LEVENE AND WELCH TEST

[samkelomotsa]

When doing an analysis of variance we assume that the data meet the assumptions of ANOVA. we even use the Levene's test of homogeneity of variances to test such. 

assuming that I have done a test of Homogeneity of (age at first egg) AFE Variance and received the output below from the Levens and Welch test respectively. 

Levene's Test for Homogeneity of AFE Variance ANOVA of Squared Deviations from Group Means
Source	DF	Sum of Squares	Mean Square	F Value	Pr > F
genotype	6	21146861	3524477	4.30	0.0003
Error	411	3.3714E8	820285

Because of significant differences between variable group means I then proceeded to the welch test results and they’re presented here:

Welch's ANOVA for AFE
Source	DF	F Value	Pr > F
genotype	6.0000	33.08	<.0001
Error	110.4

What is the next step now? do I just report the results of the welch test instead of the ANOVA results?

[Ksharp]

/*What code are you using? Like this : */
proc glm data=sashelp.heart(obs=1000);
class bp_status;
model weight=bp_status/solution;
means bp_status/hovtest;
ods output HOVFTest=HOVFTest OverallANOVA=OverallANOVA;
quit;

If  a test of Homogeneity of  Variance is significant(a.k.a  P value <0.05 ),that means Variance from different groups is different, or means you could not do ANOVA for this data.

But You could try non-parameter version of ANOVA (a.k.a  Kruskal-Wallis Test)

proc npar1way data=dataset ;  
class treat;
var var;
output out=WILCOXON_ANOVA  WILCOXON ;
run;

Also @StatDave   @SteveDenham @lvm  could give you better code / idea .

[samkelomotsa]

Thank you so much let me try.