Analysis of longitudinal "time series" data in toxicology

Christopher Cox, Deborah A. Cory-slechta

Research output: Contribution to journalArticlepeer-review

Abstract

Analysis of Longitudinal "Time Series" Data in Toxicology. COX, C, AND CORY-SLECHTA, D. A. (1987). Fundam. Appt. Toxicol. 8, 159-169. Studies focusing on chronic toxicity or on the time course of toxicant effect often involve repeated measurements or longitudinal observations of endpoints of interest. Experimental design considerations frequently necessitate between-group comparisons of the resulting trends. Typically, procedures such as the repeated-measures analysis of variance have been used for statistical analysis, even though the required assumptions may not be satisfied in some circumstances. This paper describes an alternative analytical approach which summarizes curvilinear trends by fitting cubic orthogonal polynomials to individual profiles of effect. The resulting regression coefficients serve as quantitative descriptors which can be subjected to group significance testing. Randomization tests based on medians are proposed to provide a comparison of treatment and control groups. Examples from the behavioral toxicology literature are considered, and the results are compared to more traditional approaches, such as repeated-measures analysis of variance.

Original languageEnglish (US)
Pages (from-to)159-169
Number of pages11
JournalToxicological Sciences
Volume8
Issue number2
DOIs
StatePublished - Feb 1987
Externally publishedYes

ASJC Scopus subject areas

  • Toxicology

Fingerprint

Dive into the research topics of 'Analysis of longitudinal "time series" data in toxicology'. Together they form a unique fingerprint.

Cite this