Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: -test- after -anova , sequential-

Subject   Re: st: -test- after -anova , sequential-
Date   Tue, 22 Jul 2003 07:22:58 -0500

Joseph Coveney <> asks:

> Is there a reason why -test- after -anova , sequential- uses partial sums of 
> squares?

There isn't a really good reason, but for background as to why it
acts the way it does, consider how the -sequential- option (and
the default option -partial-) of -anova- work.  These options are
allowed on replay.  You can type

    . anova y a b a*b

which by default displays the Partial SS results.  You can replay
the anova and have it switch to showing Sequential SS results

    . anova , sequential

The default is always -partial-.  So, even if you say

    . anova y a b a*b, sequential

when you then replay that with

    . anova

you will be shown the partial SS results.  To replay the anova
showing sequential SS you have to say

    . anova , sequential

Stata's -anova- stores the same information regardless of which
of the -sequential- and -partial- options are specified.  What is
displayed is different -- derived from the underlying information
based on which of the options is specified.

By the time that -test- comes along after -anova-, it does not
know if -anova- was specified with the -sequential- option or
not. -anova- does not save that information since the
-sequential- and -partial- options are considered replay options.

There probably should be an option for -test- to use sequential
SS, just as there is for -anova-, but there currently is not.

Ken Higbee
StataCorp     1-800-STATAPC

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index