A unified approach based on multidimensional scaling for calibration estimation in survey sampling with qualitative auxiliary information

Survey calibration is a widely used method to estimate the population mean or total score of a target variable, particularly in medical research. In this procedure, auxiliary information related to the variable of interest is used to recalibrate the estimation weights. However, when the auxiliary information includes qualitative variables, traditional calibration techniques may be not feasible or the optimisation procedure may fail. In this article, we propose the use of linear calibration in conjunction with a multidimensional scaling-based set of continuous, uncorrelated auxiliary variables along with a suitable metric in a distance-based regression framework. The calibration weights are estimated using a projection of the auxiliary information on a low-dimensional Euclidean space. The approach becomes one of the linear calibration with quantitative variables avoiding the usual computational problems in the presence of qualitative auxiliary information. The new variables preserve the underlying assumption in linear calibration of a linear relationship between the auxiliary and target variables, and therefore the optimal properties of the linear calibration method remain true. The behaviour of this approach is examined using a Monte Carlo procedure and its value is illustrated by analysing real data sets and by comparing its performance with that of traditional calibration procedures.

Modeling and presentation of vaccination coverage estimates using data from household surveys

It is becoming increasingly popular to produce high-resolution maps of vaccination coverage by fitting Bayesian geostatistical models to data from household surveys. Usually, the surveys adopt a stratified cluster sampling design. We discuss a number of crucial choices with respect to two key aspects of the map production process: the acknowledgement of the survey design in modeling, and the appropriate presentation of estimates and their uncertainties. Specifically, we consider the importance of accounting for urban/rural stratification and cluster-level non-spatial excess variation in survey outcomes, when fitting geostatistical models. We also discuss the trade-off between the geographical scale and precision of model-based estimates, and demonstrate visualization methods for mapping and ranking that emphasize the probabilistic interpretation of results. A novel approach to coverage map presentation is proposed to allow comparison and control of the overall map uncertainty. We use measles vaccination coverage in Nigeria as a motivating example and illustrate the different issues using data from the 2018 Nigeria Demographic and Health Survey.

Cost-efficient case-control cluster sampling designs for population-based epidemiological studies

Cost-efficient sampling schemes for population-based case-control studies are necessary for sampling subjects in geographically dispersed populations where in-house surveys are expensive to conduct due to high interviewer travel costs that may be associated with simple random sampling. Motivated by the original study conducted by Cockburn et al. (2011) that investigated the relationship between exposure to pesticides and prostate carcinogenesis, a set of cluster-based individually matched case-control designs is presented for cost-efficient sampling of additional controls. Based on cluster sampling from the field of survey sampling, the case-control study designs presented, where one case is individually matched to three controls, use case-control status in the sampling of the primary sampling clusters. In the secondary stage, interviewer travel costs are reduced by subsampling additional controls within primary sampling clusters as opposed to selecting additional controls purely at random, which would be highly inefficient from a cost perspective. Compared to the simple random sampling (SRS) 1:1 and SRS 1:3 (one case matched to: n SRS control(s)) designs, computer simulations demonstrate that these cluster-based designs provide unbiased rate ratio estimation and statistical efficiencies that are no worse than the SRS 1:1 design and moderately less than the SRS 1:3 design. Even under situations where the intracluster correlation for the exposure variable is extremely high for the exposure of interest, the cluster-based designs have statistical efficiencies that are comparable to that of the SRS 1:1 design. Furthermore, a cost-efficiency analysis is presented that demonstrates that the cluster-based designs are more cost-efficient compared to the SRS 1:3 design.

Motivational Benefits of Social Support and Behavioural Interventions for Smoking Cessation

This study examined the role of social support and behavioral interventions used during the last unsuccessful quit attempt for smokers' intentions to quit smoking within the next 6 months, and identified smokers' attributes associated with use of social support and behavioral interventions. The analytic sample included 7,195 adult daily smokers who responded to the 2010-2011 Tobacco Use Supplement to the Current Population Survey, conducted in the US, and indicated having a serious quit attempt in the past 12 months. Smokers who relied on social support from friends and family had higher odds of intending to quit than those who did not (OR= 1.39, 95% CI= 1.22:1.58), and smokers who used interventions had higher odds of intending to quit than those who did not (OR= 1.36, 95% CI= 1.07:1.74). These associations were similar for both sexes, all age groups, and nicotine dependence levels. Both, relying on social support and use of behavioral interventions were more common among smokers who were female, higher educated, residing in the Western US region, and those who used pharmacological aids for smoking cessation. Social support and behavioral interventions are associated with higher intentions to quit among attempters who relapsed and thus, may aid future smoking cessation.

Empirical likelihood method for non-ignorable missing data problems

Missing response problem is ubiquitous in survey sampling, medical, social science and epidemiology studies. It is well known that non-ignorable missing is the most difficult missing data problem where the missing of a response depends on its own value. In statistical literature, unlike the ignorable missing data problem, not many papers on non-ignorable missing data are available except for the full parametric model based approach. In this paper we study a semiparametric model for non-ignorable missing data in which the missing probability is known up to some parameters, but the underlying distributions are not specified. By employing Owen (1988)'s empirical likelihood method we can obtain the constrained maximum empirical likelihood estimators of the parameters in the missing probability and the mean response which are shown to be asymptotically normal. Moreover the likelihood ratio statistic can be used to test whether the missing of the responses is non-ignorable or completely at random. The theoretical results are confirmed by a simulation study. As an illustration, the analysis of a real AIDS trial data shows that the missing of CD4 counts around two years are non-ignorable and the sample mean based on observed data only is biased.

National weighting of data from the Behavioral Risk Factor Surveillance System (BRFSS)

Background

The Behavioral Risk Factor Surveillance System (BRFSS) is a network of health-related telephone surveys--conducted by all 50 states, the District of Columbia, and participating US territories—that receive technical assistance from CDC. Data users often aggregate BRFSS state samples for national estimates without accounting for state-level sampling, a practice that could introduce bias because the weighted distributions of the state samples do not always adhere to national demographic distributions.

Methods

This article examines six methods of reweighting, which are then compared with key health indicator estimates from the National Health Interview Survey (NHIS) based on 2013 data.

Results

Compared to the usual stacking approach, all of the six new methods reduce the variance of weights and design effect at the national level, and some also reduce the estimated bias. This article also provides a comparison of the methods based on the variances induced by unequal weighting as well as the bias reduction induced by raking at the national level, and recommends a preferred method.

Conclusions

The new method leads to weighted distributions that more accurately reproduce national demographic characteristics. While the empirical results for key estimates were limited to a few health indicators, they also suggest reduction in potential bias and mean squared error. To the extent that survey outcomes are associated with these demographic characteristics, matching the national distributions will reduce bias in estimates of these outcomes at the national level.

The China Mental Health Survey: II. Design and field procedures

China Mental Health Survey (CMHS), which was carried out from July 2013 to March 2015, was the first national representative community survey of mental disorders and mental health services in China using computer-assisted personal interview (CAPI). Face-to-face interviews were finished in the homes of respondents who were selected from a nationally representative multi-stage disproportionate stratified sampling procedure. Sample selection was integrated with the National Chronic Disease and Risk Factor Surveillance Survey administered by the National Centre for Chronic and Non-communicable Disease Control and Prevention in 2013, which made it possible to obtain both physical and mental health information of Chinese community population. One-stage design of data collection was used in the CMHS to obtain the information of mental disorders, including mood disorders, anxiety disorders, and substance use disorders, while two-stage design was applied for schizophrenia and other psychotic disorders, and dementia. A total of 28,140 respondents finished the survey with 72.9% of the overall response rate. This paper describes the survey mode, fieldwork organization, procedures, and the sample design and weighting of the CMHS. Detailed information is presented on the establishment of a new payment scheme for interviewers, results of the quality control in both stages, and evaluations to the weighting.

Calibrated propensity score method for survey nonresponse in cluster sampling

Weighting adjustment is commonly used in survey sampling to correct for unit nonresponse. In cluster sampling, the missingness indicators are often correlated within clusters and the response mechanism is subject to cluster-specific nonignorable missingness. Based on a parametric working model for the response mechanism that incorporates cluster-specific nonignorable missingness, we propose a method of weighting adjustment. We provide a consistent estimator of the mean or totals in cases where the study variable follows a generalized linear mixed-effects model. The proposed method is robust in the sense that the consistency of the estimator does not require correct specification of the functional forms of the response and outcome models. A consistent variance estimator based on Taylor linearization is also proposed. Numerical results, including a simulation and a real-data application, are presented.

Space-Time Smoothing of Complex Survey Data: Small Area Estimation for Child Mortality

Many people living in low and middle-income countries are not covered by civil registration and vital statistics systems. Consequently, a wide variety of other types of data including many household sample surveys are used to estimate health and population indicators. In this paper we combine data from sample surveys and demographic surveillance systems to produce small area estimates of child mortality through time. Small area estimates are necessary to understand geographical heterogeneity in health indicators when full-coverage vital statistics are not available. For this endeavor spatio-temporal smoothing is beneficial to alleviate problems of data sparsity. The use of conventional hierarchical models requires careful thought since the survey weights may need to be considered to alleviate bias due to non-random sampling and non-response. The application that motivated this work is estimation of child mortality rates in five-year time intervals in regions of Tanzania. Data come from Demographic and Health Surveys conducted over the period 1991-2010 and two demographic surveillance system sites. We derive a variance estimator of under five years child mortality that accounts for the complex survey weighting. For our application, the hierarchical models we consider include random effects for area, time and survey and we compare models using a variety of measures including the conditional predictive ordinate (CPO). The method we propose is implemented via the fast and accurate integrated nested Laplace approximation (INLA).

Z-estimation and stratified samples: application to survival models

The infinite dimensional Z-estimation theorem offers a systematic approach to joint estimation of both Euclidean and non-Euclidean parameters in probability models for data. It is easily adapted for stratified sampling designs. This is important in applications to censored survival data because the inverse probability weights that modify the standard estimating equations often depend on the entire follow-up history. Since the weights are not predictable, they complicate the usual theory based on martingales. This paper considers joint estimation of regression coefficients and baseline hazard functions in the Cox proportional and Lin-Ying additive hazards models. Weighted likelihood equations are used for the former and weighted estimating equations for the latter. Regression coefficients and baseline hazards may be combined to estimate individual survival probabilities. Efficiency is improved by calibrating or estimating the weights using information available for all subjects. Although inefficient in comparison with likelihood inference for incomplete data, which is often difficult to implement, the approach provides consistent estimates of desired population parameters even under model misspecification.

A log-linear modelling approach to assessing the consistency of ego reports of dyadic outcomes with applications to fertility and sexual partnerships

We propose a log-linear model to assess the consistency of ego reports of dyadic outcomes. We do so specifically in the context where males and females report on shared events, and we demonstrate how inconsistencies can be assessed by using a log-linear model that estimates separate mixing totals for each set of reports. This modelling approach immediately allows us to determine where inconsistencies in reports occur. To demonstrate how our method can be easily implemented for survey data, we apply it to both the 1992 National Health and Social Life Survey and the 2002 National Survey of Family Growth. Our analysis identifies inconsistencies in male and female reports of concurrent partnerships and the number of biological children.

Smoking ban policies and their influence on smoking behaviors among current California smokers: a population-based study

OBJECTIVE

To assess whether smoking ban policies are associated with smoking reduction and quit attempts among California smokers.

METHODS

Data were examined for 1718 current smokers from follow-up telephone interviews conducted in 2011 of persons previously identified as smokers in a representative sample of the adult population of California. Population weighted logistic regressions controlling for demographic and other variables were used to evaluate the association between smoking ban policies (home, work, and town) and changes in tobacco use (past year quit attempt or reduction in smoking rate).

RESULTS

Living in a home with a total ban was significantly associated with smoking reduction (adjusted odds ratio, AOR: 2.4, 95% CI: 1.4-4.2) and making a quit attempt (AOR: 2.3, 95% CI: 1.3-3.9) compared to living in a home with no home ban. Self-reported perception of an outdoor ban in one's city/town was associated with smoking reduction (AOR: 1.7, 95% CI: 1.02-2.7) and making a quit attempt (AOR: 1.8, 95% CI: 1.05-2.9).

CONCLUSION

These results indicate that smoking bans not only protect nonsmokers from the harms of secondhand smoke, but are also associated with smoking reduction and cessation.

Approaches to retrospective sampling for longitudinal transition regression models

For binary diseases that relapse and remit, it is often of interest to estimate the effect of covariates on the transition process between disease states over time. The transition process can be characterized by modeling the probability of the binary event given the individual's history. Designing studies that examine the impact of time varying covariates over time can lead to collection of extensive amounts of data. Sometimes it may be possible to collect and store tissue, blood or images and retrospectively analyze this covariate information. In this paper we consider efficient sampling designs that do not require biomarker measurements on all subjects. We describe appropriate estimation methods for transition probabilities and functions of these probabilities, and evaluate efficiency of the estimates from the proposed sampling designs. These new methods are illustrated with data from a longitudinal study of bacterial vaginosis, a common relapsing-remitting vaginal infection of women of child bearing age.

Empirical Likelihood-Based Estimation of the Treatment Effect in a Pretest-Posttest Study

The pretest-posttest study design is commonly used in medical and social science research to assess the effect of a treatment or an intervention. Recently, interest has been rising in developing inference procedures that improve efficiency while relaxing assumptions used in the pretest-posttest data analysis, especially when the posttest measurement might be missing. In this article we propose a semiparametric estimation procedure based on empirical likelihood (EL) that incorporates the common baseline covariate information to improve efficiency. The proposed method also yields an asymptotically unbiased estimate of the response distribution. Thus functions of the response distribution, such as the median, can be estimated straightforwardly, and the EL method can provide a more appealing estimate of the treatment effect for skewed data. We show that, compared with existing methods, the proposed EL estimator has appealing theoretical properties, especially when the working model for the underlying relationship between the pretest and posttest measurements is misspecified. A series of simulation studies demonstrates that the EL-based estimator outperforms its competitors when the working model is misspecified and the data are missing at random. We illustrate the methods by analyzing data from an AIDS clinical trial (ACTG 175).