Re: Constructivist Risk

Travis Gee (
Thu, 5 Oct 95 9:17:26 EDT

Lindsay Oades wrote:

>my Doctoral research involves developing a constructivist model of
>adolescent sexual risk taking. I have been toying with a definition of
>risk in PCP terms. Threat comes to mind immediately. However I
>believe that it is not synonomous with how we generally use the term
>risk. Is the notion of perceived risk the same as a high level of
>awareness of possible threat- and how does the choice corollary help
>here? I would appreciate any ideas on this one.

What I think you're interested in is attitudes towards sexual
activities, which would require some sort of attitude measure that
could distinguish PCP patterns from a grid. With a standard measure,
you could then locate particular grid patterns on a scale.

Guttman's Law of Attitudes states that "An item belongs to the
universe of attitude items if and only if its domain asks about
behavior in a (cognitive/affective/instrumental) modality towards an
object, and its range is ordered from (very positive to very negative)
towards that object."

Thinking this way, using Facet Theory (I've got some refs if this line
of thought interests you) one can construct a mapping sentence for the
area of risky sexual behaviour, holding constant a population facet
(adolescents), and sampling items from the universe of sexual
activities that are conceivably ordered from low to high-risk in some
kind of objective way. Try this on for size:


The extent to which (adolescent(x)) {feels} a sexual activity
{engages in}

Type of Activity

{clothed contact}
{unclothed contact}
{unprotected non-intercourse gratification - e.g. oral sex}
involving {protected non-intercourse gratification - e.g. dental dam}
{unprotected intercourse}
{protected intercourse}

Range of Possible

-----> : perceived risk.

Using this model, you can generate the Cartesian set of the products
of the domain facets as items and scale the range, for instance:

Rate the following items according to their riskiness:

Low Risk High Risk
I think that cuddling is 1 2 3 4 5
(Modality) (activity) (Range)

I feel that unbuttoning our clothes and
touching the skin is 1 2 3 4 5

and so on. You may want to add a facet based on your knowledge of the
research, but in essence, the Range is the set of responses that you
can observe, and the domain facets (modality, object, population
group, etc.) classify those responses. In fact, I just realized that
the "object" facet is really two, and without re-constructing the
whole mapping sentence, just edited in the "protected/unprotected"
bits. Essentially, there is a "protection" facet and an "activity"
facet, because each activity can be engaged in with or without
protection, be it clothing or condoms, etc. So really the first item
would be

I think that fully-dressed cuddling is 1 2 3 4 5
(Modality) (protection level) (activity) (Range)

and the multidimensional space of the itercorrelations of the items
would parse itself quite cleanly into two or three dimensions
(probably two).

With a standard against which to work (assuming that the scale
validates OK in your pilot), you can examine the way personal
constructs of identifiable groups come out. As an hypothesis, for
instance, maybe those who see everything above unbuttoning the clothes
as high-risk are more prone to apply sexual constructs to all elements
of the opposite sex. Or maybe those who think things are low-risk but
feel that the same things are higher-risk have a split in their
construct system between thinking and feeling-type constructs. Lots
of possibilities, but if I were on your committee, I'd be very curious
about what kind of standard measure you're basing your generalizations
on ;-)

As an aside, there's a paper all by itself in the contradictions that
you can study with this kind of measure, for example, does 'how far
you go' impact on the perceived risk of the activity? Objectively,
some types of unprotected non-intercourse gratification are probably
more risky than protected intercourse, but adolescents may well see it
the other way 'round!


Travis Gee () ()
() () ()()()()
() () ()
() ()()()()()()()()()
"In science, the more we know the more extensive the
contact with nescience." -Spencer