Overdispersion in the poisson regression model: A comparative simulation study

Alfonso Palmer; J. M. Losilla; J. Vives; R. Jiménez

doi:https://doi.org/10.1027/1614-2241.3.3.89

Overdispersion in the poisson regression model: A comparative simulation study

Alfonso Palmer, J. M. Losilla, J. Vives, R. Jiménez

Department of Psychobiology and Methodology of Health Sciences

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

14 Citations (Scopus)

Abstract

This simulation study compares different strategies to solve the problem of underestimating standard errors in the Poisson regression model when overdispersion is present. The study analyses the importance of sample size, Poisson distribution mean, and dispersion parameter in choosing the best index or estimate. Results show that standard error (SE) estimates obtained by resampling (nonparametric bootstrap and jackknife) are the least biased, followed by the direct index based on the χ 2, and the so-called robust indexes, in third place. Nevertheless, the inefficiency of resampling estimates is also evident, especially in small samples. © 2007 Hogrefe & Huber Publishers.

Original language	English
Pages (from-to)	89-99
Journal	Methodology
Volume	3
DOIs	https://doi.org/10.1027/1614-2241.3.3.89
Publication status	Published - 1 Jan 2007

Keywords

Direct indexes
Jackknife
Nonparametric bootstrap
Overdispersion
Robust indexes

Access to Document

https://doi.org/10.1027/1614-2241.3.3.89

Fingerprint Dive into the research topics of 'Overdispersion in the poisson regression model: A comparative simulation study'. Together they form a unique fingerprint.

Cite this

@article{1ca9217ac9c44ae5b72fcfc5e4d4e77b,

title = "Overdispersion in the poisson regression model: A comparative simulation study",

abstract = "This simulation study compares different strategies to solve the problem of underestimating standard errors in the Poisson regression model when overdispersion is present. The study analyses the importance of sample size, Poisson distribution mean, and dispersion parameter in choosing the best index or estimate. Results show that standard error (SE) estimates obtained by resampling (nonparametric bootstrap and jackknife) are the least biased, followed by the direct index based on the χ 2, and the so-called robust indexes, in third place. Nevertheless, the inefficiency of resampling estimates is also evident, especially in small samples. {\textcopyright} 2007 Hogrefe & Huber Publishers.",

keywords = "Direct indexes, Jackknife, Nonparametric bootstrap, Overdispersion, Robust indexes",

author = "Alfonso Palmer and Losilla, {J. M.} and J. Vives and R. Jim{\'e}nez",

year = "2007",

month = jan,

day = "1",

doi = "https://doi.org/10.1027/1614-2241.3.3.89",

language = "English",

volume = "3",

pages = "89--99",

}

TY - JOUR

T1 - Overdispersion in the poisson regression model: A comparative simulation study

AU - Palmer, Alfonso

AU - Losilla, J. M.

AU - Vives, J.

AU - Jiménez, R.

PY - 2007/1/1

Y1 - 2007/1/1

N2 - This simulation study compares different strategies to solve the problem of underestimating standard errors in the Poisson regression model when overdispersion is present. The study analyses the importance of sample size, Poisson distribution mean, and dispersion parameter in choosing the best index or estimate. Results show that standard error (SE) estimates obtained by resampling (nonparametric bootstrap and jackknife) are the least biased, followed by the direct index based on the χ 2, and the so-called robust indexes, in third place. Nevertheless, the inefficiency of resampling estimates is also evident, especially in small samples. © 2007 Hogrefe & Huber Publishers.

AB - This simulation study compares different strategies to solve the problem of underestimating standard errors in the Poisson regression model when overdispersion is present. The study analyses the importance of sample size, Poisson distribution mean, and dispersion parameter in choosing the best index or estimate. Results show that standard error (SE) estimates obtained by resampling (nonparametric bootstrap and jackknife) are the least biased, followed by the direct index based on the χ 2, and the so-called robust indexes, in third place. Nevertheless, the inefficiency of resampling estimates is also evident, especially in small samples. © 2007 Hogrefe & Huber Publishers.

KW - Direct indexes

KW - Jackknife

KW - Nonparametric bootstrap

KW - Overdispersion

KW - Robust indexes

U2 - https://doi.org/10.1027/1614-2241.3.3.89

DO - https://doi.org/10.1027/1614-2241.3.3.89

M3 - Article

VL - 3

SP - 89

EP - 99

ER -