You can do a tertiary split easily (although curiously there is not
consensus on exactly how to do this and how to handle things like
ties).
However, I would ask why you might want to do this - you are going to
lose power by throwing people away, and that's never a good idea.
What analysis are you doing that does not allow you to treat the
measure as continuous and analyse all of the data?
Jeremy
On 07/11/06, Tannaze Tinati <[log in to unmask]> wrote:
> Hi all...
>
> Wondering how to perform a tertiary split, either in spss or excel.
>
> I have taken a measure of anxiety from participants, and would like to
> create 3 groups..low, medium, and high anxious, but not sure how to create
> the group boundaries in terms of their anxiety scores. Essentially, I want
> to look at only the high and low people, and disregard anyone in the medium
> group to look for maximal effects. (When I initially wanted high and low
> groups i simply found the median, and scores above where assigned to high
> etc). Can i perform a tertiary split (if that's what it's called!) in spss
> or excel?
>
> Thank you in advance for your help!
>
> Tannaze
>
>
>
> Postgraduate research student
> Email: [log in to unmask]
> Office Telephone: +44 (0)23 8059 5078
>
> School of Psychology
> University of Southampton
> Highfield Campus
> SO17 IBJ
> United Kingdom
>
> Tel: +44 (023) 8059 3995
>
--
--
Jeremy Miles
www.jeremymiles.co.uk
|