4
International Approaches to Measuring Prevalence
This chapter covers the workshop discussions on international counter-trafficking work being conducted by select government and nongovernmental organizations in the United States, as well as counter-trafficking collaborations involving actors from the United States and Europe.
SECTOR- AND INDICATOR-BASED APPROACHES
U.S. Bureau of International Labor
Lauren Damme (U.S. Department of Labor) explained that the Office of Child Labor, Forced Labor, and Human Trafficking, which is housed in the department’s Bureau of International Labor Affairs (ILAB), was formed to promote a fair global playing field for workers in the United States and around the world by enforcing trade commitments, strengthening labor standards, and combating international child labor, forced labor, and human trafficking. Although the office is well known for its work in child labor, one-third of its $259 million annual budget is currently allocated to combating forced labor. One of the office’s major research products is an annual report of the findings on the worst forms of child labor, authorized by the Trafficking Victims Protection Reauthorization Act, which as of 2018 included a list of 148 goods produced by forced labor, child labor, and forced child labor in 76 countries. Information from the report also appears in a mobile app ILAB created called “Sweat and Toil.”
Damme is part of ILAB’s monitoring and evaluation team and personally oversees the agency’s portfolio of impact evaluations—most of which
are randomized controlled trials conducted to understand what works to combat child and forced labor. She talked about three impact evaluations ILAB conducted, two in Nepal and one in Hong Kong, to measure the effectiveness of different mass media campaigns in decreasing vulnerability to child labor, forced labor, and forced child labor. The evaluations tested posters, graphic novels, radio campaigns, and multiple audiovisual campaigns. Damme said they measured vulnerability using an index based on the Hard to See, Harder to Count framework of the International Labour Organization (ILO) (see Chapter 2) and performed list experiments—asking program participants if they identify with particular statements presented at random, instead of asking them questions outright—to get clarity on participants’ labor conditions. Overall, the studies found that danger narratives that portray the victims in hopeless situations were much less effective in reducing vulnerability than empowerment narratives that showed victims finding support to get out of trafficking situations.
Carolyn Huang (U.S. Department of Labor), who serves on the research and policy team at ILAB, told the workshop participants that ILAB’s technical cooperation projects create tools for forced labor research and help countries increase their capacity to work with the ILO to produce data at nationally representative levels. ILAB has also provided technical support for external research tools, such as Hard to See, Harder to Count, and for pilot studies seeking to understand whether national statistical systems can be leveraged to derive prevalence estimates on forced labor.
In Brazil, Huang said, ILAB funded a study of 2,000 households that looked at forced labor and hereditary slavery. As a result of this survey, Brazil is now trying to incorporate a forced labor module as part of its National Health Survey and to learn more about the linkages between forced labor and health. Huang also described a study in Malaysia examining forced labor in the production of palm oil and the sampling strategy used to protect the privacy and confidentiality of the survey respondents. She talked about similar studies that ILAB supports in Nepal and Peru and about a collaboration between the ILO, UNICEF, the International Organization for Migration (IOM), and the Walk Free Foundation to understand children in armed recruitment, in which they used information gathered from internationally displaced person camps in three countries to estimate the number of children. ILAB is providing support to expand and house these various research tools on the UN’s Delta 8.7 platform.
Huang underscored the importance of considering the safety of the population of interest from a study’s inception and of striving to decrease harm even in the process of accessing the population and creating a sample. She acknowledged that household survey methods used to measure prevalence of forced labor may be less effective in some contexts, especially in populations that are less accessible or visible. She noted that the cost
of collecting data on nationally representative levels—both money and resources—needs to be considered; accessing and identifying a sufficient sample of forced labor victims through household surveys can be a costly process because the population is relatively rare. She concluded by saying that the ultimate goals should be to obtain information and develop interventions that address and prevent harm and to use the findings to inform policy makers.
Verité
Erin Klett (Verité) told the workshop participants that her organization is a global nongovernmental organization that has worked since 1995 with companies, governments, and civil society to ensure that people around the world work under safe, fair, and legal conditions. Verité partners with major multinational companies to help them better understand labor conditions in their supply chains through a variety of approaches, including workplace assessments, training, capacity building, and independent research. It uses a sector-based approach to illuminate labor rights problems with the goal of better understanding risk and vulnerability. Verité has conducted research in 15 sectors in 20 countries, and the results of the research have been used to develop recommendations and advocate for new policy. The 15 sectors are apparel, cattle, cocoa, coffee, corn, electronics, fishing, gold, palm oil, rubber, shrimp, sugar, tea, tobacco, and tree nuts.
One of Verité’s biggest activities is ResponsibleSourcingTool.org, which is funded by the Office to Monitor and Combat Trafficking in Persons in the U.S. Department of State. The website describes trafficking risk in different sectors and geographies and provides tools for compliance management and risk mitigation. Verité uses the ILO’s Forced Labour Convention definition for its sector-based approach. Klett said Verité has recently begun looking at human trafficking from an indicators-based approach in order to create profiles of vulnerability; for this approach the ILO’s Hard to See, Harder to Count is used as the guiding principle.
When Verité begins a study, the investigators first identify vulnerability indicators that may be relevant to the labor sector of the study. Verité uses the ILO guidance to develop preliminary definitions of how those indicators may present themselves, and it then creates a survey that asks respondents about the presence or absence of each indicator. This approach allows the organization to get a sense of whether there is a risk of forced labor or human trafficking among a certain population of workers, as well as how that risk could be reduced or eliminated. It also allows Verité to document the presence of risk even in cases in which a worker might technically not be in a condition of forced labor or human trafficking at that moment, but could be highly vulnerable to it.
Klett described a Verité study in Malaysia1 that focused on seeing if forced labor in the electronics sector was an isolated or a widespread issue. The study used a cross-sectional research design that used a variety of data collection methods: the survey included a closed-ended questionnaire, open-ended questions, and personal interviews, and Verité used a combination of nonprobability sampling and snowball sampling to recruit participants. The research team interviewed 501 workers, including people from the countries that are primary consumers of electronics from Malaysia and people from major producing regions within Malaysia, aiming to interview no more than five workers from any given factory and trying to cover workers on all major Malaysian electronic products. Klett said the team acknowledged that the method may not have yielded a statistically representative sample, but the primary goal was to have a sample that was reasonably reflective of the different key dynamics in creating vulnerability.
To develop the survey, Verité first came up with sector-specific and context-specific definitions explaining how something like limited freedom of movement and communication would manifest in the electronics sector. Researchers found that many of the restrictions for workers in this region and sector centered around victims’ having their daily movements heavily monitored while living in employer-provided housing or around having their passports withheld to restrict travel. The team also found that fees and debt were other major contributors to forced vulnerability. As such, Verité designed survey questions to measure the presence of these indicators.
After the Verité report was published, there were substantial improvements in this sector—although, Klett said, one cannot give full credit to Verité for this improvement because there were other counter-trafficking efforts occurring in tandem. The Electronics Industry Citizenship Coalition and many major multinational electronics brands adopted no-fees policies that prohibited the charging of any recruitment fees to workers in their supply chains, which Klett said is a very substantial policy change for the electronics industry. The Malaysian government officially deemed it illegal to withhold workers’ passports. Verité also saw the proliferation of hotlines and grievance channels for foreign workers.
Klett also talked about a study in Côte d’Ivoire where Verité took a more qualitative approach to identifying deceptive recruitment practices for male migrant workers; Verité is using the results to develop specific recommendations for the industry. She said that companies regularly show interest in this work and request prevalence estimates to assess their own supply chains.
___________________
1 Forced Labor in the Production of Electronic Goods in Malaysia: A Comprehensive Study of Scope and Characteristics. (2014). Available: http://verite.org/wp-content/uploads/2016/11/VeriteForcedLaborMalaysianElectronics2014.pdf.
INTERAGENCY COLLABORATIONS
Several of the organizations with representatives at the workshop have long-standing histories of collaboration on counter-trafficking enterprises, which they hope to build on in order to achieve the goal of creating a cohesive strategy for research on the prevalence of human trafficking.
Global Estimates of Modern Slavery
Michaëlle De Cock (International Labour Organization) told the group that in 2017 the ILO, in collaboration with the Walk Free Foundation and IOM, published the Global Estimates of Modern Slavery.2 To obtain the estimate, they used a combination of methodologies on national datasets in more than 50 countries and on data from IOM’s database on victims of trafficking for forced labor and sexual exploitation. In addition to these global estimates, the ILO is continuing to work with IOM and the UN’s Office of Drugs and Crime to develop a standardized definition and specific tools for the measurement of trafficking for forced labor.
International Conference of Labor Statisticians
De Cock also described work of the International Conference of Labor Statisticians (ICLS), which is a standard-setting body of labor statisticians that has met every 5 years since 1923. She noted that several workshop participants are also part of ICLS. After the ILO presented pilot work on the measurement of trafficking and forced labor at the ICLS meeting in 2013, ICLS member states gave the ILO a mandate to set up a working group to test new measurement tools and create a set of guidelines, taking into account national contexts as well as different economic systems. Those guidelines were formally adopted in 2018, along with a unique statistical definition of forced labor. De Cock said that when member states ask for support with prevalence surveys, it is often with the objective of translating the results into efficient and targeted policies. As such, she feels it is important to have an idea of the scale of trafficking in order to give appropriate guidance.
IOM and the Counter Trafficking Data Collaborative
Sara Crowe (Polaris) said that the idea for the Counter Trafficking Data Collaborative (CTDC) was to create a global data repository of de-identified
___________________
2 International Labour Organization. (2017). Global Estimates of Modern Slavery: Forced Labour and Forced Marriage. Available: https://www.ilo.org/wcmsp5/groups/public/--dgreports/---dcomm/documents/publication/wcms_575479.pdf.
data for access by researchers and policy makers (see Chapter 5). This repository required normalizing very large datasets from Polaris and from IOM that had been collected for more than 10 and, in some cases, more than 20 years. She said CTDC acknowledged and was transparent about the limitations associated with combining sensitive data from multiple sources, and she stressed the primacy of not compromising subjects’ data privacy.
CTDC is a combined dataset of 91,000 records; its two primary outputs are a visual dashboard for aggregate data and a downloadable, de-identified and disaggregated dataset. Crowe said that creating the repository required the three primary organizations, Polaris, IOM, and Liberty Shared,3 to reconcile definitional differences: Polaris uses the definitions in the U.S. federal Trafficking Victims Protection Act; IOM and Liberty Shared use European definitions that include forced marriage and organ trafficking. There are also differences in how the data are collected—in person or by telephone or other means. The agencies ultimately agreed on 73 shared variables and aligned values, including demographic information and such contextual information as the type of trafficking, means of control, and relationship to the recruiter. The other definitional differences are accounted for in the CTDC codebook.
Harry Cook (International Organization for Migration) told participants that IOM, which is the U.N.’s migration agency, provides assistance to upwards of 8,000 trafficking victims every year in more than 150 countries. The objective of IOM, and CTDC to which it belongs, is to create a greater culture of openness around data. He said it is important to be clear about the reason the data are being collected, by whom, and for what purpose, and that IOM has been working on anonymization standards to help navigate data-sharing obstacles.
Cook noted that CTDC is aiming to establish standards around more nuanced forms of exploitation, such as begging and sex work, which includes accounting for selection bias in the case of the data. Cook defined data standards as documented agreements on representation, format, definition, structuring, tagging, transmission, manipulation, use, and management. He said that standards provide the ability to combine data from different sources, to interpret the data together, and to ensure that they are actually interoperable—three components Cook thinks are critical to understanding the causes of human trafficking. He added that measuring the extent to which there is bias is challenging when the data are being compared with a figure that is yet unknown.
___________________
3 The goal of Liberty Shared is to prevent human trafficking through legal advocacy, technological interventions, and strategic collaborations with NGOs and corporations around the world. See https://libertyshared.org.
Cook noted that CTDC has done a lot of research around anonymization. The datasets are de-identified by removing personal identifying information and other hard identifiers, and smaller categories (such as small geographic areas) are grouped into larger regions to further preserve anonymity. After the de-identification, they K-anonymize the dataset with a K equal to 11 so that it is impossible to query the data and return fewer than 11 cases. Specific combinations of values that yield less than 11 results are removed. Cook said that a K equal to 11 is very high and results in severe data loss—a dataset of more than 91,000 cases gets reduced to around 59,000 after redaction—but the loss ratio decreases as the amount of detail added to the dataset increases.
Cook showed participants examples of interactive maps on the CTDC website that display trafficking data by geography, type of trafficking, and certain demographic characteristics. The visualizations require additional anonymization protocol to protect the identities of unique cases, such as those in remote geographic areas or areas where the figures are low for a particular demographic characteristic. CTDC keeps the detail at a high level to prevent the type of cross-filtering that would enable specific queries. He noted that applying differential privacy techniques would bolster anonymity, but the process is costly.
Cook told participants that data governance across sectors is another key component to maintaining high standards for data collection. CTDC has also considered how to implement standards and structures that could improve the use of multiple systems estimation for human trafficking data. One issue that has been identified is massive directionality in the referral identification system; accounting for the directionality in the data contingency tables is one way to mitigate its effect.
Delta 8.7, Earthtime, and Code 8.7
Kelly Gleason (United Nations University Centre for Policy Research [UNU-CPR]) told the group about Delta 8.7, an innovative online tool created by the research center to help policy actors efficiently access and digest data most relevant to policy-making efforts and to identify what type of policy efforts have been most effective eradicating forced labor, child labor, modern slavery, and human trafficking. The Delta 8.7 website has three primary components: a forum with articles by researchers that showcase methods, measures, tools, and organizational efforts; a resources page that provides details on upcoming learning opportunities, events, and a glossary of related terms; and the data and measurement section, which contains country data dashboards and guidance references on measuring concepts related to Sustainable Development Goal Target 8.7 (see Chapter 1).
Gleason said Delta 8.7 has data dashboards for 45 countries that display an overview of nationally representative prevalence measures, sourced from the ILO, at the country level. Each dashboard displays data on contextual factors relevant to the region, such as the U.S.’s Human Development Index,4 and various labor indicators and social protections. Along with the visualizations, developers built in summaries to help users of all expertise levels understand the scope of the problem and how it has changed over time. The dashboards also contain data from partnering institutions regarding policy response and relevant government programs. Gleason said UNU-CPR has solicited constructive feedback on the data presented in Delta 8.7 from statistical offices in partner countries.
Gleason described two more initiatives at UNU-CPR. The first initiative—Earthtime—is a collaborative project with Carnegie Mellon University that shows progress on Target 8.7 issues by geographic area over time. She said Earthtime overlays data onto the world map in a way that tells a story and showcases relationships among datasets. Earthtime tells two stories: one that shows progress on estimating the prevalence of human trafficking and one that shows the ways governments have responded to human trafficking issues through aid, policy, and other activities, and how those interventions have changed over time.
The second initiative—Code 8.7—is a collaboration among multiple institutions focused on finding technological solutions from the computational science and artificial intelligence research communities to eradicate modern slavery. Code 8.7 was designed with the purpose of building out shared data sources and collection frameworks, mapping new technology pipelines, and mapping new research opportunities. Gleason said that the primary objective of all UNU-CPR collaborations is to support the collection and provision of the data in a way that benefits both the research community and data providers while always keeping the focus on the victims and their sustained freedom.
Research to Action Project
Gleason also described the Research to Action Project, which is an ILO initiative funded by the U.S. Department of Labor designed to examine the state of available child labor, human trafficking, and forced labor data and research; highlight gaps in knowledge; and communicate findings to policy makers. The project, which involves input from IOM and UNU-CPR, is intended to build greater research capacity and interest among researchers across disciplines to study forced labor, child labor, and human trafficking through grants and online educational tools. Gleason
___________________
4 See http://hdr.undp.org/en/content/human-development-index-hdi.
said that the project is mapping past and current research efforts to identify funding sources, which will inform national and global research agendas. The project’s output will be consolidated into a tool, intended to educate policy actors on how different types of evidence can illuminate different characteristics of a research issue.
Gleason mentioned four challenges associated with successfully accessing, using, and communicating about existing data or administrative records: siloed information, lack of standardized definitions, privacy and confidentiality for sensitive data, and communication. Siloed information, whether separated by institution or topic area, can prevent a researcher from understanding the extent of data collection efforts for a given topic. The lack of standardized definitions of human trafficking makes it difficult to synthesize the research. Privacy and confidentiality protections for sensitive data, while necessary, create barriers to access. Lastly, even when the data are available and of good quality, it can be difficult to communicate the information in a way that conveys a meaningful message.
Gleason then noted five opportunities to enhance interagency coordination. First, she emphasized, is the importance of making sure collaborative efforts reach across disciplines and across the community, both to combine resources and mitigate the duplication of research efforts. Second, once datasets are combined, she said it is important to carefully investigate the relationships that are observed and to interpret the findings, being mindful of context-specific factors (i.e., units of measurement, timing) and discerning between correlation and causation. Third, she said, the research should directly benefit those providing the data. Care providers are a crucial resource; in order to support the collection and provision of data in a way that researchers can use, Gleason said communication should be a two-way street. Fourth, she encouraged researchers to remember that the goal in collaborating and seeking patterns in the data is to promote sustained freedom for the victims and not simply reduce the prevalence of human trafficking. Lastly, researchers should identify novel sources of data to help understand what is still unknown, and they should think about innovative and responsible new ways to communicate about issues of vulnerability.
This page intentionally left blank.