the various elements that comprise safe harbor legislation.

2. Constructing an instrument to capture the status of all states for each element.

3. Subjecting the coding scheme to an expert review panel.

4. Coding all states for each element by tracking the legislative history of relevant bills to pinpoint the effective date of each public law related to safe harbor legislation.

5. Calibrating findings of the coding process by presenting it to the expert review panel for comment.

A team of five researchers coded the safe harbor status of all 50 states (excluding the District of Columbia, Puerto Rico, and other territories) to create a safe harbor state dataset that included the 11 elements: (1) immunity, (2) diversion, (3) mandatory referral, (4) demand-side penalties, (5) trafficker-side penalties, (6) investigation, (7) training, (8) task force, (9) specialized services, (10) civil action, and (11) vacating convictions.

Phase 2: A nonrandomized quasi-experimental design was used to compare prostitution-related crime trends and sexual abuse maltreatment trends at the county level in states that have implemented safe harbor laws with prostitution-related arrest and sexual abuse maltreatment trends in a control group of counties in states that have not implemented safe harbor laws. A county-level core dataset was created to account for each county in the United States monthly, over the 11-year study period (2005-15). Next, a variable was added to identify treatment units by indicating the monthly safe harbor status of each county. Then the yearly sociodemographic estimates from the American Community Survey were merged to this dataset to create the core safe harbor dataset. Finally, two sets of outcome data were integrated, separately, with the core dataset to create two independent datasets that each shared common elements. The first dataset merged monthly crime data from the Uniform Crime Reports (UCR). The second integrated monthly maltreatment cases from the National Child Abuse and Neglect Data System (NCANDS).

Phase 3: The research team used a comparative case design to understand the implementation processes within two states that had different outcomes and different safe harbor law structures. The two states were selected based on matching criteria that allowed for a comparison of a state that demonstrated effective outcomes (state A) and a state that demonstrated ineffective outcomes (state B). Interviews with key informants were conducted in these two states. In each state, two categories of key informants were recruited: those who contributed to constructing the safe harbor policies, and those who had been involved in the implementation of programs created by the policies. The interview sample includes a total of 17 interviews: 11 completed interviews from state A and 6 completed interviews from state B.

Sample View help for Sample

Phase 1: A legal review protocol was executed to code the safe harbor status of all 50 states (excluding the District of Columbia, Puerto Rico, and other territories).

Phase 2: The United States includes 3,236 unique jurisdictions. Because the Census does not produce yearly estimates for smaller entities, the sample was restricted to counties and county equivalents with a population of 65,000 or more, reducing the number of eligible entities to 819 jurisdictions.

In terms of crime trend data, Alaska and Florida, however, do not report any prostitution-related crimes during the specified timespan. Moreover, prostitution-related crimes in Alabama and Minnesota spiked unusually high annually each December, suggesting a reporting anomaly. Similarly, Illinois spiked in June and December of every year. Consequently, each of these five states was removed from the sample, further reducing the number of eligible entities to 718.

In terms of maltreatment data, Vermont did not report to NCANDS by county, instead using a statewide code, and thus was excluded. In addition, a unique feature of the NCANDS data is that cases are anonymized if they come from a jurisdiction with fewer than 1,000 reports in a reporting year, to protect the anonymity of those involved. These anonymized cases were not included in the analysis. Consequently, the final number of eligible entities contributing data to the analysis was 688.

Phase 3: For each state, an initial list of potential interviewees was created. This intentional sampling frame produced 40 potential interviewees, 20 per state. Snowballing was also used to identify potential interviewees by asking at the conclusion of each interview if the interviewee could identify additional people to contact. The snowball sampling method produced six additional interviewees. In total, this process yielded an interviewee sample of 46 individuals, 23 for each state. In state A 11 interviews were conducted (48 percent response rate). In state B 6 interviews were conducted (26 percent response rate).

Time Method View help for Time Method

Universe View help for Universe

All U.S. county and county equivalents with a population of 65,000 or more.

Unit(s) of Observation View help for Unit(s) of Observation

Data Type(s) View help for Data Type(s)

Mode of Data Collection View help for Mode of Data Collection

Description of Variables View help for Description of Variables

Dataset 1: There are 101 variables included in this dataset. Some of the different variable grouping included in this dataset are Research, UCR, and Census variables. Research variables include original coding by research staff on the type and timing of safe harbor legislation in each state as well as some structural time and place variables. Census variables include questions about demographic and geographic data obtained from the US Census. UCR variables include questions about crime outcome variables recoded from from the UCR Program.

Dataset 2: There are 103 variables included in this dataset. Some of the different variable grouping included in this dataset are Research, Census, and Maltreatment variables. Research variables include original coding by research staff on the type and timing of safe harbor legislation in each state as well as some structural time and place variables. Census variables include questions about demographic and geographic data obtained from the US Census. The Maltreatment variables are child abuse and neglect report data from the NCANDS which was recoded to only contain reports of child sex abuse as a maltreatment type.

Qualitative Data: Interviews with key informants were conducted in two states. In each state, two categories of key informants were recruited: those who contributed to constructing the safe harbor policies, and those who had been involved in the implementation of programs created by the policies. The interview sample includes a total of 17 interviews: 11 completed interviews from state A and 6 completed interviews from state B. Though there were a total of 17 interviews, only 15 interview transcripts were deposited and will be released as part of this study. Question topics for these interviews include Respondent Background, Agenda Setting, Access to Decision-Making, Achieving Favorable Policies, Monitoring and Shaping Implementation, and Influence on Long-Term Priorities. All the questions asked to the interviewees were deposited with the data and will be released as part of this study.

Response Rates View help for Response Rates

Phase 1: Response rates are not applicable.

Phase 2: Response rates are not applicable.

Phase 3: The potential interviewees included 46 individuals, 23 for each state. In state A, 11 interviews were conducted (48 percent response rate). In state B, 6 interviews were conducted (26 percent response rate).

Presence of Common Scales View help for Presence of Common Scales

None

Hide

Original Release Date View help for Original Release Date

2022-04-28

Version History View help for Version History

Hide