Well, you've set yourself a lulu of a problem, especially if (as it seems -
forgive me if I misread you) this is your first experience with Grid and
PCP. I think you need to do some background reading first - if you go to
www.enquirewithin.co.nz you'll find a free download of my Business
Applications of the Repertory Grid, which is a pretty good guide to the
practicalities of Grid (less about PCP, but you don't get to PCP without
Grid) and lots of free hints and tips which should answer several of your
questions. One of the issues you'll find discussed in detail are the
difficulties inherent in using 'activities' and 'events' as elements -
basically, the more concrete your elements the easier they are for you and
for the interviewee to handle.

And - forgive me, I just don't want you to get deep in data without knowing
what they mean - there's a bunch of cultural/methodological issues to
consider. If you know Myers-Briggs, you might understand what I mean when I
say that having used it with Maori a great deal, they often ask 'is this me
at work or me on the marae?' and can present quite different profiles as
between the two. (Typically, 'me at work' is more low-risk and gives less
away). My Maori friends told me that different iwi (that's home tribes for
the audience) had different ways of construing behaviour. Then, I'm a
little puzzled about whether there's something recursive about having
schizophrenics construe schizophrenic experiences - are you going to use a
control group? And one more question: do you propose to use as an element
'A person sees someone or something that others cannot?' or 'An occasion
when you have seen someone or something that others cannot?' Very, very
different questions. One looks outward to the big wide world; the other
examines personal experience.

All of which takes me to the following two crucial questions: (i) have you
got access to a supervisor or other experienced person who can help you
with the ongoing issues you're bound to run into, and (ii) have you done a
full pilot? Before you sign up sixty people, take three or four through the
whole process and see what happens.


As to the question of how many elements and constructs you need: the
helpful answer is enough to fulfill your purpose. It's usual to start with
at least eight or nine elements, but if you look at what's said on the
website about the many different methods of analysis and their suitability
for different purposes that ought to clarify the question about the number
of constructs. BUT - Grid is first and foremost a conversation. The Grid
process generates a conversation which can be analysed in many different
ways, and every project should begin with an idea of what analysis method
will be used, but that shouldn't be one's first focus - as you should
discover when you've done a few pilots.

It seems as if you have factor analysis in mind. Factor analysis is only
one method of analysing Grid data, and in my view (I've been doing and
teaching Grid for more than 30 years) it can very easily lose much of the
subtlety which the Grid process can elicit; personally I would never
recommend it. There's really no substitute for looking at the detail of the
information - which takes me back to (i) what's your research hypothesis,
and (ii) pilot, pilot, pilot. In case that starts a huge row, consider the
following: way back in the 1950s, Osgood Suci and Tannenbaum invented a
technique called 'semantic differential.' The conclusion - very firmly
researched - was that if you boil all the words and phrases in the language
into a factor analysis, at the root there are simply three factors:
evaluation, which accounts for about 50% of the variance, and power and
activity which take about 25% each. Applying factor analysis to Grid data
is the first step down this road.

I may make Schopenhauer look like Pollyanna with a response like this, but
you (and your sixty schizophrenic subjects) deserve honesty and respect.
Over the years I've accumulated a large file which could be labelled 'I
spent eighteen months gathering Grid data and I've run it through seven
kinds of analysis and I still don't know what it means,' and I take a
little time out from earning a living to help them rescue something useful.
Your project is doable, and has a good deal of potential, but it needs a
lot of prior research and it will need access to supervision.

By the way, there have been many times when I've seen someone or something
that others cannot.

Take care,

Valerie Stewart.