On Mon, 25 Nov 2002, siobhain mcardle wrote: > Could someone please advise if, when standardizing the independent > variables in moderated hierarchical regression, the resulting > unstandardized beta weights should be the same as the standardized beta > weights when no product (interaction) terms are significant.
Short answer: No. Longer answer: If those are in fact _product_ terms (rather than pure _interaction_ terms), they are almost certainly correlated with the variables that were multiplied to produce them. And the regression weights in any model are affected by correlations between their proper variable and any others. Reference to an even longer answer: See my paper on modelling and interpreting interactions in multple regression, on the Minitab web site: it's one of the "White Papers" on the site. HTH. Cheers! -- DFB. ----------------------------------------------------------------------- Donald F. Burrill [EMAIL PROTECTED] 56 Sebbins Pond Drive, Bedford, NH 03110 (603) 626-0816 [was: 184 Nashua Road, Bedford, NH 03110 (603) 471-7128] . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
