When Log-Normal and Gamma Models Give Different Results: A Case Study
- 1 May 1999
- journal article
- research article
- Published by Taylor & Francis in The American Statistician
- Vol. 53 (2) , 89-93
- https://doi.org/10.1080/00031305.1999.10474437
Abstract
The analysis of a simple dataset with two similar models is considered. A generalized linear model assuming a log-normal distribution and a generalized linear model assuming a gamma distribution are two models assuming constant coefficient of variation (CCV). Sources in the literature indicate that these two models are often interchangeable. However, in this real dataset—obtained from a clinical trial of a vaccine product—the two models do not agree. Reasons for this lack of agreement are explored. It is proposed that analyzing a dataset with both of the models may be an ad hoc robustness analysis of the dependence of the conclusions on the assumed model.Keywords
This publication has 6 references indexed in Scilit:
- Concurrent administration of inactivated hepatitis A vaccine with immune globulin in healthy adultsVaccine, 1999
- The Reverse Cumulative Distribution Plot: A Graphic Method for Exploratory Analysis of Antibody DataPediatrics, 1995
- Generalized Linear ModelsPublished by Springer Nature ,1989
- Multiplicative Errors: Log-Normal or Gamma?Journal of the Royal Statistical Society Series B: Statistical Methodology, 1988
- Science and StatisticsJournal of the American Statistical Association, 1976
- Generalized Linear ModelsJournal of the Royal Statistical Society. Series A (General), 1972