II. MATERIALS AND METHODS 



Fisher Exact/Cochran-Armitage Trend Analy- 

 ses -In addition to survival-adjusted methods, 

 the results of the Fisher exact test for pairwise 

 comparisons and the Cochran-Armitage linear 

 trend test (Armitage, 1971; Gart et al., 1979) are 

 given in the appendixes containing the analyses 

 of tumor incidence. These two tests are based on 

 the overall proportion of tumor-bearing animals 

 and do not adjust for survival differences. 



Historical Control Data: Although the concur- 

 rent control group is always the first and most 

 appropriate control group used for evaluation, 

 there are certain instances in which historical 

 control data can be helpful in the overall assess- 

 ment of tumor incidence. Consequently, control 

 tumor incidences from the NTP historical con- 

 trol data base (Haseman et al., 1984, 1985) are 

 included for those tumors appearing to show 

 compound-related effects. 



Nitrofurantoin, NTP TR 341 



34 



