>So for example, the
>construct "low pay -v- high pay", found participants construing high pay
>as positive, and low pay as negative. This is done, of course, so that
>when using Likert ratings (say seven point) the negative pole would start
>at a rating of (-3) and the positive pole would ceiling at (+3)
If that's what you want to do, fine. But why?
> Here is my problem. Some participants ended up nominating
>constructs for which they insisted both poles be positive. For example
>the construct "works indoors -v- works outdoors" was seen by one fellow as
>embodying two desirable states of affairs. He liked the outdoors, and he
>also worked well indoors. To him, the two poles were opposite in meaning,
>but equal in value.
If you task is to reflect the world as he sees it, better think of
alternative way of analysing your data that drops the above assumption.
>I think this measure runs counter to the whole
>philosophy behind the grid, which is to elicit as faithful as possible a
>representation of the individual's construct system.
If you mean that the (-3) to (+3) arranegment does, yes, exactly.
> This is not an isolated case in my recent experience, but I have
>not seen anything about it in the literature. Has anyone encountered this
>one before?
Yes, it happens all the time. If you need, for computational reasons, to
force people into a straightjacket, I'd tend to avoid grids.
You'll find the following very interesting and relevant:
Yorke, M. (1989) "The intolerable wrestle: words, numbers, and meanings"
_International Journal of Personal Construct Psychology_, 2, 1, 65-76.
and the followup:
Adams-Webber, J. (1989) "Some reflections on the 'meaning' of repertory
grid responses"
_Journal of Personal Construct Psychology_, 2, 1, 77-92.
see also:
Yorke, M. (1983) "Straight or bent: an inquiry into rating scales in
repertory grids" _British Educational Research Journal_, 9, 2, 141-151.
Kind regards,
Devi Jankowicz
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%