A Bayesian approach to account for misclassification in prevalence and trend estimation

Martijn van Hasselt; Christopher R. Bollinger; Jeremy W. Bray

doi:10.1002/jae.2879

A Bayesian approach to account for misclassification in prevalence and trend estimation

Martijn van Hasselt, Christopher R. Bollinger, Jeremy W. Bray

Economics

Research output: Contribution to journal › Article › peer-review

1 Scopus citations

Abstract

In this paper, we present a Bayesian approach to estimate the mean of a binary variable and changes in the mean over time, when the variable is subject to misclassification error. These parameters are partially identified, and we derive identified sets under various assumptions about the misclassification rates. We apply our method to estimating the prevalence and trend of prescription opioid misuse, using data from the 2002–2014 National Survey on Drug Use and Health. Using a range of priors, the posterior distribution provides evidence that among middle-aged White men, the prevalence of opioid misuse increased multiple times between 2002 and 2012.

Original language	English
Pages (from-to)	351-367
Number of pages	17
Journal	Journal of Applied Econometrics
Volume	37
Issue number	2
DOIs	https://doi.org/10.1002/jae.2879
State	Published - Mar 2022

Bibliographical note

Publisher Copyright:
© 2021 John Wiley & Sons, Ltd.

Funding

We thank John Pepper and Gary Koop for helpful comments on an earlier version of this paper.

Funders	Funder number
John Pepper and Gary Koop

Keywords

Bayesian estimation
Misclassication
partial identication

ASJC Scopus subject areas

Social Sciences (miscellaneous)
Economics and Econometrics

Access to Document

10.1002/jae.2879

Cite this

@article{5f2edcfc361b4f10bbb134af2fdaf6ab,

title = "A Bayesian approach to account for misclassification in prevalence and trend estimation",

abstract = "In this paper, we present a Bayesian approach to estimate the mean of a binary variable and changes in the mean over time, when the variable is subject to misclassification error. These parameters are partially identified, and we derive identified sets under various assumptions about the misclassification rates. We apply our method to estimating the prevalence and trend of prescription opioid misuse, using data from the 2002–2014 National Survey on Drug Use and Health. Using a range of priors, the posterior distribution provides evidence that among middle-aged White men, the prevalence of opioid misuse increased multiple times between 2002 and 2012.",

keywords = "Bayesian estimation, Misclassication, partial identication",

author = "{van Hasselt}, Martijn and Bollinger, {Christopher R.} and Bray, {Jeremy W.}",

note = "Publisher Copyright: {\textcopyright} 2021 John Wiley & Sons, Ltd.",

year = "2022",

month = mar,

doi = "10.1002/jae.2879",

language = "English",

volume = "37",

pages = "351--367",

number = "2",

}

TY - JOUR

T1 - A Bayesian approach to account for misclassification in prevalence and trend estimation

AU - van Hasselt, Martijn

AU - Bollinger, Christopher R.

AU - Bray, Jeremy W.

PY - 2022/3

Y1 - 2022/3

N2 - In this paper, we present a Bayesian approach to estimate the mean of a binary variable and changes in the mean over time, when the variable is subject to misclassification error. These parameters are partially identified, and we derive identified sets under various assumptions about the misclassification rates. We apply our method to estimating the prevalence and trend of prescription opioid misuse, using data from the 2002–2014 National Survey on Drug Use and Health. Using a range of priors, the posterior distribution provides evidence that among middle-aged White men, the prevalence of opioid misuse increased multiple times between 2002 and 2012.

AB - In this paper, we present a Bayesian approach to estimate the mean of a binary variable and changes in the mean over time, when the variable is subject to misclassification error. These parameters are partially identified, and we derive identified sets under various assumptions about the misclassification rates. We apply our method to estimating the prevalence and trend of prescription opioid misuse, using data from the 2002–2014 National Survey on Drug Use and Health. Using a range of priors, the posterior distribution provides evidence that among middle-aged White men, the prevalence of opioid misuse increased multiple times between 2002 and 2012.

KW - Bayesian estimation

KW - Misclassication

KW - partial identication

UR - http://www.scopus.com/inward/record.url?scp=85115807849&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85115807849&partnerID=8YFLogxK

U2 - 10.1002/jae.2879

DO - 10.1002/jae.2879

M3 - Article

AN - SCOPUS:85115807849

SN - 0883-7252

VL - 37

SP - 351

EP - 367

JO - Journal of Applied Econometrics

JF - Journal of Applied Econometrics

IS - 2

ER -

A Bayesian approach to account for misclassification in prevalence and trend estimation

Abstract

Bibliographical note

Funding

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this