On 8 Jul 2002 04:45:53 -0700, [EMAIL PROTECTED] (David Emery) wrote: > > Rich Ulrich <[EMAIL PROTECTED]> wrote in message >news:<[EMAIL PROTECTED]>... [ snip, some of his and some of mine ... ] > > 'Ordinal' works out as pretty close to 'interval' for a lot of > > purposes, especially whenever you start with a scale with > > only a few points. You could consider using your scores > > of (integers, 1-to-5) for correlation. On the other hand, > > there is the really odd result that you write, where var#4 > > show the (relatively) huge difference, and the groups are > > badly out of order. (Unless that is a typographical error > > where 3.09 was 2.09.)
DE > > This is an interesting result but not a typographical error. [ ... snip some detail ] > Back to your first point, I think that treating interest ratings as > ordinal variables is better than treating them as interval variables. > There is no reason to assume each rating of interest is evenly > disbursed, as I believe an interval variable would imply. - Well, assuming that there was decent design (based on precedent) and pilot-data (to iron out wrinkles) going into the study, there is usually both an assumption and evidence that the simple ratings on items with 4 or 5 points do get used (pretty well) as good intervals. Unfortunately, I think I read your *evidence* that you failed to achieve *ordinal* results. -- That is, assuming that I remember that those 'groups' were really supposed to be in order, you can't use ANOVA on the scores or on the ranks of the scores for 'interesting'. Sorry, I have gotten too confused by the details posted, and the details in your paper when I read it. I can fathom a single dimension that ranges from "very interesting" to "uninteresting" but I can't add on "very uninteresting" as a symmetrical category -- I can say that something is stimulating or not stimulating, but to say that it is VERY MUCH not-stimulating ... implies what? numbing? creating an aversion? On the face of it [and this invokes 'face validity', ] to offer a choice <very interesting> seems to invite or require a potential response of hostility. Or something different from the simple continuum of 'interesting'. The results that were cited seem to support the interpretation that the subjects did not treat it as one continuum, either. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
