[EpiData-list] "Anova = Means command" in Epidata V.1Rel.1

epidata-list at lists.umanitoba.ca epidata-list at lists.umanitoba.ca
Thu Sep 15 00:28:58 CDT 2005


Hello ,
 
For comparing "means" of two independent samples , Epidata performs Anova (= t-test).
But it assumes that the two samples have equal variances.( I have compared the results of Epidata with SPSS and Epiinfo).It does Bartlett's test and the result of this test ,say, shows that two samples have unequal variances but it doesn't notice to Bartlett's test ! Then why does it do Bartlett's test ?
( In SPSS , for comparing two independent samples ,Levene's test is done first and then we can interpret our results for "means" according to the result for Levene's test. In Epiinfo , Bartlett's test  is done and if it is significant , Kruskal-Wallis is done. )
 
Thanks 
Reza

		
---------------------------------
Yahoo! for Good
 Click here to donate to the Hurricane Katrina relief effort. 


More information about the EpiData-list mailing list