On 28 Sep 2000 08:06:54 -0700, [EMAIL PROTECTED] (Jenny Goodspeed)
wrote:
>
> I have a question regarding the creation of an
> overall scale from several likert-type items.
>
> I have a group of items that I'd like to sum to
> create the scale. Results of a factor analysis show
> that the items load high on one factor. The items
> also have a high alpha coefficient.
>
> However, some of the items have a 4-point scale,
> while the remaining have a 7-point scale. Is this a
> problem?
(You don't mention whether there are the same number of items in each
style; or why one range is 3, the other 6....; are they equally
important?)
If this is just a matter of numeric tags for the same verbal labels,
where one is (0,1,2,3) and the other is (0,2,4,6), then you can
rescore one in the obvious way.
Otherwise. The convention in psychometry is to use a simple average
the scores using the same motif. Here, an easy first-step is to get
two scores, as total or average. Then figure how you want to weight
them when you total or average THOSE two.
You might try to do your write-up before spending much time with the
data, since you probably want to write something that no one will leap
to criticize, while you cover it in 10 or 15 words.
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
