ANOVA procedures and t-test equal variance
 

In looking through Excel I see that it has both the ANOVA procedure and a
t-test with equal variances and a t-test for unequal variances. One of the
assumptions that is made in both the "standard t test" and ANOVA is that the
variances are equal. Yet, Excel seems to have no test for homoscedasicity
(oh what a cool word). How do you know these procedures are ok to use if you
have not checked the assumption of equal variances?

Noosa -

In looking through Excel I see that it has both the ANOVA procedure and a
t-test with equal variances and a t-test for unequal variances. One of
the assumptions that is made in both the "standard t test" and ANOVA is
that the variances are equal. Yet, Excel seems to have no test for
homoscedasicity (oh what a cool word). How do you know these procedures
are ok to use if you have not checked the assumption of equal variances? <

Excel's Analysis ToolPak includes an F-Test Two-Sample for Variances.

- Mike
www.mikemiddleton.com