Affordable Access

Statistical methods for analyzing developmental toxicity data.

Authors
  • Piegorsch, W W
  • Haseman, J K
Type
Published Article
Journal
Teratogenesis, carcinogenesis, and mutagenesis
Publication Date
Jan 01, 1991
Volume
11
Issue
3
Pages
115–133
Identifiers
PMID: 1686820
Source
Medline
License
Unknown

Abstract

A description and review of methods for performing per-litter analyses involving extrabinomial proportion response is provided. It is stressed that the litter should be regarded as the appropriate experimental unit for quantitative analysis in studies for teratogenic or heritable mutagenic effects. Attention is directed at statistical identification of possible treatment effects, such as a positive dose response to a chemical stimulus. The methods range from distribution-free, nonparametric analyses to models involving parametric distributions such as the beta-binomial density. It is seen that most current methods require computer implementation. When concern is raised over misspecification of assumptions critical to the statistical analysis, it is argued that relatively parameter-free methods are appropriate for use. These include statistical bootstrapping and rank-based analyses.

Report this publication

Statistics

Seen <100 times