Date:         Wed, 2 Apr 2003 21:42:09 -0500
Reply-To:     Christina Cutshaw <ccutsha1@JHEM.JHMI.EDU>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Christina Cutshaw <ccutsha1@JHEM.JHMI.EDU>
Subject:      Skewness
Content-Type: text/plain; charset=us-ascii

Joe, William, Michael and David-

Thanks for the help with the question about skewness.  To be more
specific about the sample sizes. I have a variable A with a sample size
of 242 a skewness of 1.113 and a SE (0.156), and is not normally
distributed.  I want to run two analyses:

state organizations: compare Variable A between three groups with n=19,
25, and 29 and
local organizations: compare Variable A between three groups with
n=44,55,and 70 (did not satisfy the assumption of homogeneous variances
for ANOVA)

I was planning on using the Kruskal-Wallis test to determine if their is
a difference between the mean ranks and then Mann-Whitney for pairwise
comparisons.

Does this strategy make sense?

Best,

Chris C.