Date:         Thu, 15 May 1997 01:32:30 GMT
Reply-To:     Ana Marques <dmfl.etal@MAIL.TELEPAC.PT>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         Ana Marques <dmfl.etal@MAIL.TELEPAC.PT>
Subject:      homogeneity of variance
Content-Type: text/plain; charset="us-ascii"

Hi!
 
I am currently analising data resulting of a 2 X 2 factorial design.
Nevertheless, and as I performed test for homogeneity of variance, I came
across with highly significant p values (.000) both in Cochrans C test as
well as on Bartlett-Box F test. This, of course, means that I am dealing
with heterogeneous variances in my sample, and that I am violating a basic
principle for performing analysis of variance (ANOVA). What is more, I tryed
to create new dependent variables by affecting the old ones with "base e
logarithms", but the heterogeneity of variances stays the same...
Can you give me a hint about another way of transforming the old dependent
variables into new ones, so as to ensure homogeneity of variance.
 
Thanking you in advance !
 
Diniz Lopes