Statistical evaluation of sister chromatid exchanges: refined method.

The statistical problems in evaluating the results of sister chromatid exchange (SCE) scores can be solved by fitting log-linear models to the number of SCE scores (2). In each single test all measurements from all chromosome groups, both from the control and from the experimental sets are utilized. Previously, we analyzed for each chromosome group the number of metaphases with specified total SCE score for the group. We have refined this method to an analysis of the total number of chromosomes with specified SCE score over all metaphases. The results show that the refined method uses more information and can lead to conclusions which are at variance with the conclusions of the first method. By proceeding in this way, better use is made of all available information on the SCE scores and the shortcomings of the classical Student's t-test and chi-square (X2) tests are avoided.