Hi all,

I am trying to understand the formulae used by SPSS to calculate the SE and

p-values for the Tukey's HSD test.

The Tukey-Cramer SE formula is SE_Q= SQRT(MSE/2*(1/ni+1/nj))... However, in

the Tukey's multiple comparison output SPSS uses SE = SQRT(MSE*(1/ni+1/nj))

- it drops the two -. I thought that in this later case the p-value would be

obtained from the t-distribution with N-k degrees of freedom. It is not...

It is not also the Tukey's q distribution ... so I am lost here (specially

because for the Confidence interval the SE_Q and the q distribution are

used....)

Can any of the members of the list help me sort this out?

I will appreciate your thoughts on this.

Warm regards,

João

--

Sent from:

http://spssx-discussion.1045642.n5.nabble.com/=====================

To manage your subscription to SPSSX-L, send a message to

[hidden email] (not to SPSSX-L), with no body text except the

command. To leave the list, send the command

SIGNOFF SPSSX-L

For a list of commands to manage subscriptions, send the command

INFO REFCARD