Document Detail


Accounting for misclassification error in retrospective smoking data.
MedLine Citation:
PMID:  15386689     Owner:  NLM     Status:  MEDLINE    
Abstract/OtherAbstract:
Recent waves of major longitudinal surveys in the US and other countries include retrospective questions about the timing of smoking initiation and cessation, creating a potentially important but under-utilized source of information on smoking behavior over the life course. In this paper, we explore the extent of, consequences of, and possible solutions to misclassification errors in models of smoking participation that use data generated from retrospective reports. In our empirical work, we exploit the fact that the National Longitudinal Survey of Youth 1979 provides both contemporaneous and retrospective information about smoking status in certain years. We compare the results from four sets of models of smoking participation. The first set of results are from baseline probit models of smoking participation from contemporaneously reported information. The second set of results are from models that are identical except that the dependent variable is based on retrospective information. The last two sets of results are from models that take a parametric approach to account for a simple form of misclassification error. Our preliminary results suggest that accounting for misclassification error is important. However, the adjusted maximum likelihood estimation approach to account for misclassification does not always perform as expected.
Authors:
Donald S Kenkel; Dean R Lillard; Alan D Mathios
Related Documents :
21085449 - Determination of line strengths by fourier-transform spectroscopy.
16895819 - Automated medication error studies with audit supplementation were effectively designed...
11453269 - Accuracy and modifying factors of the sonographic estimation of fetal weight in a high-...
18252649 - On the regularization of forgetting recursive least square.
16441959 - Procedure to normalize data for benchmarking.
3427179 - Statistical diagnosis and the influence of diagnostic error.
2692999 - Animal studies of surfactant replacement therapy.
21421429 - Recognizing architectural distortion in mammogram: a multi-scale texture modeling appro...
10966759 - The implications of variable remigration intervals for the assessment of population siz...
Publication Detail:
Type:  Journal Article; Research Support, U.S. Gov't, P.H.S.    
Journal Detail:
Title:  Health economics     Volume:  13     ISSN:  1057-9230     ISO Abbreviation:  Health Econ     Publication Date:  2004 Oct 
Date Detail:
Created Date:  2004-10-06     Completed Date:  2005-01-27     Revised Date:  2007-11-14    
Medline Journal Info:
Nlm Unique ID:  9306780     Medline TA:  Health Econ     Country:  England    
Other Details:
Languages:  eng     Pagination:  1031-44     Citation Subset:  IM    
Copyright Information:
Copyright 2004 John Wiley & Sons, Ltd.
Affiliation:
Department of Policy Analysis and Management, Cornell University, Ithaca, NY 14853-4401, USA. dsk10@cornell.edu
Export Citation:
APA/MLA Format     Download EndNote     Download BibTex
MeSH Terms
Descriptor/Qualifier:
Data Interpretation, Statistical*
Humans
Longitudinal Studies
Models, Econometric
Retrospective Studies
Smoking / epidemiology*
United States
Grant Support
ID/Acronym/Agency:
R01 CA094020-01/CA/NCI NIH HHS

From MEDLINE®/PubMed®, a database of the U.S. National Library of Medicine


Previous Document:  The sources of hospital cost variability.
Next Document:  The effect of the tobacco settlement and smoking bans on alcohol consumption.