Generalized Linear Models for Small-Area Estimation

Ghosh, Malay; Natarajan, Kannan; Stroud, T. W. F.; Carlin, Bradley P.

doi:10.2307/2669623

Cited by 88 publications

(81 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For example, generalized linear models with binomial or Poisson errors, as suggested in Ghosh et al (1998) andFarrell (2000), could be used but would complicate the modeling (e.g., in incorporating complex design features) and computational tasks greatly. For example, use of the logistic-normal model would require rejection sampling techniques such as the Metropolis-Hastings algorithm within the Gibbs sampler for over 37,000 parameters, and the resulting increased computing time needed would likely render the estimation procedure infeasible.…”

Section: Use Of the Arcsine-square Root Transformationmentioning

confidence: 99%

Combining Information From Two Surveys to Estimate County-Level Prevalence Rates of Cancer Risk Factors and Screening

Raghunathan

Xie

Schenker

et al. 2007

Journal of the American Statistical Association

109

View full text Add to dashboard Cite

Cancer surveillance requires estimates of the prevalence of cancer risk factors and screening for small areas such as counties. Two popular data sources are the Behavioral Risk Factor Surveillance System (BRFSS), a telephone survey conducted by state agencies, and the National Health Interview Survey (NHIS), an area probability sample survey conducted through face-to-face interviews. Both data sources have advantages and disadvantages. The BRFSS is a larger survey, and almost every county is included in the survey; but it has lower response rates as is typical with telephone surveys, and it does not include subjects who live in households with no telephones. On the other hand, the NHIS is a smaller survey, with the majority of counties not included; but it includes both telephone and non-telephone households and has higher response rates. A preliminary analysis shows that the distributions of cancer screening and risk factors are different for telephone and non-telephone households. Thus, information from the two surveys may be combined to address both nonresponse and noncoverage errors. A hierarchical Bayesian approach that combines information from both surveys is used to construct county-level estimates. The proposed model incorporates potential noncoverage and nonresponse biases in the BRFSS as well as complex sample design features of both surveys. A Markov Chain Monte Carlo method is used to simulate draws from the joint posterior distribution of unknown quantities in the model based on the design-based direct estimates and county-level covariates. Yearly prevalence estimates at the county level for 49 states, as well as for the entire state of Alaska and the District of Columbia, are developed for six outcomes using BRFSS and NHIS data from the years 1997-2000. The outcomes include smoking and use of common cancer screening procedures. The NHIS/BRFSS combined county-level estimates are substantially different from those based on BRFSS alone. Combining Information from Two Surveys to Estimate AbstractCancer surveillance research requires estimates of the prevalence of cancer risk factors and screening for small areas such as counties. Two popular data sources are the Behavioral Risk Factor Surveillance System (BRFSS), a telephone survey conducted by state agencies, and the National Health Interview Survey (NHIS), an area probability sample survey conducted through face-to-face interviews. Both data sources have advantages and disadvantages. The BRFSS is a larger survey, and almost every county is included in the survey; but it has lower response rates as is typical with with telephone surveys, and it does not include subjects who live in households with no telephones.On the other hand, the NHIS is a smaller survey, with the majority of counties not included; but it includes both telephone and non-telephone households and has higher response rates. A preliminary analysis shows that the distributions of cancer screening and risk factors are different for telephone and non-telephone households. Thus,...

show abstract

Section: Use Of the Arcsine-square Root Transformationmentioning

confidence: 99%

Combining Information From Two Surveys to Estimate County-Level Prevalence Rates of Cancer Risk Factors and Screening

Raghunathan

Xie

Schenker

et al. 2007

Journal of the American Statistical Association

109

View full text Add to dashboard Cite

show abstract

“…Much of the literature focuses on continuous outcome measures such as income, but there is a growing literature where the outcome is a qualitative or discrete variable (see, for example, Ghosh et al, 1998). Though much of the literature uses Bayesian methods (see, for example, Ghosh and Rao,1994;Malec et al,1997), Elbers, Lanjouw, and Lanjouw (2003) suggest building a model of the outcome variable, estimating the model parameters using classical methods, and then using the estimated relationship to project locally.…”

Section: Literature Reviewmentioning

confidence: 99%

Estimating local prevalence of mental health problems

Stern

2014

Health Serv Outcomes Res Method

View full text Add to dashboard Cite

“…In the context of small area estimation, Ghosh et al (1998) and Sun et al (1999) show that despite the singularity of matrix Q such autoregressive priors lead to a proper posterior distribution under mild conditions on the model likelihood function. Our ''pseudolikelihood'' (7) does not satisfy these conditions when C s ¼ 0 for all s, or when C s ¼ T s for all s. Although very unlikely, such values of recombination counts do have strictly positive probability mass a posteriori.…”

Section: Synchronizing Recombinant and Parental Sequencesmentioning

confidence: 99%

Phylogenetic Mapping of Recombination Hotspots in Human Immunodeficiency Virus via Spatially Smoothed Change-Point Processes

Minin¹,

Dorman

Fang³

et al. 2007

Genetics

View full text Add to dashboard Cite

We present a Bayesian framework for inferring spatial preferences of recombination from multiple putative recombinant nucleotide sequences. Phylogenetic recombination detection has been an active area of research for the last 15 years. However, only recently attempts to summarize information from several instances of recombination have been made. We propose a hierarchical model that allows for simultaneous inference of recombination breakpoint locations and spatial variation in recombination frequency. The dual multiple change-point model for phylogenetic recombination detection resides at the lowest level of our hierarchy under the umbrella of a common prior on breakpoint locations. The hierarchical prior allows for information about spatial preferences of recombination to be shared among individual data sets. To overcome the sparseness of breakpoint data, dictated by the modest number of available recombinant sequences, we a priori impose a biologically relevant correlation structure on recombination location log odds via a Gaussian Markov random field hyperprior. To examine the capabilities of our model to recover spatial variation in recombination frequency, we simulate recombination from a predefined distribution of breakpoint locations. We then proceed with the analysis of 42 human immunodeficiency virus (HIV) intersubtype gag recombinants and identify a putative recombination hotspot.

show abstract

Generalized Linear Models for Small-Area Estimation

Cited by 88 publications

References 0 publications

Combining Information From Two Surveys to Estimate County-Level Prevalence Rates of Cancer Risk Factors and Screening

Combining Information From Two Surveys to Estimate County-Level Prevalence Rates of Cancer Risk Factors and Screening

Estimating local prevalence of mental health problems

Phylogenetic Mapping of Recombination Hotspots in Human Immunodeficiency Virus via Spatially Smoothed Change-Point Processes

Contact Info

Product

Resources

About