Working with living models sometimes the response is reflected in the variability more than the middle parameters, i.e., the variability can reflected the system response, p.ej., in toxicology models the variability in the response means something about the selective pressure, buy how to evaluated this variability?. The ANOVA Test evaluated difference between the medians using the variance between groups and behind groups, but i can??t use ANOVA if the two suposes don??t exists, normalitiy and variance homogeniety, then how to evaluated the differences in variance?, Barttlet Test evaluated the homoscedascity using the square sum and not the variance as a response variable. How could I evaluated the difference in groups with differents variances?. I don??t want use Kruskall-Wallis Test because my variable is continue and numerical. Do someone know what do must do it? Jos?? Herrera Baz??n Science Faculty UNAM http://search.t1msn.com.mx/