Date:         Mon, 31 Jan 2000 16:35:26 +0100
Reply-To:     "Christian F.G. Schendera" <schendera@NIKOCITY.DE>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "Christian F.G. Schendera" <schendera@NIKOCITY.DE>
Subject:      Re: Q:multivariate stats
Comments: To: Alex Gray <Alex.Gray@RTP.PPDI.COM>, SASL <SAS-L@UGA.CC.UGA.EDU>
Content-Type: text/plain; charset="iso-8859-1"

Alex,

Just some notes:
First: In GLM / ANOVA independent vars are nominal / ordinal vars. Their
different classes cannot interact. Imagine the ind.var. GENDER: "Male"
cannot interact with "female".
Second: This question depends on your theoretical concept. Time effects may
depend from certain vars, or may represent independent "causality". Choose
your pick (and decide how to scale time: interval or the like).
Third: Mixed question: Coding time as an independent classification variable
is a valid approach. Just decide whether both categories are independent
from each other. To code 8 times of responses (of one var, I presume)
transforms one var into eight. If this provides you with more info, do it.
Please note that you transform the scaling of this var by this approach. Now
you might want to pick PROC CATMOD.
Fourth: Sorry, still working with V6.12
HTH,Chris