A bridging model to reconcile statistics based on data from multiple surveys

Erciulescu, Andreea L.; Opsomer, Jean D.; Breidt, F. Jay

doi:10.1214/20-aoas1437

Cited by 4 publications

(12 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this article, we consider two surveys and assume that the corresponding pairs of estimates for a common variable are unbiased but subject to different, independent sampling errors. This assumption is present in the Raghunathan et al (2007) and Erciulescu, Opsomer & Breidt (2020) studies. Unlike the work of Erciulescu, Opsomer & Breidt (2020), we assume that the pairs of estimates are measured on the same scale (no bridging level is needed), and at the same level of aggregation (no aggregation is needed), but are available for different sets of domains (there are domains with data available from only one of the two surveys).…”

Section: Introductionmentioning

confidence: 89%

“…This assumption is present in the Raghunathan et al (2007) and Erciulescu, Opsomer & Breidt (2020) studies. Unlike the work of Erciulescu, Opsomer & Breidt (2020), we assume that the pairs of estimates are measured on the same scale (no bridging level is needed), and at the same level of aggregation (no aggregation is needed), but are available for different sets of domains (there are domains with data available from only one of the two surveys). In addition, unlike the studies mentioned above, we combine information available on more than one quantity of interest.…”

Section: Introductionmentioning

confidence: 89%

“…The authors reconciled state‐level yield forecasts available from two surveys at multiple time points within the crop season, based on a third survey collecting yield information at the end of the crop season. Recently, Erciulescu, Opsomer & Breidt (2021) proposed a hierarchical Bayes model for reconciling recreational activity estimates available from two surveys fielded concurrently. Similar to the application in Lohr & Brick (2012), one survey, the National Survey of Fishing, Hunting, and Wildlife‐Associated Recreation, was not sufficient to provide reliable estimates at granular levels such as states.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Statistical data integration using multilevel models to predict employee compensation

2022

View full text Add to dashboard Cite

This article considers the case where two surveys collect data on a common variable, with one survey being much smaller than the other. The smaller survey collects data on an additional variable of interest, related to the common variable collected in the two surveys, and out‐of‐scope with respect to the larger survey. Estimation of the two related variables is of interest at domains defined at a granular level. We propose a multilevel model for integrating data from the two surveys, by reconciling survey estimates available for the common variable, accounting for the relationship between the two variables, and expanding estimation for the other variable, for all the domains of interest. The model is specified as a hierarchical Bayes model for domain‐level survey data, and posterior distributions are constructed for the two variables of interest. A synthetic estimation approach is considered as an alternative to the hierarchical modelling approach. The methodology is applied to wage and benefits estimation using data from the National Compensation Survey and the Occupational Employment Statistics Survey, available from the Bureau of Labor Statistics, Department of Labor, United States.

show abstract

Section: Introductionmentioning

confidence: 89%

Section: Introductionmentioning

confidence: 89%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Statistical data integration using multilevel models to predict employee compensation

2022

View full text Add to dashboard Cite

show abstract

“…The bridging problem and the estimation at finer levels problem have been recently considered in [5]. In this study, data from one survey are available at the census division level, while data from another survey are available at the state level.…”

Section: Introductionmentioning

confidence: 99%

“…Similarly, we develop a model that integrates the health insurance coverage estimates at the state level and produces estimates at the county level on the scales of both BRFSS and SAHIE. Unlike the scenario in [5], only one of the sources is a survey, the other being the result of modelbased estimation and estimates from both sources are publicly available as health official statistics at distinct levels of aggregation.…”

Section: Introductionmentioning

confidence: 99%

Statistical data integration models to bridge health official statistics

Erciulescu¹

2022

SJI

Self Cite

View full text Add to dashboard Cite

There are at least two sources of health insurance coverage estimates in the United States: the Behavioral Risk Factor Surveillance System (BRFSS) and the Small Area Health Insurance Estimates (SAHIE) program. This paper addresses the integration of BRFSS and SAHIE data using multilevel models that account for the different levels of aggregation at which data are available and for the different errors to which data are subject to. The uncertainty in the initial state-level estimates available from BRFSS and SAHIE is improved by borrowing information from both sources and across geographies. County-level model estimates are produced on both BRFSS and SAHIE scales, improving the usability of the BRFSS public-use data and inspiring possible extensions to estimation of BRFSS quantities other than health insurance coverage. The application uses 2018 public-use data. Parallels to small area estimation models and measurement error models are briefly discussed.

show abstract

Combining surveys in small area estimation using area‐level models

Franco

Maitra

2023

WIREs Computational Stats

View full text Add to dashboard Cite

For many sample surveys, researchers, policymakers, and other stakeholders are interested in obtaining estimates for various domains, such as for geographic levels, for demographic groups, or a cross‐classification of both. Often, the demand for estimates at a disaggregated level exceeds what the sample size and survey design can support when estimation is done by traditional design‐based estimation methods. Small area estimation involves exploiting relationships among domains and borrowing strength from multiple sources of information to improve inference relative to direct survey methods. This typically involves the use of models whose success depends heavily on the quality and predictive ability of the sources of information used. Possible sources of auxiliary information include administrative records, Censuses, big data such as traffic or cell phone data, or previous vintages of the same survey. One rich source of information is that of other surveys, especially in countries like the United States, where multiple surveys exist that cover related topics. We will provide an introduction to the topic of combining information from multiple surveys in small area estimation using area‐level models, including practical advice and a technical introduction, and illustrating with applications. We will discuss reasons to combine surveys and give an overview of some of the most common types of models.This article is categorized under: Statistical Models > Multivariate Models

show abstract

A bridging model to reconcile statistics based on data from multiple surveys

Cited by 4 publications

References 8 publications

Statistical data integration using multilevel models to predict employee compensation

Statistical data integration using multilevel models to predict employee compensation

Statistical data integration models to bridge health official statistics

Combining surveys in small area estimation using area‐level models

Contact Info

Product

Resources

About