Bias decomposition and estimator performance in respondent-driven sampling

Respondent-Driven Sampling (RDS) is a method of network sampling that is used to sample hard-to-reach populations. The resultant sample is non-random, but different weighting methods can account for the over-sampling of (1) high-degree individuals and (2) homophilous groups that recruit members more effectively. While accounting for degree-bias is almost universally agreed upon, accounting for recruitment-bias has been debated as it can further increase estimate variance without substantially reducing bias. Simulation-based research has examined which weighting procedures perform best given underlying population network structures, group recruitment differences, and sampling processes. Yet, in the field, analysts do not have a priori knowledge of the network they are sampling. We show that the RDS sample data itself can determine whether a degree-based estimator is sufficient. Formulas derived from the decomposition of a ‘dual-component’ estimator can approximate the ‘recruitment component’ (RC) and ‘degree component’ (DC) of a sample’s bias. Simulations show that RC and DC values can predict the performance of different classes of estimators. Samples with extreme ‘RC’ values, a consequence of network homophily and differential recruitment, are better served by a classical estimator. The use of sample data to improve estimator selection is a promising innovation for RDS, as the population network features that should guide estimator selection are typically unknown.     

Social network analysis with respondent-driven sampling data: A study of racial integration on campus

This paper presents respondent-driven sampling (RDS) as a viable method of sampling and analyzing social networks with survey data. RDS is a network based sampling and analysis method that provides a middle ground compliment to ego-centric and saturated methods of social network analysis. The method provides survey data, similar to ego-centric approaches, on individuals who are connected by behaviorally documented ties, allowing for macro-level analysis of network structure, similar to that supported by saturated approaches. Using racial interaction of university undergraduates as an empirical example, the paper examines whether and to what extent racial diversity at the institutional level is reflected as racial integration at the interpersonal level by testing hypotheses regarding the quantity and quality of cross-race friendships. The primary goal of this article, however, is to introduce RDS to the network community and to stimulate further research toward the goal of expanding the analytical capacity of RDS. Advantages, limitations, and areas for future research to network analysis using RDS are discussed.     

Community understanding of respondent-driven sampling in a medical research setting in Uganda: importance for the use of RDS for public health research

Respondent-driven sampling (RDS) is a widely-used variant of snowball sampling. Respondents are selected not from a sampling frame, but from a social network of existing members of the sample. Incentives are provided for participation and for the recruitment of others. Ethical and methodological criticisms have been raised about RDS. In this study RDS was used to recruit male household heads in rural Uganda. We investigated community members' understanding and experience of the method, and explored how these may have affected the quality of the RDS survey data. Our findings suggest that because participants recruit participants, the use of RDS in medical research may result in increased difficulties in gaining informed consent, and data collected using RDS may be particularly susceptible to bias due to differences in the understanding of key concepts between researchers and members of the community.     

Linked Ego Networks: Improving estimate reliability and validity with respondent-driven sampling

Respondent-driven sampling (RDS) is currently widely used for the study of HIV/AIDS-related high risk populations. However, recent studies have shown that traditional RDS methods are likely to generate large variances and may be severely biased since the assumptions behind RDS are seldom fully met in real life. To improve estimation in RDS studies, we propose a new method to generate estimates with ego network data, which is collected by asking respondents about the composition of their personal networks, such as “what proportion of your friends are married?”. By simulations on an extracted real-world social network of gay men as well as on artificial networks with varying structural properties, we show that the precision of estimates for population characteristics is greatly improved. The proposed estimator shows superior advantages over traditional RDS estimators, and most importantly, the method exhibits strong robustness to the recruitment preference of respondents and degree reporting error, which commonly happen in RDS practice and may generate large estimate biases and errors for traditional RDS estimators. The positive results henceforth encourage researchers to collect ego network data for variables of interests by RDS, for both hard-to-access populations and general populations when random sampling is not applicable.     

Ties,sites and irregularities: pitfalls and benefits in using respondent-driven sampling for surveying a homeless population

Respondent-driven sampling (RDS) is a survey method for hidden populations and, as such, it offers a suitable approach for sampling the homeless. Surprisingly, the practical use of RDS in surveying homeless populations has only sporadically been described in the professional literature so far, and the specifics of using RDS for sampling this group have not been reflected in depth. The goal of the article is to investigate how three specific concerns, namely the transitional character of social ties among the homeless, problems with locating the interview site, and the irregular use of coupons, affect the sampling process. The findings of the article are based on experience using RDS for creating two independent samples of homeless persons in two Czech cities. Worries of poor referral effectiveness were disproved, but poor validity of the network size indicator, tendency towards irregular use of coupons, and negative effects arising from selecting service centres for the homeless as survey sites were confirmed.     

Estimating network structure via random sampling: Cognitive social structures and the adaptive threshold method

This paper introduces and tests a novel methodology for measuring networks. Rather than collecting data to observe a network or several networks in full, which is typically costly or impossible, we randomly sample a portion of individuals in the network and estimate the network based on the sampled individuals’ perceptions on all possible ties. We find the methodology produces accurate estimates of social structure and network level indices in five different datasets. In order to illustrate the performance of our approach we compare its results with the traditional roster and ego network methods of data collection. Across all five datasets, our methodology outperforms these standard social network data collection methods. We offer ideas on applications of our methodology, and find it especially promising in cross-network settings.     

Using respondent-driven sampling with ‘hard to reach’ marginalised young people: problems with slow recruitment and small network size

This paper documents an experience of using respondent-driven sampling (RDS) to recruit socially marginalised young people in Sydney, Australia. Respondents were young people aged 16–24?years who were current illicit drug users and who reported at least one feature of social marginalisation (e.g. recent homelessness or juvenile detention). Four seeds initiated the sampling and 61 respondents were recruited until the sampling was closed due to slow progress at week nine. The paper examines: (1) the overall success of RDS and compares this with similar RDS studies; and (2) the sufficiency of network ties among respondents. The analyses suggest that RDS was generally successful in that, despite its small size, the sample achieved adequately long recruitment chains and variables converged to equilibrium. Nevertheless, recruitment was much slower than comparable studies. This could be due to the study population having reduced willingness to participate, a high proportion of respondents who did not fit the selection criteria, and small and disparate networks. Using RDS with marginalised youth may require generous resourcing to allow large incentives to increase willingness, and a lengthy recruitment period. Moreover, the small networks suggest that researchers should start the sampling with a large number of seeds.     

Migration Plans and Received Remittances: Evidence from Fiji and Tonga1

This paper investigates the relationship between remittances and migration intentions in the source country using data from households in the Pacific island countries of Tonga and Fiji. Unlike previous research it accounts for potential endogeneity in the relationship between remittances and intentions and also includes households that receive remittances but do not have current migrants. We find a positive impact of remittances on intentions. The impact of remittances on intentions is stronger in a country with a longer history of migration and stronger in households with current migrants than those without migrants. These results provide new evidence on the role of remittances in the development of migration chains.     

One world is not enough: the structured phenomenology of lifestyle migrants in East Asia

The paper is based on original empirical research into the lifestyle migration of European migrants, primarily British, to Thailand and Malaysia, and of Hong Kong Chinese migrants to Mainland China. We combine strong structuration theory (SST) with Heideggerian phenomenology to develop a distinctive approach to the interplay between social structures and the lived experience of migrants. The approach enables a rich engagement with the subjectivities of migrants, an engagement that is powerfully enhanced by close attention to how these inner lives are deeply interwoven with relevant structural contexts. The approach is presented as one that could be fruitfully adopted to explore parallel issues within all types of migration. As is intrinsic to lifestyle migration, commitment to a better quality of life is central to the East Asian migrants, but they seek an uncomplicated, physically enhanced texture of life, framed more by a phenomenology of prosaic well‐being than of self‐realization or transcendence. In spite of possessing economic and status privileges due to their relatively elite position within global structures the reality for a good number of the lifestyle migrants falls short of their prior expectations. They are subject to particular kinds of socio‐structural marginaliszation as a consequence of the character of their migration, and they find themselves relatively isolated and facing a distinct range of challenges. A comparison with research into various groups of migrants to the USA brings into relief the specificities of the socio‐structural positioning of the lifestyle migrants of the study. Those East Asian migrants who express the greatest sense of ease and contentment seem to be those who have responded creatively to the specific challenges of their socio‐structural situation. Often, this appears to have been achieved through understated but active involvements with their new settings and through sustaining focused transnational connections and relationships.     

Estimating network properties from snowball sampled data

This article addresses the estimation of topological network parameters from data obtained with a snowball sampling design. An approximate expression for the probability of a vertex to be included in the sample is derived. Based on this sampling distribution, estimators for the mean degree, the degree correlation, and the clustering coefficient are proposed. The performance of these estimators and their sensitivity with respect to the response rate are validated through Monte Carlo simulations on several test networks. Our approach has no complex computational requirements and is straightforward to apply to real-world survey data. In a snowball sample design, each respondent is typically enquired only once. Different from the widely used estimator for Respondent-Driven Sampling (RDS), which assumes sampling with replacement, the proposed approach relies on sampling without replacement and is thus also applicable for large sample fractions. From the simulation experiments, we conclude that the estimation quality decreases with increasing variance of the network degree distribution. Yet, if the degree distribution is not to broad, our approach results in good estimates for the mean degree and the clustering coefficient, which, moreover, are almost independent from the response rate. The estimates for the degree correlation are of moderated quality.     

The Quagmire of Return and Reintegration: Challenges to Multi-Stakeholder Co-ordination of Involuntary Returns

The institutional aspect of return migration has received little attention in the theoretical and empirical literature on return migration. This research fills the apparent lacuna by unearthing institutional challenges to multi-stakeholder coordination, at different spatial levels in crisis situations and negative effects on reintegration of forcibly returned migrants. We use the evacuation of Ghanaian migrants from Libya who occupied very low socio-economic positions, experienced racism and discrimination, including physical attacks and arbitrary arrests in 2011, as a case study to understand institutional challenges to forced return when migrants’ carefully tailored plans are thrown into disarray and they are forced to return unprepared. This study employed mainly qualitative research methods among six different categories of actors and engaged an adaptation of Cassarino's “returnee's preparedness framework” to expand theoretical understandings of return migration from the institutional perspective and to highlight what can go wrong when institutions are unprepared for involuntary returnees.     

THE POT RUSH: IS LEGALIZED MARIJUANA A POSITIVE LOCAL AMENITY?

This paper examines the amenity value of legalized marijuana by analyzing the impact of marijuana legalization on migration to Colorado. Colorado is the pioneering state in this area having legalized medical marijuana in 2000 and recreational marijuana in 2012. We test whether potential migrants to Colorado view legalized marijuana as a positive or negative local amenity. We use the synthetic control methodology to examine in- and out-migration to/from Colorado versus migration to/from counterfactual versions of Colorado that have not legalized marijuana. We find strong evidence that potential migrants view legalized marijuana as a positive amenity with in-migration significantly higher in Colorado compared with synthetic-Colorado after the writing of the Ogden memo in 2009 that effectively allowed state laws already in place to be activated, and additionally after marijuana was legalized in 2013 for recreational use. When we employ permutation methods to assess the statistical likelihood of our results given our sample, we find that Colorado is a clear and significant outlier. We find no evidence for changes in out-migration from Colorado suggesting that marijuana legalization did not change the equilibrium for individuals already living in the state. (JEL I18, R23, K42, C22)     

Statistical adjustment of network degree in respondent-driven sampling estimators: Venue attendance as a proxy for network size among young MSM

We introduce a new venue-informed network degree measure, which we applied to respondent-driven sampling (RDS) estimators. Using data collected from 746 young MSM in 2014–2016 in Chicago, IL, and Houston, TX, we estimated the population seroprevalence of HIV and syphilis and risk/protective behaviors, using RDS estimates with self-reported network size as a standard degree measure as well as our proposed venue-informed degree measure. The results indicate that the venue-informed degree measure tended to be more efficient (smaller variance) and less biased than the other measure in both cities sampled. Venue attendance-adjusted network size may provide a more reliable and accurate degree measure for RDS estimates of the outcomes of interest.     

AN EMPIRICAL TEST OF RESPONDENT-DRIVEN SAMPLING: POINT ESTIMATES, VARIANCE, DEGREE MEASURES, AND OUT-OF-EQUILIBRIUM DATA

This paper, which is the first large-scale application of respondent-driven sampling (RDS) to nonhidden populations, tests three factors related to RDS estimation against institutional data using two WebRDS samples of university undergraduates. First, two methods of calculating RDS point estimates are compared. RDS estimates calculated using both methods coincide closely, but variance estimation, especially for small groups, is problematic for both methods. In one method, the bootstrap algorithm used to generate confidence intervals is found to underestimate variance. In the other method, where analytical variance estimation is possible, confidence intervals tend to overestimate variance. Second, RDS estimates are found to be robust against varying measures of individual degree. Results suggest the standard degree measure currently employed in most RDS studies is among the best-performing degree measures. Finally, RDS is found to be robust against the inclusion of out-of-equilibrium data. The results show that valid point estimates can be generated with RDS analysis using real data, but that further research is needed to improve variance estimation techniques.     

Self-selection,earnings and Chicano migration: differences between return and onward migrants

"The purpose of this article is to extend the empirical literature on Chicano return migration by examining earnings differentials between return and onward Chicano migrants. Our approach reflects the complexity of estimating such effects in terms of selectivity biases and the interaction between individual and locational attributes. We use data derived from the public use microdata sample (PUMS) of the 1990 U.S. census. After controlling for migration and labor force self-selection, results indicate that Chicano return migrants are not negatively self-selected. Chicano return migrants have smaller earnings profiles largely due to the negative effects of living in areas with higher concentrations of co-ethnics. Apparently, return migrants, at least in the short run, are willing to accept lower earnings for the nonpecuniary benefits of living in the Southwest."     

Rural Migration,Rapid Growth,and Fear of Crime*

Abstract Although much research on rural “boomtowns” explores differences between rapid‐growth communities and more stable communities, it is logical to consider that residents within rural boomtowns experience community transitions in different ways. We examine a specific outcome, fear of crime, across three categories of community residents with different migration histories: lifetime residents, migrants who joined the boomtown community during its period of rapid growth, and post‐boom period migrants. This perspective is particularly interesting, given the likelihood that these three different categories of residents have had substantially different community experiences. Making use of survey data from two intermountain West communities that represent resource‐dependent transitions during the 1970s and 1980s (Evanston, Wyoming and Delta, Utah), we find that boom migrants express greater fear of crime than longer‐term residents or post‐boom migrants. The findings suggest that the longer‐term decline in fear of crime in “post‐boom” periods is not equal among residents.     

EXTENSIONS OF RESPONDENT-DRIVEN SAMPLING: ANALYZING CONTINUOUS VARIABLES AND CONTROLLING FOR DIFFERENTIAL RECRUITMENT

Respondent-driven sampling (RDS) is a network-based method for sampling hidden and hard-to-reach populations that has been shown to produce asymptotically unbiased population estimates when its assumptions are satisfied. This includes resolving a major concern regarding bias in chain-referral samples—that is, producing a population estimate that is independent of the seeds (initial subjects) with which sampling began. However, RDS estimates are limited to nominal variables, and one of the assumptions required for the proof of lack of bias is the absence of differential recruitment. One aim of this paper is to analyze the role of differential recruitment, quantify the bias it produces, and propose a new estimator that controls for it. The second aim is to extend RDS so that it can be employed to analyze continuous variables in a manner that controls for differential recruitment. The third aim is to describe means for carrying out multivariate analyses using RDS data. The analyses employ data from an RDS sample of 264 jazz musicians in the greater New York metropolitan area, taken in 2002.     

Assessing respondent-driven sampling: A simulation study across different networks

The purpose was to assess RDS estimators in populations simulated with diverse connectivity characteristics, incorporating the putative influence of misreported degrees and transmission processes. Four populations were simulated using different random graph models. Each population was “infected” using four different transmission processes. From each combination of population x transmission, one thousand samples were obtained using a RDS-like sampling strategy. Three estimators were used to predict the population-level prevalence of the “infection”. Several types of misreported degrees were simulated. Also, samples were generated using the standard random sampling method and the respective prevalence estimates, using the classical frequentist estimator. Estimation biases in relation to population parameters were assessed, as well as the variance. Variability was associated with the connectivity characteristics of each simulated population. Clustered populations yield greater variability and no RDS-based strategy could address the estimation biases. Misreporting degrees had modest effects, especially when RDS estimators were used. The best results for RDS-based samples were observed when the “infection” was randomly attributed, without any relation with the underlying network structure.     

The State,Unauthorized Mexican Migration,and Vulnerability in the Workplace

For the last 20 years, migration scholars have generated a number of important empirical insights about the ways in which the state, through the enactment of immigration policies, creates workplace vulnerabilities such as discrimination, harassment, wage theft, workplace raids, and the threat of deportation. Recent studies of illegality also examine the role of the state but do so in a way that explores what legal status means and how it is experienced in everyday lives of migrants marked as “illegal” by the state. This article reviews recent research that shows that the state operates in a gray zone of enforcement that puts migrants in ambiguous social spaces and heightens their vulnerability at work. However, research also finds that migrants find ways to exert their agency in challenging work environments.     

Evaluating the Quality of Online Survey Data Collected in 2018 in the USA: Univariate,Bivariate, and Multivariate Analyses

Collecting survey data in Japan and the USA is difficult for many reasons, including the difficulty of creating complete sampling frames and low response rates. Online surveys, which collect data through online access panels and recruit members using nonprobability sampling, is becoming a popular alternative to traditional survey techniques, both in market and social science research. However, sociological researchers are hesitant to adopt online survey data collection techniques as it is unclear how representative such data are. This article examines the quality of online survey data by comparing online surveys with the two datasets collected using probability sampling, the American Community Survey and the General Social Survey. We conducted univariate, bivariate, and multivariate analyses and investigated whether similarities improve from the univariate analyses to bivariate relationships, and to bivariate relationships with controls. The results show that similarities improved from univariate to bivariate and multivariate analyses, but the results of bivariate and multivariate analyses did not differ. We concluded that data collected via volunteer‐based access panels yielded similar results to the benchmark data, especially when the relationships among variables are examined. We argue that automatically discrediting the data collected via volunteer‐based access panels is not good practice. With careful considerations of data quality, online survey data with nonprobability sampling could advance sociological research by giving researchers more freedom to ask questions at a relatively low cost, with a quick turnaround time, and more opportunities to conduct cross‐national research.