Rich Ulrich <[EMAIL PROTECTED]> wrote: > On 7 Mar 2002 05:49:13 -0800, [EMAIL PROTECTED] (Mobile Survey) > wrote:
>> Rich Ulrich <[EMAIL PROTECTED]> wrote in message >news:<[EMAIL PROTECTED]>... >> >> > If one item has a terrible distribution, it is going to >> > correlate poorly. Okay, that means, Drop it. >> > For the statistics, you drop both: bad content and >> > bad wording. (Wording is usually easier to mend.) >> No. the univariate distributions are normal. But what I meant was that > For reporting, use Pearson. People are going to expect it, > For most purposes, you are stuck with the Pearsons; > Spearman's, especially, serve no end that I think of. > For factor analysis, and other 'latent structures', some > people have very large Ns and compute tetrachoric r for > their 2x2 contrasts. There's also a 'polychoric r' (more > than 2 categories) which is mentioned even less. If you use LISREL, this latter is/was the only option for analysis of ordinal data. You can perform some testing of the appropriateness of the model. The progam Mx, which is mainly aimed at genetic SEM, can fit a threshold model (polychorics) to up to 20 variables via full ML, and allowing partially missing data. LISREL uses a WLS approach. Polychoric r's have the advantage of dealing with (some) floor/ceiling effects. -- | David Duffy. ,-_|\ | email: [EMAIL PROTECTED] ph: INT+61+7+3362-0217 fax: -0101 / * | Epidemiology Unit, The Queensland Institute of Medical Research \_,-._/ | 300 Herston Rd, Brisbane, Queensland 4029, Australia v . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
