Nonparametric Bayes modeling with sample survey weights

Kunihama, Tsuyoshi; Herring, Amy H.; Halpern, Carolyn Tucker; Dunson, David B.

doi:10.1016/j.spl.2016.02.009

Cited by 15 publications

(24 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, as discussed by Kunihama et al . (), the vast majority of such methods are not appropriate in model‐based inferences, particularly under Bayesian frameworks, and those methods that have been proposed for Bayesian analysis require highly complex models. As identified by Gelman (), Kunihama et al .…”

Section: Discussionmentioning

confidence: 99%

“…As identified by Gelman (), Kunihama et al . () and Kang and Bernstein (), this has led to a disconnect between the analysis of survey data in practice and methods that have been developed to account for survey sample bias in Bayesian inference.…”

Section: Discussionmentioning

confidence: 99%

“…There is a rich literature on including survey weight adjustments in estimation (see Rao (2011) for a detailed review). However, as discussed by Kunihama et al (2016), the vast majority of such methods are not appropriate in model-based inferences, particularly under Bayesian frameworks, and those methods that have been proposed for Bayesian analysis require highly complex models. As identified by Gelman (2007), Kunihama et al (2016) and Kang and Bernstein (2016), this has led to a disconnect between the analysis of survey data in practice and methods that have been developed to account for survey sample bias in Bayesian inference.…”

Section: Discussionmentioning

confidence: 99%

See 2 more Smart Citations

‘What Drives Commuter Behaviour?’: A Bayesian Clustering Approach for Understanding Opposing Behaviours in Social Surveys

Dawkins

Williamson

Barr

et al. 2019

Journal of the Royal Statistical Society Series A: Statistics in Society

View full text Add to dashboard Cite

Summary The city of Exeter, UK, is experiencing unprecedented growth, putting pressure on traffic infrastructure. As well as traffic network management, understanding and influencing commuter behaviour is important for reducing congestion. Information about current commuter behaviour has been gathered through a large on‐line survey, and similar individuals have been grouped to explore distinct behaviour profiles to inform intervention design to reduce commuter congestion. Statistical analysis within societal applications benefit from incorporating available social scientist expert knowledge. Current clustering approaches for the analysis of social surveys assume that the number of groups and the within‐group narratives are unknown a priori. Here, however, informed by valuable expert knowledge, we develop a novel Bayesian approach for creating a clear opposing transport mode group narrative within survey respondents, simplifying communication with project partners and the general public. Our methodology establishes groups characterizing opposing behaviours based on a key multinomial survey question by constraining parts of our prior judgement within a Bayesian finite mixture model. Drivers of group membership and within‐group behavioural differences are modelled hierarchically by using further information from the survey. In applying the methodology we demonstrate how it can be used to understand the key drivers of opposing behaviours in any wider application.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

‘What Drives Commuter Behaviour?’: A Bayesian Clustering Approach for Understanding Opposing Behaviours in Social Surveys

Dawkins

Williamson

Barr

et al. 2019

Journal of the Royal Statistical Society Series A: Statistics in Society

View full text Add to dashboard Cite

show abstract

“…For the finite approximation of the Dirichlet process mixtures with H components, Kunihama et al . () developed an adjustment method, which modifies the mixture weights { π h } in equation by using the survey weights. In the posterior computation, to obtain distributions for a target population, we generate adjusted mixture weights

\overset{false~}{π} = {({\overset{π}{true}}_{1}, \dots, {\overset{π}{true}}_{H})}^{'}

from

\overset{false~}{π} \sim Dirichlet (a_{1} + \sum_{i : s_{i} = 1} \frac{w_{i}}{c}, \dots, a_{H} + \sum_{i : s_{i} = H} \frac{w_{i}}{c}),

where

c = {normalΣ}_{i = 1}^{n} w_{i} / N

and s i is the cluster index variable for the i th respondent.…”

Section: Proposed Modelling Of Mixed Scale Longitudinal Surveysmentioning

confidence: 99%

“…In our study, we follow the default setting in Kunihama et al . () with prior sample size 1–2% of the population size N .…”

Section: Proposed Modelling Of Mixed Scale Longitudinal Surveysmentioning

confidence: 99%

Non-parametric Bayes Models for Mixed Scale Longitudinal Surveys

Kunihama

Halpern

Herring

2019

Journal of the Royal Statistical Society Series C: Applied Statistics

View full text Add to dashboard Cite

Summary Modelling and computation for multivariate longitudinal surveys have proven challenging, particularly when data are not all continuous and Gaussian but contain discrete measurements. In many social science surveys, study participants are selected via complex survey designs such as stratified random sampling, leading to discrepancies between the sample and population, which are further compounded by missing data and loss to follow‐up. Survey weights are typically constructed to address these issues, but it is not clear how to include them in models. Motivated by data on sexual development, we propose a novel non‐parametric approach for mixed scale longitudinal data in surveys. In the approach proposed, the mixed scale multivariate response is expressed through an underlying continuous variable with dynamic latent factors inducing time varying associations. Bias from the survey design is adjusted for in posterior computation relying on a Markov chain Monte Carlo algorithm. The approach is assessed in simulation studies and applied to the National Longitudinal Study of Adolescent to Adult Health.

show abstract

Bayesian estimation methods for survey data with potential applications to health disparities research

Wu,

Stephenson

2023

WIREs Computational Stats

View full text Add to dashboard Cite

Understanding how and why certain communities bear a disproportionate burden of disease is challenging due to the scarcity of data on these communities. Surveys provide a useful avenue for accessing hard‐to‐reach populations, as many surveys specifically oversample understudied and vulnerable populations. When survey data is used for analysis, it is important to account for the complex survey design that gave rise to the data, to avoid biased conclusions. The field of Bayesian survey statistics aims to account for such survey design while leveraging the advantages of Bayesian models, which can flexibly handle sparsity through borrowing of information and provide a coherent inferential framework to easily obtain variances for complex models and data types. For these reasons, Bayesian survey methods seem uniquely well‐poised for health disparities research, where heterogeneity and sparsity are frequent considerations. This review discusses three main approaches found in the Bayesian survey methodology literature: (1) multilevel regression and post‐stratification, (2) weighted pseudolikelihood‐based methods, and (3) synthetic population generation. We discuss the advantages and disadvantages of each approach, examine recent applications and extensions, and consider how these approaches may be leveraged to improve research in population health equity.This article is categorized under: Applications of Computational Statistics > Clinical Trials Statistical and Graphical Methods of Data Analysis > Bayesian Methods and Theory Applications of Computational Statistics > Health and Medical Data/Informatics Statistical and Graphical Methods of Data Analysis > Sampling

show abstract

Nonparametric Bayes modeling with sample survey weights

Cited by 15 publications

References 12 publications

‘What Drives Commuter Behaviour?’: A Bayesian Clustering Approach for Understanding Opposing Behaviours in Social Surveys

‘What Drives Commuter Behaviour?’: A Bayesian Clustering Approach for Understanding Opposing Behaviours in Social Surveys

Non-parametric Bayes Models for Mixed Scale Longitudinal Surveys

Bayesian estimation methods for survey data with potential applications to health disparities research

Contact Info

Product

Resources

About