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

Re: st: SP/RM[1,2] EMS

From   David Airey <>
Subject   Re: st: SP/RM[1,2] EMS
Date   Fri, 22 Aug 2003 12:52:05 -0500

Ken kindly set me right, but left me wondering about Stata's ANOVA syntax:

David has misread the example in the FAQ.  He says that it uses
MS(E) for the test of P and I (his letters, or period and dial
the corresponding names for the example in the FAQ).

Look carefully at the -anova- command and at the output shown in
the FAQ.

    term                 tested using
    -----------------    --------------------
    noise                subject|noise
    period               period*subject|noise
    noise*period         period*subject|noise
    dial                 dial*subject|noise
    noise*dial           dial*subject|noise
    period*dial          residual
    noise*period*dial    residual

If I am following the mapping from his names to the FAQ example
names correctly, it appears that we are in agreement.

By the way, what I present in the FAQ as Stata output of -anova-
agrees with what Winer, Brown, & Michels show in Table 7.16 where
they show the ANOVA table for the example dataset.
I guess I would have expected the command syntax to then be:

. anova score noise/subject|noise period/period*subject|noise noise*period/perio
> d*subject|noise dial/dial*subject|noise noise*dial/dial*subject|noise period*d
> ial noise*period*dial, repeated(period dial)
repeated term

which fails,

rather than:

. anova score noise / subject|noise period noise*period / period*subject|noise
dial noise*dial / dial*subject|noise period*dial noise*period*dial ,
repeated(period dial)

I thought when a term did not specify the denominator, the residual was used. In the latter command, the terms period and dial are listed without a denominator so I thought they were using the residual. But when I look at the output, it's clear they are using a different (and correct) denominator. What am I not understanding in telling or not telling Stata what I want?


* For searches and help try:

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