A robust property of pseudo-likelihood estimation for count data

It is a common occurrence that count data exhibit extra-Poisson variation. In such cases, quasi- likelihood estimation is often employed for the analysis of the covariate effect. Here, only the first two moments of the counts need specification. The estimator retains consistency when the variance is...

Full description

Saved in:
Bibliographic Details
Published in:Journal of statistical planning and inference 1993-06, Vol.35 (3), p.309-317
Main Author: Dean, C.B.
Format: Article
Language:English
Subjects:
estimating equation
Exact sciences and technology
extra-binomial variation
extra-Poisson variation
matched data
Mathematics
mixture model
moment methods
Overdispersion
Parametric inference
Probability and statistics
quasi-likelihood
Sciences and techniques of general use
Statistics
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is a common occurrence that count data exhibit extra-Poisson variation. In such cases, quasi- likelihood estimation is often employed for the analysis of the covariate effect. Here, only the first two moments of the counts need specification. The estimator retains consistency when the variance is misspecified, but, in general, the estimate of its variance does not. This paper explores situations where the asymptotic variance of the estimator of the covariate effect is correct under misspecification of the variance of the counts. A particular case where this holds is when the pseudo-likelihood estimating function is used for estimation in a matched design.
ISSN:0378-3758
1873-1171
DOI:10.1016/0378-3758(93)90019-3