Nick Winter
> Nick Cox asks:
>
> > Why do you need a composite measure? It is often a good way
> > of blurring important distinctions. If in fact these measures
> > are highly related, then one will serve as well as any other.
> > If, as seems a little more likely, they measure rather
> > different things, it is not clear that any composite measure
> > will add much to looking separately at your different responses.
>
> Survey data often have lots of (presumed to be random) measurement
> error, in which case comparing individual items leads to
> interpretation
> of noise; scaling them gives a more reliable measure of the
> underlying
> construct.
>
I can subscribe to the general argument in many contexts, but it seems
very
pessimistic for what Seth Hannah sketched in the original question.
It's Seth's project, not mine, but I note
* There are four measures of prejudice. Lumping them into
one measure of overall prejudice seems to be moving away
from some interesting distinctions. I just asked "Why
do you need a composite measure?".
* The sample size is 8,916. If this isn't enough for
some structure to be evident in looking at single responses,
a different methodology is needed.
* A quick look at the frequency distributions Seth showed
(using among other commands -ordplot, sc(logit)- from SSC)
indicated some major location differences between
measures of prejudice.
Nick
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
