home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!ALBNYVM1.BITNET!BCD80
- Message-ID: <STAT-L%92073020343775@VM1.MCGILL.CA>
- Newsgroups: bit.listserv.stat-l
- Date: Thu, 30 Jul 1992 20:23:26 EDT
- Sender: "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
- From: "Bruce C. Dudek" <BCD80@ALBNYVM1.BITNET>
- Subject: Re: Interaction effects
- Lines: 31
-
- With regard to lower order effects in the same equation as interaction
- terms, the "correct" answer depends on more information than was givn
- in the original question. It sounds to me like the reviewer might have
- a view of interaction effects predominated by experimental design texts.
- In the situation where experimental factors form a factorial design, the
- interaction vectors are uncorrelated with the main effect vectors.
- Therefore, no concern about centering or multicollinearity exists. In
- such experimental situations, most texts recommend that main effects are
- likely to be uninterpretable, or misleading when an interaction is present.
- This doesn't necessarily imply leaving them out of the equation, but
- the approach would involve looking as "simple" main effects rather than
- main effects. I suspect this is the reviewer's orientation.
-
- However, if
- the factors (predictors) are not design vectors from a controlled
- experimental situation, then the earlier comments about centering are
- appropriate. The point is that an interaction cannot be unambiguously
- interpreted, unless the correlated main effect is controlled for. One
- way of controlling is centering. You might refer to a recent Sage
- booklet by Jaccard, et al titled something like "Interaction effects
- in Multiple Regression".
-
- ************************************************************************
- ************************************************************************
- ** Bruce C. Dudek * **
- ** Dept. Psychology * BCD80@ALBNYVM1.BITNET **
- ** SUNY-Albany * BCD80@BALDER.ALBANY.EDU **
- ** 1400 Washington Ave * Fax: 518-442-4867 **
- ** Albany, NY 12222 USA * Voice: 518-442-4824 **
- ************************************************************************
- ************************************************************************
-