Grant: Discovery Grants Program - Individual
Memorial University of Newfoundland
$21,000.00
Modeling measurement error problems using generalized quasi-likelihood method
FY 2012-13
Summary
In statistical modeling of data with dependent responses, measurement errors in covariates are often neglected, especially for data collected in socioeconomic studies. This causes estimation bias of the model parameters and usually leads to unsatisfactory efficiency of the inference. The correction of this estimation bias has recently attracted the attention of statisticians.The generalized method of moments (GMM) has been widely used to dependent data by economists to obtain consistent and efficient estimates of model parameters. It has be also demonstrated that the generalized quasilikelihood approach, which exploits the mean vector and the covariance structure of the repeated data, produces consistent and more efficient estimates compared with the GMM and other generalized estimating equations (GEE) methods.
The preliminary simulation results based on GMM and GQL approaches that we have obtained from our investigation of bias correction in linear panel data models where possibly all the covariates are subject to measurement error are very promising. In this research program we will investigate bias correction of parameter estimation and their asymptotic properties based on the GMM and GQL approaches for different kinds of dependent data. We will also conduct Monte Carlo simulation to investigate their small sample properties.
The outcome of our proposed research will provide a more efficient and easy to manipulate method to handle dependent data with covariates subject to measurement errors. It can help practitioners to better understand their underlying research in different fields, such as biomedical studies and socio-economic studies.
Program
Discovery Grants Program - Individual
Recipient
Location
Newfoundland and Labrador, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
326970-2009
Name nom
Fan, Zhaozhi
Department département
Mathematics and Statistics
Organization id
89
Institution établissement
Memorial University of Newfoundland
Province en
Newfoundland and Labrador
Province fr
Terre-Neuve-et-Labrador
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2012
Competition year année de concours
2009
Award amount
21000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
Discovery grants programs
Group fr
Programmes de subventions à la découverte
Committee code
14
Committee name en
Statistical Sciences
Committee name fr
Statistique
Area of application code
1205
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Mathematical sciences
Area of application fr
Sciences mathématiques
Research subject code
3004
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Applied statistics
Research subject fr
Statistique appliquée
Partie
4
Nb partie
5
Application title
Modeling measurement error problems using generalized quasi-likelihood method
Keyword
NSERC
Application summary
In statistical modeling of data with dependent responses, measurement errors in covariates are often neglected, especially for data collected in socioeconomic studies. This causes estimation bias of the model parameters and usually leads to unsatisfactory efficiency of the inference. The correction of this estimation bias has recently attracted the attention of statisticians.The generalized method of moments (GMM) has been widely used to dependent data by economists to obtain consistent and efficient estimates of model parameters. It has be also demonstrated that the generalized quasilikelihood approach, which exploits the mean vector and the covariance structure of the repeated data, produces consistent and more efficient estimates compared with the GMM and other generalized estimating equations (GEE) methods.
The preliminary simulation results based on GMM and GQL approaches that we have obtained from our investigation of bias correction in linear panel data models where possibly all the covariates are subject to measurement error are very promising. In this research program we will investigate bias correction of parameter estimation and their asymptotic properties based on the GMM and GQL approaches for different kinds of dependent data. We will also conduct Monte Carlo simulation to investigate their small sample properties.
The outcome of our proposed research will provide a more efficient and easy to manipulate method to handle dependent data with covariates subject to measurement errors. It can help practitioners to better understand their underlying research in different fields, such as biomedical studies and socio-economic studies.