Using data from an enumerated network of worldwide flight connections between airports, we examine how sampling designs and sample size influence network metrics. Specifically, we apply three types of sampling designs: simple random sampling, nonrandom strategic sampling (i.e., selection of the largest airports), and a variation of snowball sampling. For the latter sampling method, we design what we refer to as a controlled snowball sampling design, which selects nodes in a manner analogous to a respondent-driven sampling design. For each design, we evaluate five commonly used measures of network structure and examine the percentage of total air traffic accounted for by each design. The empirical application shows that (1) the random and controlled snowball sampling designs give rise to more efficient estimates of the true underlying structure, and (2) the strategic sampling method can account for a greater proportion of the total number of passenger movements occurring in the network.
Abstract Multiple systems estimation using a Poisson loglinear model is a standard approach to quantifying hidden populations where data sources are based on lists of known cases. Information criteria are often used for selecting between the large number of possible models. Confidence intervals are often reported conditional on the model selected, providing an over-optimistic impression of estimation accuracy. A bootstrap approach is a natural way to account for the model selection. However, because the model selection step has to be carried out for every bootstrap replication, there may be a high or even prohibitive computational burden. We explore the merit of modifying the model selection procedure in the bootstrap to look only among a subset of models, chosen on the basis of their information criterion score on the original data. This provides large computational gains with little apparent effect on inference. We also incorporate rigorous and economical ways of approaching issues of the existence of estimators when applying the method to sparse data tables.
Much media coverage in recent years has identified the abuses of migrant workers in the Gulf Cooperation Council (GCC) countries (e.g. Bahrain, Kuwait, Oman, Qatar, United Arab Emirates [UAE], and Saudi Arabia), an increasingly popular migrant destination for Kenyan workers. Due to their unique geo-political and economic profiles, these countries with widespread human rights violations have remained non-responsive primarily to pressure from the international community. Importantly, these factors mask another critical gap in human trafficking research: the lack of reliable data and how best to collect it threatens the effectiveness of existing and proposed interventions in the GCC. To address this gap, this team sought to measure the prevalence of forced labor among recently returned Kenyan migrant workers from GCC countries currently residing in the Nairobi Metro area. Using rigorous estimation strategies in our data collection and analysis, we found that forced labor was pervasive among this population. Practically every Kenyan migrant worker who worked in GCC countries during the studied period could be considered a victim of forced labor. Most striking was the consistency in the high rates of violations across all measures, regardless of which set of indicators we applied in our analysis. Although abuses among migrant workers are not uncommon in wealthy nations, such high rates of forced labor violations in GCC countries are truly rare, if not unprecedented in current prevalence estimation research, and call for massive systemic efforts to address the situation.
Photo by Grace Forrest, Walk Free Foundation, www.by-grace.org.The research community faces two major challenges in estimating the global prevalence of human trafficking for sexual exploitation: co...
We present a new design and method for estimating the size of a hidden population best reached through a link-tracing design. The design is based on selecting initial samples at random and then adaptively tracing links to add new members. The inferential procedure involves the Rao–Blackwell theorem applied to a sufficient statistic markedly different from the usual one that arises in sampling from a finite population. The strategy involves a combination of link-tracing and mark-recapture estimation methods. An empirical application is described. The result demonstrates that the strategy can efficiently incorporate adaptively selected members of the sample into the inferential procedure. Supplementary materials for this article are available online.
The goal of this paper is to compare a traditional survey method with the network scale-up method (NSUM) for the prevalence estimation of child trafficking in Sierra Leone in 2020. The traditional survey method involved a probability-based, stratified, and clustered multistage sampling design in which adult respondents in 3,070 households were interviewed about trafficking of children who reside in their household in three selected districts. This paper details the first attempt to estimate the prevalence of child trafficking using NSUM, which entailed questioning the same adult respondents about the trafficking-related activities of children in their personal networks. Findings and interpretation of these results are presented, along with implications and recommendations for future studies.
Le calage est une methode de redressement qui utilise de l’information sur la distribution d’un echantillon et de la population nationale pour determiner la ponderation des participants a une enquete. Le calage vise a ponderer un echantillon afin que sa composition demographique soit representative de la population cible.
Sample calibration is a procedure that utilizes sample and national-level demographic distribution information to weight survey participants. The objective of calibration is to weight the sample so that it is demographically representative of the target population. This technical report details our calibration analysis for the 2013 Methods-of-Payment survey questionnaire sample. The analysis makes use of a variety of variables, with corresponding distributions from the 2011 National Household Survey and 2012 Canadian Internet Use Survey. Our primary objective is to seek a sensible set of variables for calibration and to propose a set of final weights that meet a validation criterion. A raking ratio calibration procedure is used in the analysis. We base calibration on candidate variables and nesting of pairs of variables chosen within the context of the study. An imputation strategy is implemented to account for the relatively few missing observations. Three samples are obtained for the survey and we summarize an analysis that suggests that calibration should be based on the full/collapsed data set. We describe our research on several validation criteria and, after testing the calibration procedure, report our proposed set of final weights.
ABSTRACTFinancial wellbeing – broadly characterized as a liveable income, savings, and autonomy over financial decisions – has been shown to influence physical and psychological health and is therefore an important element of holistic wellbeing. The present study examines the factors that impact the financial wellbeing of survivors of human trafficking in the United States. Using survey data from a sample of trafficking survivors in the U.S. we find that both temporal distance from the exploitative experience and a stable source of income significantly predict financial wellbeing for trafficking survivors. Importantly, however, stable work that provides predictable income mitigates the impact of time in establishing financial wellbeing. The implications of this research are significant for policy and programs aimed at improving the livelihoods of trafficking survivors.KEYWORDS: TraffickingLivelihoodFinancial WellbeingRegression AnalysisParticipatory Action Research AcknowledgmentsThe authors would like to thank individual donors, corporations, and foundations who contribute to Polaris’ unrestricted fund, including partial funding for the National Survivor Study from IHG Hotels & Resorts, Match Group, PayPal, and United Way Worldwide. The authors would also like to thank the National Survivor Study core research team and community advisory group, including: Lara Powers, Katherine T. Bright, Tristan Call, Michael Chen, Hazel Fasthorse, Tawana Bandy Fattah, La Toya Gix, Elizabeth Jacobs, Forrest Jacobs, Ashley Maha’a, Erin Marsh, Namrita S. Singh, Karen Snyder, Lauren Vollinger, Charity Watters, Wade Arvizu, Marlene Carson, Harold D’Souza, Wang Fen, and Eric Harris.Disclosure statementNo potential conflict of interest was reported by the author(s).Supplementary materialSupplemental data for this article can be accessed online at https://doi.org/10.1080/23322705.2023.2257126Notes1 For example, women, children, racial and ethnic minorities, LGBTQ+, disabled and migrants have all been groups identified as being vulnerable or at higher risk to trafficking victimization and/or are disproportionately represented among trafficking victims compared to their numbers in the general population (see, for example, IOM et al., Citation2022; Jagoe et al., Citation2022; Latham-Sprinkle et al., Citation2019; UNODC, Citation2020; Whitbeck et al., Citation2004).2 While we expect there to be a relationship between stable work and financial wellbeing, when one does not feel competent in their work, does not feel adequately supported, or the environment becomes unduly demanding, overall wellbeing may be jeopardized, which could have negative implications for financial wellbeing. While it is beyond the scope of the present study, we acknowledge that the quality of the stable employment may be just as important as the stable employment itself.3 The survey instrument is available upon request.4 The survey and sampling methodology received ethical approval from the Biomedical Research Alliance of New York (BRANY) Institutional Review Board (IRB). The study also received Certificate of Confidentiality (CoC) through the National Institutes of Health (NIH), which protects the privacy of research participants by prohibiting disclosure of identifiable, sensitive information to anyone not connected to the research team.5 In addition to running these models with the imputed dataset, we also ran all of the models with the original non-imputed dataset. The results were very similar across all of the models making a strong case for the accuracy of our imputation method. Output for these models is available in the appendix.6 There is ample cross-disciplinary precedent for using factor scores as dependent variables. See, for example, Grasso and Simons (Citation2011), Salcioglu et al. (Citation2007), and Vogel et al. (Citation2019).7 There were 13 respondents who answered that they had no income, making the denominator in this equation zero. To avoid this, we recoded $0 as $1000 in income per year.8 To impute missing data for the financial wellbeing dependent variable, which was created from the composite of three factors, two separate imputation procedures were investigated. The first was to directly impute missing data of the underlying factor variables and then create the composite variable. The second was to create the composite variable first and then impute the missing data in the composite variable. It was found that the first approach resulted in a greater imputation accuracy score and was therefore used for imputing the missing entries.9 There were not enough non-binary responses to treat gender as non-binary for the purpose of the statistical analysis. Therefore, we include gender in the model as a binary rather than categorical variable.10 Termed “the motherhood penalty,” the financial impact of children is greater for females than males (e.g., Budig & England, Citation2001; Budig & Hodges, Citation2010; Gangl & Ziefle, Citation2009). Further, single women of children are more deeply impacted financially than men and married women (Bartfeld, Citation2000; De Vaus et al., Citation2017; Leopold, Citation2018; Lichtenstein & Johnson, Citation2019).11 In addition to running these models with the composite measure of financial wellbeing as the dependent variable, we also ran all of the models with each of the factor variables as the dependent variable. The results were very similar across all of the models making a strong case for the financial wellbeing composite measure. Output for these models is available in the appendix.12 The marginal means plots provided are for the most common profile in the sample; however, we provide additional marginal means plots in the Appendix for minority females without children and without a disability for comparison.13 We did not find consistently or strong statistically significant results based on age, race, or gender.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)
