Count data are commonly assumed to have a Poisson distribution, especially when there is no diagnostic procedure for checking this assumption. However, count data rarely fit the restrictive assumptions of the Poisson distribution. The violation of much of such assumptions commonly results in overdispersion, which invalidates the Poisson distribution. Undetected overdispersion may entail important misleading inferences, so its detection is essential. In this study, different overdispersion diagnostic tests are evaluated through two simulation studies. In Exp. 1, the nominal error rate is compared under different sample sizes and lamda conditions. Analysis shows a remarkable performance of the chi2 df test. In Exp. 2 and 3, statistical power is compared under different sample sizes, lamda, and overdispersion conditions. Chi2 and LR tests provide the highest statistical power.
%0 Journal Article
%1 vives_overdispersion_2008
%A Vives, Jaume
%A Losilla, Josep-Maria
%A Rodrigo, Maria-Florencia
%A Portell, Mariona
%D 2008
%J Psychological Reports
%K \& data humans, numerical psychology psychology, statistics
%N 1
%P 145--160
%R 10.2466/PR0.103.5.145-160
%T Overdispersion tests in count-data analysis.
%U http://www.ncbi.nlm.nih.gov/pubmed/18982948
%V 103
%X Count data are commonly assumed to have a Poisson distribution, especially when there is no diagnostic procedure for checking this assumption. However, count data rarely fit the restrictive assumptions of the Poisson distribution. The violation of much of such assumptions commonly results in overdispersion, which invalidates the Poisson distribution. Undetected overdispersion may entail important misleading inferences, so its detection is essential. In this study, different overdispersion diagnostic tests are evaluated through two simulation studies. In Exp. 1, the nominal error rate is compared under different sample sizes and lamda conditions. Analysis shows a remarkable performance of the chi2 df test. In Exp. 2 and 3, statistical power is compared under different sample sizes, lamda, and overdispersion conditions. Chi2 and LR tests provide the highest statistical power.
@article{vives_overdispersion_2008,
abstract = {Count data are commonly assumed to have a Poisson distribution, especially when there is no diagnostic procedure for checking this assumption. However, count data rarely fit the restrictive assumptions of the Poisson distribution. The violation of much of such assumptions commonly results in overdispersion, which invalidates the Poisson distribution. Undetected overdispersion may entail important misleading inferences, so its detection is essential. In this study, different overdispersion diagnostic tests are evaluated through two simulation studies. In Exp. 1, the nominal error rate is compared under different sample sizes and lamda conditions. Analysis shows a remarkable performance of the chi2 df test. In Exp. 2 and 3, statistical power is compared under different sample sizes, lamda, and overdispersion conditions. Chi2 and LR tests provide the highest statistical power.},
added-at = {2017-01-09T13:57:26.000+0100},
author = {Vives, Jaume and Losilla, Josep-Maria and Rodrigo, Maria-Florencia and Portell, Mariona},
biburl = {https://www.bibsonomy.org/bibtex/2655f780aad77a0070a383538067b4965/yourwelcome},
doi = {10.2466/PR0.103.5.145-160},
interhash = {feaa0ee6e799bd7869cb46bbf45d148e},
intrahash = {655f780aad77a0070a383538067b4965},
journal = {Psychological Reports},
keywords = {\& data humans, numerical psychology psychology, statistics},
number = 1,
pages = {145--160},
timestamp = {2017-01-09T14:01:11.000+0100},
title = {Overdispersion tests in count-data analysis.},
url = {http://www.ncbi.nlm.nih.gov/pubmed/18982948},
volume = 103,
year = 2008
}