re: abstractions measureable?

Devi (
Fri, 13 Sep 1996 22:03:43 +0100

Duane Steward gives some background:

>I wish to make use of an individual's personal constructs regarding
>healthiness in decision support technology. More specifically, I am
>attempting to utilize a person's constructs to compute a utility value
>(micro-economic theory) for specific health states.

and then asks the intriguing question:

>The above leads to a desire to find measureable attributes or bipolar
>constructs for any abstractions elucidated in the personal constructus. The
>question then arises, "Is there such a thing as an abstraction that cannot
>be described in measurable constructs (attributes)?"

bearing in mind that he is prepared to "Ladder Down" to find more
behaviourally explicit operationalisations of initially "too-abstract"
constructs; in short,

>how safe is it to say we will always find measurable constructs if we dig deep

IMHO, this is a question to be put to each and every respondent in your
study as part of the process of construct elicitation, rather than
something that can be answered a priori.

While we speak as if constructs "existed in there", when researching or
interacting in other ways with people, in fact they don't: all that it's
possible to do is to negotiate a shared meaning with the person providing
the information, which we then call a (measurable) construct; and all the
techniques for construct elicitation, repertory grids or anything else, are
simply ways of recording the outcomes of the negotiation over our
respondent's meaning in a mutually satisfactory way.

So laddering down is useful only because, in elucidating "in what way" is a
set of elements "THIS" as opposed to "THAT", our respondent expresses his
or her initially "abstract" construct in a way which is more meaningful to
ourself while at the same time being helped to confirm/negate that we have
understood the distinction he or she is making. It isn't a technique for
getting at something "in there" which exists, sui generis, qua "measurable

I'm taking this purist line because of something you said later in your e-mail:

>I give room for this not to hold in the individual who is "not of this
>world" (may I say it that way, not being a psychologist?), but I confess my
>ambition is not to accomodate the personal constructs from the extremes of
>irrational humanity. Rather, I seek to provide a pragmatic tool for the bulk
>of physiologic medical decision making.

I know what you mean! The most frustrating rep grid I ever obtained was
from an accountant (_the_ prototype extraterrestrial for all organizational
psychologists who work in business schools, you understand) who couldn't
understand what I was on about and only succeeded in providing 2 constructs
after 40 minutes' work.

BUT BUT BUT! all that means is that he didn't find it helpful to think
about his understanding of things as bipolar this or construed that;
"hoping to get a picture of how you think about things / construe events /
give meaning to things" was gobbledegook to him no matter how user-friendly
the way I expressed it; as was triadic elicitation, self- characterisation
technique, or anything else we tried.

The observation that he was an thick, unthinking, obtuse, stereotyping,
brainless pinstriped bloody SUIT!! was okay as a way for me to let off
steam in the privacy of my own mind, but not terribly useful otherwise.

Come to think of it, there's another advantage to seeing elicitation and
laddering down as a process of negotiation: viz., are the results
meaningful to other people or just something which you and the client have
mutually intuited on a "mhmm, know where you're at, Man" basis during your
interaction? Seems to me that to arrive at some fairly behaviourally
explicit expressions of a construct is likely to assist you to achieve the
reliability and verifiability of any content analysis which you find
yourself doing, explicitly or only in the way you interpret your data to
yourself, in thinking about the implications of your findings.


You know that both Colin Eden (uses directed graph techniques and is keen
on the negotiation paradigm) and Patrick Humphreys (uses grids prior to
automated multiattribute decision theory computations on inter-construct
preference choices) have covered what appears to be similar ground to
yourself? The references are a bit hoary now, but if they're any use to
you, do please give me a shout.

Kindest regards,

Devi Jankowicz