Re: Counselor Constructs

Chris Evans (
Wed, 9 Aug 1995 16:23:56 bst

Jack Gerber said:
> I have been following this discussion with interest. One of the
> suggestions was to use a rep grid to investigate. I have been wrestling
> for years with the problem of trying to use a grid to investigate the
> personal contructs of a group of people. The problem is that (at least
> as I understand it) each person uses his or her OWN constructs. How does
> one compare them between people then?
> For instance, if you were to use a grid to investigate the PCs people
> have about their counselors, how do you compare one with another? Am I
> missing something?
I too have been following this thread with interest. I'd like to pick up
Jack's question as I have handled this with a study someone else started
before I was involved. She had collected grids from supervisors of
psychotherapists/counsellors and the grids had eight supervisees of their
experience and triadically (if I remember rightly) elicited constructs, i.e.
no commonality or "alignability" of either constructs or elements. However,
the researcher had a theory about superordinate constructs (or groupings of
constructs) she expected to see and which she expected to differ in their
prevalences between groups of supervisors of different theoretical
orientations. What we did was for her and two of us who were essentially
blind to her ideas to sort the constructs (written out onto small cards)
according to the superordinate groups to which we thought they belonged. We
did some conventional work to look for significant level of agreement between
us (this was actually pretty significant) and then tested her expectations
about theoretical orientation differences conventionally using majority
allocations of constructs to determine their membership of superordinate
groups. The sample sizes weren't huge and we haven't pursued it to
publication anywhere but I think the methodology is robust. You could adapt
it to a much less pre-ordained system by allowing free groupings of constructs
on cards by your fellow researchers (blind to your thoughts or not) and then
looking for emergent superordinate groupings.

I liked the method. I thought it kept the idiography but allowed some
sociality commonalities to be extracted fairly clearly. Some constructs used
were clearly entirely idiosyncratic but a fair proportion did fit her system.

If was going to have a look at within-grid inter-construct correlations to see
if those within superordinate groups tended to be higher than those across
groups but didn't do this as there were rather few repetitions of constructs
within superordinate groups so there would have been little to analyse. (Was
going to involve some rather nasty SAS or SAS/IML code if I remember rightly
too!) That means she didn't actually need to get people to complete the grids
ratings for her final analyses but she did look at some of the individual
grids to get a feel of inter-construct correlations/clusters.

I hope this contributes to the thread.


Chris Evans, Senior Lecturer, ||| Psychotherapy Section
Cranmer Terrace ||| Dept.Ment.Health.Sci.
London SW17 0RE ||| St. George's Hosp.Med.Sch.
Britain ||| University of London
Tel/fax.: (+44|0) 181 725 2540 ||| Email:
World Wide Web: