Actionable, Cost-Effective Prevalence Measurement to End Modern Slavery
New and better estimates of the scale of modern slavery at the global level have helped galvanize global commitment to end this scourge. But for the global anti-slavery field to move from commitments to effective interventions and investments, data must be even more precise. Data that highlights the prevalence of slavery within specific sectors and geographies can help interventions target the most vulnerable populations and make the case for private sector action against specific supply chain risks. It can also help shed light on which anti-slavery interventions work.
Organizations engaged in the global fight against trafficking have increased efforts to measure prevalence of slavery within specific industries and geographies, drawing on innovations to address the intrinsic challenges of measuring hidden populations. An early example is the United Nations Inter-Agency Project on Human Trafficking’s use of the sentinel surveillance methodology—using interviews with randomized samples of victims and migrants in key hotspots—in 2010 to study prevalence of human trafficking among Cambodian migrants deported from Thailand. More recently, International Justice Mission (IJM) and its partners conducted a series of sector and sub-national studies, including a 2016 study of Commercial Sexual Exploitation of Children in Mumbai. These studies used respondent-driven sampling (RDS)—a type of snowball sampling that is used to analyse hidden populations—and time-space sampling—a venue-based method that identifies likely hotspots, randomly selects a date and time and asks members of the target population to participate in the research. The Freedom Fund and its partners, including Institute of Development Studies and Praxis in a 2017 baseline study, have also deployed participatory statistics and social mapping in research in Northern India.
All of this work is rapidly advancing understanding in the field, but many more studies are needed to properly inform the efforts required to move the needle on modern slavery. Unfortunately, the cost and time to conduct these studies can be prohibitive. For example, we estimate that a typical prevalence study currently costs $200,000-$500,000 and can take 15-24 months, for a baseline alone.
Inputting data. Unsplash/Stephen Dawson.
The Global Fund to End Modern Slavery (GFEMS) seeks to address this challenge and rapidly expand the evidence base by reducing the time and cost of measuring prevalence by an order of magnitude. We believe it is possible to make this process more efficient and cost-effective through three main levers:
- exploring new estimation methodologies;
- leveraging mobile technology wherever possible; and
- collaborating strategically to leverage skills across different fields, all while working to ensure statistical robustness.
Exploring new estimation methodologies
Among other approaches, GFEMS is currently testing the application of the network scale-up method (NSUM) to estimate the percentage of a given population in particular forms of forced labour. NSUM has been used in the field of public health to estimate the size of hidden populations at risk of contracting HIV and can increase estimation capability and reduce costs because it does not require direct access to hidden populations. Instead, survey respondents in the general population are surveyed to understand the ratio of (a) how many people they know in a hidden population, e.g., intravenous drug users or workers whose wages are regularly withheld, to (b) the size of their personal networks, based on the number of people they know in subgroups of a known size. Statisticians use this ratio to understand the percentage of the overall population that is in the hidden population.
There are, of course, potential biases to account for, such as the barrier effect (it may be more difficult to “know” people in the hidden population) and transmission bias (members of a hidden population may not reveal this to people they know). However, there are ways to correct for bias, such as through Bayesian modelling to understand likelihood of under/overestimation based on other data provided by the respondent. This modelling improves with each repetition of the survey, and the survey can be easily repeated with a different random sample of the general population each time. We do not believe that NSUM will be a silver bullet, but it may be an important new option on a menu of methodologies.
Data that highlights the prevalence of slavery within specific sectors and geographies can help interventions target the most vulnerable populations and make the case for private sector action against specific supply chain risks.
Leveraging mobile technology wherever possible
The increasing penetration of mobile technology provides an unprecedented opportunity to detect forced labour and commercial sexual exploitation, sometimes in real time, especially among migrants. Victims of modern slavery may retain mobile devices, even under conditions of forced labour. In these cases, targeted short-form questions (via phone call, SMS or interactive voice response) can detect incidence of forced labour within weeks of a migrant’s journey and flag the need for direct intervention.
Using mobile technology effectively to detect modern slavery requires large-scale registration of migrants and their family members at a centralized point of departure and asking simple and clear questions appropriate to the specific slavery context.
More broadly, even where victims do not have access to technology, the ability to reach the general population via social media can reduce the time and cost to conduct surveys, as can the use of mobile devices by enumerators. In a pilot test of NSUM in Vietnam, we completed 400 surveys in the first 24 hours over the Facebook mobile app with the support of targeted Facebook ads. Recognizing the biases that may come from Facebook users, we also conducted face-to-face surveys to balance and validate the data.
Leveraging skills across a number of fields
To further streamline prevalence measurement at the Fund, we aim to reduce time and cost through the strategic coordination of multiple partners, methods and tools for data collection, cleaning, analysis and storage. We create a coalition of global and local partners that includes university professors, data scientists and programmers, each with a core strength and a clearly delineated role that supports our prevalence estimations. To maximize coordination, we have enlisted a single global data collection partner who will have a consistent presence across all of the Fund’s research and development activities.
Cost and time remain significant hurdles to rapidly generate robust measures of modern slavery prevalence. And, prevalence measurement is only part of a broader set of essential capabilities, including assessing the impact of specific interventions and analysing return on investment. But we are committed to developing and sharing these capabilities. If we do not rapidly expand the evidence base and understanding of what works, efforts to combat modern slavery will continue to be underfunded and sub-scale, doing a disservice to the millions of people who need it to end.
The Global Fund to End Modern Slavery is a public-private partnership that seeks to catalyse and coordinate a coherent global strategy to end modern slavery by making it economically unprofitable. GFEMS’s strategy includes increasing resources, engaging government and the private sector, funding transformative programmes and technologies and ensuring robust assessment of impact across all partners and programmes.
This article has been prepared by the Global Fund to End Modern Slavery as a contribution to Delta 8.7. As provided for in the Terms and Conditions of Use of Delta 8.7, the opinions expressed in this article are those of the author and do not necessarily reflect those of UNU or its partners.