Complex Survey Data Research Articles

Intuitive Joint Priors for Variance Parameters

Variance parameters in additive models are typically assigned independent priors that do not account for model structure. We present a new framework for prior selection based on a hierarchical decomposition of the total variance along a tree structure to the individual model components. For each split in the tree, an analyst may be ignorant or have a sound intuition on how to attribute variance to the branches. In the former case a Dirichlet prior is appropriate to use, while in the latter case a penalised complexity (PC) prior provides robust shrinkage. A bottom-up combination of the conditional priors results in a proper joint prior. We suggest default values for the hyperparameters and offer intuitive statements for eliciting the hyperparameters based on expert knowledge. The prior framework is applicable for R packages for Bayesian inference such as INLA and RStan. Three simulation studies show that, in terms of the application-specific measures of interest, PC priors improve inference over Dirichlet priors when used to penalise different levels of complexity in splits. However, when expressing ignorance in a split, Dirichlet priors perform equally well and are preferred for their simplicity. We find that assigning current state-of-the-art default priors for each variance parameter individually is less transparent and does not perform better than using the proposed joint priors. We demonstrate practical use of the new framework by analysing spatial heterogeneity in neonatal mortality in Kenya in 2010–2014 based on complex survey data.

Bayesian Empirical Likelihood Inference with Complex Survey Data

SummaryWe propose a Bayesian empirical likelihood approach to survey data analysis on a vector of finite population parameters defined through estimating equations. Our method allows overidentified estimating equation systems and is applicable to both smooth and non-differentiable estimating functions. Our proposed Bayesian estimator is design consistent for general sampling designs and the Bayesian credible intervals are calibrated in the sense of having asymptotically valid design-based frequentist properties under single-stage unequal probability sampling designs with small sampling fractions. Large sample properties of the Bayesian inference proposed are established for both non-informative and informative priors under the design-based framework. We also propose a Bayesian model selection procedure with complex survey data and show that it works for general sampling designs. An efficient Markov chain Monte Carlo procedure is described for the required computation of the posterior distribution for general vector parameters. Simulation studies and an application to a real survey data set are included to examine the finite sample performances of the methods proposed as well as the effect of different types of prior and different types of sampling design.

Design-based single-mediator approach for complex survey data

We discuss a two-step approach to test for a mediated effect using data gathered via complex sampling. The approach incorporates design-based multiple linear regressions and a generalized Sobel’s method to test for significance of a mediated effect. We illustrate the applications to a study of nicotine dependence, race/ethnicity and cigarette purchase price among daily smokers in the U.S. The study goal was to assess significance of cigarette purchase price as a mediator in the association between race/ethnicity (non-Hispanic Black/African American, non-Hispanic White) and nicotine dependence measured in terms of the average number of cigarettes smoked per day. The single-mediator model incorporated 18 covariates as control factors. The results indicated a significant mediated effect of cigarette purchase price on the association. However, the relative effect size of 5% indicated low practical significance of the cigarette purchase price as a mediator in the association between race/ethnicity and nicotine dependence. The approach can be modified to studies where data are gathered via other types of complex sampling.

Goodness‐of‐fit tests for distributions estimated from complex survey data

AbstractWe propose nonparametric procedures for comparing the empirical distribution function of data from a complex survey with a hypothesized parametric reference distribution. The hypothesized distribution may be fully specified, or it may be a family with the parameters to be estimated from the data. Of the procedures studied, a modification of the Cramér–von Mises test proposed by Lockhart, Spinelli & Stephens [Lockhart, Spinelli and Stephens, The Canadian Journal of Statistics 2007; 35, 125–133] is supported theoretically and performs well in two simulation studies. The methods are applied to examine the distribution of body mass index in the U.S. National Health and Nutrition Examination Survey. The Canadian Journal of Statistics 47: 409–425; 2019 © 2019 Statistical Society of Canada

Clustering asthma symptoms and cleaning and disinfecting activities and evaluating their associations among healthcare workers.

Asthma is a heterogeneous disease with varying severity and subtypes. Recent reviews of epidemiologic studies have identified cleaning and disinfecting activities (CDAs) as important risk factors for asthma-related outcomes among healthcare workers. However, the complexity of CDAs in healthcare settings has rarely been examined. This study utilized a complex survey dataset and data reduction approaches to identify and group healthcare workers with similar patterns of asthma symptoms, and then explored their associations with groups of participants with similar patterns of CDAs. Self-reported information on asthma symptoms/care, CDAs, demographics, smoking status, allergic status, and other characteristics were collected from 2030 healthcare workers within nine selected occupations in New York City. Hierarchical clustering was conducted to systematically group participants based on similarity of patterns of the 27 asthma symptom/care variables, and 14 product applications during CDAs, separately. Word clouds were used to visualize the complex information on the resulting clusters. The associations of asthma health clusters (HCs) with exposure clusters (ECs) were evaluated using multinomial logistic regression. Five HCs were identified (HC-1 to HC-5), labelled based on predominant features as: "no symptoms", "winter cough/phlegm", "mild asthma symptoms", "undiagnosed/untreated asthma", and "asthma attacks/exacerbations". For CDAs, five ECs were identified (EC-1 to EC-5), labelled as: "no products", "housekeeping/chlorine", "patient care", "general cleaning/laboratory", and "disinfection products". Using HC-1 and EC-1 as the reference groups, EC-2 was associated with HC-4 (odds ratio (OR) = 3.11, 95% confidence interval (95% CI) = 1.46-6.63) and HC-5 (OR = 2.71, 95% CI = 1.25-5.86). EC-3 was associated with HC-5 (OR = 2.34, 95% CI = 1.16-4.72). EC-4 was associated with HC-5 (OR = 2.35, 95% CI = 1.07-5.13). EC-5 was associated with HC-3 (OR = 1.81, 95% CI = 1.09-2.99) and HC-4 (OR = 3.42, 95% CI = 1.24-9.39). Various combinations of product applications like using alcohols, bleach, high-level disinfectants, and enzymes to disinfect instruments and clean surfaces captured by the ECs were identified as risk factors for the different asthma symptoms clusters, indicating that prevention efforts may require targeting multiple products. The associations of HCs with EC can be used to better inform prevention strategies and treatment options to avoid disease progression. This study demonstrated hierarchical clustering and word clouds were useful techniques for analyzing and visualizing a complex dataset with a large number of potentially correlated variables to generate practical information that can inform prevention activities.

Postpartum Cardiovascular Outcomes Among Women With Heart Disease from A Nationwide Study

There is limited data on postpartum maternal postpartum major adverse cardiovascular and cerebrovascular events (MACCE) among women with heart disease (HD) in the US. Therefore, we aimed to determine the prevalence and predictors of MACCE in the US. The Nationwide Readmissions Databases (2010 to 2014) were screened for patients with and without HD undergoing delivery. HD subtypes included cardiomyopathy (CDM), congenital heart disease, valvular heart disease, and pulmonary hypertension. Rates and reasons of 42-day readmission were determined using weighted national estimates. Independent predictors of postpartum MACCE were determined using multivariable logistic regression for complex survey data. We found among 15,273,247 patients hospitalized for delivery, 33,827 had HD (CDM 22.78%, congenital heart disease 45.98%, valvular heart disease 24.81%, and pulmonary hypertension 6.41%). Of these, 5.2% of HD patients and 1.4% of No HD were readmitted. MACCE was higher in HD vs No HD (2.68% vs 0.17%, p <0.0001). Median time to MACCE was 5.6 days (interquartile range 3 to 15 days). CDM had >10% readmission at 42 days. Among HD patients, cardiovascular, infectious, hypertensive syndromes, and complications of pregnancy were the most common reasons for 42-day readmission. MACCE predictors in women with HD included HD subtype, age, insurance status, obesity, eclampsia, postpartum hemorrhage, MACCE during delivery, preterm delivery, and thrombotic complications. In conclusion, among a nationwide analysis, postpartum MACCE was more common among patients with HD especially within 1 week of discharge from delivery. Predictors can be easily screened for by clinicians, including presence of any HD, hypertensive syndromes, age, obesity, and obstetrical events during index hospitalization.

How aggression-related mindsets explain SES-differences in bullying behavior

ABSTRACTPrevious research suggests that lower socioeconomic status (SES) adolescents bully more than their higher-SES peers. This paper tests whether aggression-related mindsets, defined as mindsets that theoretically influence aggressive behavior, explain the relationship between SES and bullying engagement among adolescents. Using a large and diverse dataset of survey responses from secondary students in the U.S. (N = 146,044 students, 30% White, 70% students of color from 5th-12th grade), this study applies structural regression modeling with complex survey data analysis. Results suggest that differences in aggression-related mindsets, including feelings of academic efficacy, feelings of purpose, global self-esteem, academic-status insecurity, school-related anger, and school-related happiness account for almost half of the relationship between SES and bullying. Students’ school-related anger is the strongest direct predictor of bullying (0.88 standard deviation), which suggests that strategies to reduce adolescent bullying are more effective if they teach anger-reducing skills or eliminate the root causes of students’ school-related anger.

The Impact of School Bullying on Physical Activity in Overweight Youth: Exploring Race and Ethnic Differences

About one third of youth in the United States are overweight or obese and African American youth are at an increased risk for pediatric overweight and obesity as well as their complications. Physical activity has been identified as one determinant of overweight and obesity, and school bullying has been found to be associated with decreased physical activity. Guided by the Transactional Stress and Coping Model, this study examines how school bullying might impact the physical activity of white and African American healthy weight and overweight youth. Existing, nationally representative, and complex survey data (N = 4509) from the 2005-2006 United States Health Behavior in School-Aged Children (HBSC) were analyzed using multiple group structural equation modeling to evaluate study questions. Support for the hypothesized model was found such that bullying negatively impacted physical activity by way of increasing internalizing symptoms. Possible evidence for parental support, but not peer support, as a protective factor was also found. Results were generally similar for all groups, though some differences are discussed. School bullying is a risk factor for reduced physical activity, regardless of race-ethnicity and weight. Implications for school health professionals are discussed.

Anaemia among adolescent girls in three districts in Ethiopia

BackgroundAdolescence is characterized by rapid growth and development with a significantly increased need for macro and micronutrients. However, there is little empirical evidence on the burden of anaemia among adolescent girls in developing countries such as Ethiopia. This study aims to address this gap by evaluating the magnitude of anaemia with an aim to guide design of intervention modalities to address anaemia among adolescent girls.MethodsThe study employed a community based cross sectional design. The study was conducted on weekends to capture both in school and out of school adolescent girls. Data was collected from a total 1323 adolescent girls. From each district, we randomly selected villages and ensured that the sampled households had a range geographical spread (lowlands, highlands) within the larger category of rural and urban. We performed anaemia testing using HemoCue B-Haemoglobin analyser. We applied a complex survey data analysis method to estimate the level of anaemia. The hemoglobin level was adjusted for altitude and smoking status. We ran a logistic regression model to evaluate predictors of anaemia.ResultsThe overall anaemia prevalence ranged from 24 to 38%, with an average rate of 29%. Less than half of the girls heard the term anaemia, and about one third knew the relationship between anaemia and the intake of iron rich foods. The risk of anaemia is higher among adolescent girls in their early adolescence period (10–14 years) (Adjusted Odds Ratio (AOR); 1.98; 95% CI; 1.03, 3.82] and among adolescent girls who lived in moderately food insecure households (AOR 1.48; 95% CI; 1.05–2.09). However, knowing the term “anaemia” was found to be protective against the risk of anaemia.ConclusionsThe risk of anaemia was particularly high among adolescent girls in their early age and among those living in food insecure households. The prevalence of anaemia among adolescent girls is a moderate public health problem. According to the WHO set criteria, the districts could be candidates for intermittent iron and Folic acid supplementation program.

The 12 city HIV Surveillance Survey among MSM in Brazil 2016 using respondent-driven sampling: a description of methods and RDS diagnostics.

This paper details the methods used in the second national Biological and Behavioral Surveillance Survey (BBSS) of HIV, syphilis, and hepatitis B and C among men who have sex with men in Brazil. Respondent-driven sampling (RDS) was used in 12 cities in 2016. The targeted sample size was initiated with five to six seeds in each city. HIV, syphilis, and Hepatitis B and C rapid tests were offered to participants. RDS Analyst with Gile's successive sampling (SS) estimator was used to adjust results as recommended and a weight for each individual was generated for further analysis. Data for the 12 cities were merged and analyzed using Stata 14.0 complex survey data tools with each city treated as its own stratum. Duration of data collection varied from 5.9 to 17.6 weeks. 4,176 men were recruited in the 12 cities. Two sites failed to achieve targeted sample size due to a six-month delay in local IRB approval. No city failed to reach convergence in our major outcome variable (HIV). The comprehensive BBSS was completed as planned and on budget. The description of methods here is more detailed than usual, due to new diagnostic tools and requirements of the new STROBE-RDS guidelines.

Using mixed effects logistic regression models for complex survey data on malaria rapid diagnostic test results

BackgroundThe effect of malaria in Nigeria is still worrisome and has remained a leading public health issue in the country. In 2016, Nigeria was the highest malaria burden country among the 15 countries in sub-Saharan Africa that accounted for the 80% global malaria cases. The purpose of this study is to utilize appropriate statistical models in identifying socio-economic, demographic and geographic risk factors that have influenced malaria transmission in Nigeria, based on malaria rapid diagnostic test survey results. This study contributes towards re-designing intervention strategies to achieve the target of meeting the Sustainable Development Goals 2030 Agenda for total malaria elimination.MethodsThis study adopted the generalized linear mixed models approach which accounts for the complexity of the sample survey design associated with the data. The 2015 Nigeria malaria indicator survey data of children between 6 and 59 months are used in the study.ResultsFrom the findings of this study, the cluster effect is significant (P<0.0001) which has suggested evidence of heterogeneity among the clusters. It was found that the vulnerability of a child to malaria infection increases as the child advances in age. Other major significant factors were; the presence of anaemia in a child, an area where a child resides (urban or rural), the level of the mother’s education, poverty level, number of household members, sanitation, age of head of household, availability of electricity and the type of material for roofing. Moreover, children from Northern and South-West regions were also found to be at higher risk of malaria disease and re-infection.ConclusionImprovement of socio-economic development and quality of life is paramount to achieving malaria free Nigeria. There is a strong link of malaria risk with poverty, under-development and the mother’s educational level.

Quantile Regression Analysis of Survey Data Under Informative Sampling.

For complex survey data, the parameters in a quantile regression can be estimated by minimizing an objective function with units weighted by the original design weights. However, when the complex survey sampling design is informative (i.e., when the design weights are correlated with the study variable even after conditioning on other covariates), the efficiency of this design-weighted estimator may be improved. In this article, we propose several weight-smoothing estimators for quantile regression analysis of complex survey data collected with an informative sampling design. Our new estimators incorporate nonparametric methods for modeling the weight functions and pseudo-population bootstrap methods for variance estimation. A simulation study compares, our proposed methods with the original design-based method in terms of bias, standard error, mean squared error, and confidence coverage. Our proposed estimators have smaller bias and mean squared error than does the design-based estimator. We further illustrate and compare estimators for the 1988 US National Maternal and Infant Health Survey.

Some Theoretical and Practical Aspects of Empirical Likelihood Methods for Complex Surveys

SummaryThis paper provides an overview on two parallel approaches to design‐based inference with complex survey data: the pseudo empirical likelihood methods and the sample empirical likelihood methods. The general framework covers parameters defined through smooth or non‐differentiable estimating functions for analytic use of survey data as well as descriptive finite population parameters, and the theory focuses on point estimation, hypothesis tests and variable selection under an arbitrary sampling design. Major practical issues for the implementation of the methods, including computational algorithms, are briefly discussed. Results from simulation studies to compare the finite sample performances of the two approaches are presented.

Determinants of inadequate minimum dietary diversity among children aged 6\u201323\xa0months in Ethiopia: secondary data analysis from Ethiopian Demographic and Health Survey 2016

BackgroundInadequate feeding practices are a significant reason for the onset of malnutrition in young children, and their consequences are one of the major obstacles to sustainable socioeconomic development and poverty reduction. Dietary diversity is one of the useful indicators to assess the nutrient adequacy and can examine how different food groups contribute to the nutrient adequacy of the diet in a specific area. Minimum dietary diversity is the intake of at least four food types from the seven categories.MethodsSecondary data analysis of Ethiopian Demographic health survey of 2016 was conducted to explore significant predictors that make children inappropriate to meet minimum dietary diversity. There were 2972 weighted samples, and we have used “SVY” command by STATA 14.0 during data analysis to run the complex survey data. This study has identified the possible factors of inadequate minimum dietary diversity of children.ResultsThe proportion of inadequate minimum dietary diversity in Ethiopia was found 85.1%. Frequency of reading newspaper or magazine, frequency of listening to radio, father’s educational level and household wealth index were found significant predictors to determine the minimum dietary diversity of children. Dairy products and grain, roots and tubers account more than half of consumed foods. Among breastfed children who attained minimum dietary diversity, majority of them were in the age group of 6–11 months.ConclusionsMinimum dietary diversity is still low in Ethiopia, and most of mothers feed their child the most and easy accessible food rather than of diverse food. In the way of addressing the Sustainable Development Goal, Ethiopia requires substantial improvement in complementary feeding practices. Appropriate infant and young child feeding messages should to be developed and delivered through mass media.

Statistical Inference on Health Disparity Indices for Complex Surveys.

The National Cancer Institute developed the Health Disparities Calculator (HD*Calc) to facilitate research on health disparities. HD*Calc calculates multiple measures of health disparities using data collected from population-based disease surveillance systems, such as cancer registries. In this paper, we extend the use of HD*Calc to complex survey data by developing plug-in point estimators and Taylor linearization variance estimators that consider complex designs: stratification, multistage clustering, and differential weighting. Our simulation indicates that the plug-in estimators are approximately unbiased and the Taylor linearization variance estimators are accurate. Using 2011-2016 data from the National Health and Nutrition Examination Survey, we demonstrate the use of these estimators in evaluating socioeconomic disparities in the prevalence of child and adolescent (ages 2-18 years) obesity in the United States. Statistical software has been developed for ease of disparity analyses using complex survey data.

Methods for analysis of complex survey data: An application using the Tanzanian 2015 Demographic and Health Survey and Service Provision Assessment

Methods for analysis of complex survey data: An application using the Tanzanian 2015 Demographic and Health Survey and Service Provision Assessment

Clustering analysis of line indices for LAMOST spectra with AstroStat

The application of data mining in astronomical surveys, such as the Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST) survey, provides an effective approach to automatically analyze a large amount of complex survey data. Unsupervised clustering could help astronomers find the associations and outliers in a big data set. In this paper, we employ the k-means method to perform clustering for the line index of LAMOST spectra with the powerful software AstroStat. Implementing the line index approach for analyzing astronomical spectra is an effective way to extract spectral features for low resolution spectra, which can represent the main spectral characteristics of stars. A total of 144 340 line indices for A type stars is analyzed through calculating their intra and inter distances between pairs of stars. For intra distance, we use the definition of Mahalanobis distance to explore the degree of clustering for each class, while for outlier detection, we define a local outlier factor for each spectrum. AstroStat furnishes a set of visualization tools for illustrating the analysis results. Checking the spectra detected as outliers, we find that most of them are problematic data and only a few correspond to rare astronomical objects. We show two examples of these outliers, a spectrum with abnormal continuumand a spectrum with emission lines. Our work demonstrates that line index clustering is a good method for examining data quality and identifying rare objects.

Measuring model misspecification: Application to propensity score methods with complex survey data

Model misspecification is a potential problem for any parametric-model based analysis. However, the measurement and consequences of model misspecification have not been well formalized in the context of causal inference. A measure of model misspecification is proposed, and the consequences of model misspecification in non-experimental causal inference methods are investigated. The metric is then used to explore which estimators are more sensitive to misspecification of the outcome and/or treatment assignment model. Three frequently used estimators of the treatment effect are considered, all of which rely on the propensity score: (1) full matching, (2) 1:1 nearest neighbor matching, and (3) weighting. The performance of these estimators is evaluated under two different sampling designs: (1) simple random sampling (SRS) and (2) a two-stage stratified survey. As the degree of misspecification of either the propensity score or outcome model increases, so does the bias and the root mean square error, while the coverage decreases. Results are similar for the simple random sample and a complex survey design.

HIV prevalence among men who have sex with men in Brazil: results of the 2nd national survey using respondent-driven sampling.

This paper reports human immuno-deficiency virus (HIV) prevalence in the 2nd National Biological and Behavioral Surveillance Survey (BBSS) among men who have sex with men (MSM) in 12 cities in Brazil using respondent-driven sampling (RDS).Following formative research, RDS was applied in 12 cities in the 5 macroregions of Brazil between June and December 2016 to recruit MSM for BBSS. The target sample size was 350 per city. Five to 6 seeds were initially selected to initiate recruitment and coupons and interviews were managed online. On-site rapid testing was used for HIV screening, and confirmed by a 2nd test. Participants were weighted using Gile estimator. Data from all 12 cities were merged and analyzed with Stata 14.0 complex survey data analysis tools in which each city was treated as its own strata. Missing data for those who did not test were imputed HIV+ if they reported testing positive before and were taking antiretroviral therapy.A total of 4176 men were recruited in the 12 cities. The average time to completion was 10.2 weeks. The longest chain length varied from 8 to 21 waves. The sample size was achieved in all but 2 cities.A total of 3958 of the 4176 respondents agreed to test for HIV (90.2%). For results without imputation, 17.5% (95%CI: 14.7–20.7) of our sample was HIV positive. With imputation, 18.4% (95%CI: 15.4–21.7) were seropositive.HIV prevalence increased beyond expectations from the results of the 2009 survey (12.1%; 95%CI: 10.0–14.5) to 18.4%; CI95%: 15.4 to 21.7 in 2016. This increase accompanies Brazil's focus on the treatment to prevention strategy, and a decrease in support for community-based organizations and community prevention programs.

DINÂMICA INTERGERACIONAL DE EDUCAÇÃO E CORRESIDÊNCIA ENTRE PAIS E FILHOS ADULTOS NO BRASIL

Este artigo busca analisar a mobilidade intergeracional de educação no Brasil considerando sua possível inter-relação com a decisão de corresidência entre filhos adultos e pais no mesmo domicílio. Foram usados os dados da Pesquisa Nacional por Amostra de Domicílio (PNAD) de 1996 incorporando sua amostragem complexa e um modelo empírico de decisão conjunta de corresidência e acumulação de capital humano com autosseleção de filhos de acordo com sua posição familiar. Os resultados sugerem que os coeficientes de persistência educacional não podem ser generalizados devido ao viés de seleção amostral. Ademais, foram encontradas evidências de uma maior mobilidade para o grupo de filhos corresidentes com os pais na vida adulta em relação a filhos que são chefes ou cônjuges nos domicílios pesquisados.