Abstract #300451


The views expressed here are those of the individual authors
and not necessarily those of the ASA or its board, officers, or staff.


Back to main JSM 2002 Program page



JSM 2002 Abstract #300451
Activity Number: 317
Type: Contributed
Date/Time: Wednesday, August 14, 2002 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section*
Abstract - #300451
Title: Critical Values and Power for a Test of Group Differences in Rodent Carcinogenicity Bioassays Based on the Beta-Binomial Distribution with Small Samples
Author(s): Ben Ahlstrom*+ and John Lawson+
Affiliation(s): Bandag Inc. and Brigham Young University
Address: 6500 49th Street, Muscatine, IA, 52761, USA 223B TMCB BYU, Provo, UT, 84602, USA
Keywords: likelihood ratio test ; rodent carcinogenicity bioassay ; two control groups ; Fisher's exact test ; overdispersion ; significance level
Abstract:

In this paper we show that false positives are possible when analyzing dichotomous data from rodent carcinogenicity bioassays using Fisher's exact test. The false positives occur when there are multiple control groups in a study and overdispersion is present. When comparing tumor counts from two control groups to one treated group in a rodent bioassay, 3×2 tables of counts are analyzed. Under various overdispersion scenarios (represented by combinations of beta-binomial parameters), we found 7-85% false positives by enumerating the distribution of Fisher's exact test. To reduce false positives, in the analysis of 3×2 tables, we propose the use of a likelihood ratio test for treatment-effect based on the beta-binomial distribution. Small sample sizes in rodent bioassays do not justify an asymptotic Chi-square distribution of the statistic. Therefore, critical values for the proposed test statistic were found by enumerating its distribution, and a small Monte Carlo Simulation was used to study its power. The simulation results indicate this likelihood ratio statistic controls the 0.05 significance level between .040 and .070 and that it has desirable power properties.


  • The address information is for the authors that have a + after their name.
  • Authors who are presenting talks have a * after their name.

Back to the full JSM 2002 program

JSM 2002

For information, contact meetings@amstat.org or phone (703) 684-1221.

If you have questions about the Continuing Education program, please contact the Education Department.

Revised March 2002