Date: Mon, 1 Sep 2003 11:57:39 -0400 "William B. Ware" "SPSSX(r) Discussion" "William B. Ware" Re: ANOVA or Kruskal-Wallis ANOVA To: Christina Cutshaw <3F52A805.A61A5A0@jhem.jhmi.edu> TEXT/PLAIN; charset=US-ASCII

If you have four groups, it seems to me that there should be four tests for normality, one for each group... The assumption pertains to the within cell distributions, not to the total sample...

WBW

__________________________________________________________________________ William B. Ware, Professor Educational Psychology, CB# 3500 Measurement, and Evaluation University of North Carolina PHONE (919)-962-7848 Chapel Hill, NC 27599-3500 FAX: (919)-962-1533 http://www.unc.edu/~wbware/ EMAIL: wbware@unc.edu __________________________________________________________________________

On Sun, 31 Aug 2003, Christina Cutshaw wrote:

> Dear List: > > I have a basic statistics question. I want to determine whether four > groups: A, B, C, D (total N=73) differ in their number of > child-related policy goals. I need to decide between using a one-way > ANOVA to compare means or the Kruskal-Wallis ANOVA to compare mean > ranks: > > Number of child-related policy goals: > Normality: Kolmogorov-Smirnov (p=0.007), Shapiro-Wilk=(p=0.015) > Normality was rejected > > Heterogeneity of variances: Levene Test 3.17, df1=3, df2=69, p=0.030 > Heterogeneity of variances was rejected > > With this information I decided to use a Kruskal-Wallis test and then > pair-wise Mann-Whitney tests to look at the differences between groups > in their number of goals. Did I make the correct decision? > > And, when I do the Mann-Whitney post-hoc comparisons, because of my > small sample size, is it appropriate to eyeball the means/mean ranks and > just compare the groups that look like they are significantly different > in order to avoid having to divide 0.05 by conducting all 6 comparisons? > > Thank you, > > Chris Cutshaw >

Back to: Top of message | Previous page | Main SPSSX-L page