Attach file.
--PART.BOUNDARY.0.29563.emout07.mail.aol.com.826991619
Content-ID: <0_29563_826991619@emout07.mail.aol.com.220796>
Content-type: text/plain;
name="RATINGS"
Content-Transfer-Encoding: quoted-printable
Hemant,
=0D
It is true that ranks are widely perceived
as more primitive than are ratings. Ratings =
are more abstract. They assume that a
common essense has been refined and is =
used as the common denominator in a =
general system. But often we give up more
by shifting to ratings than we gain. People
may be able to better construe whole figures using
rankings and in ways that still add up. The
use of ratings requires that the whole figure
constructs be reduced from their true =
complexity. Such ostensible parsimony may =
alienate us from the deeper truths of the
person. This may gain us a certain flexibility
as psycholgists. We reduce the person to a =
semantic network, a construct dictionary,
rather than knowing the person as a person
first. =
=0D
The use of rating scales assumes a level
of abstraction that may be both unobtained
and undesireable. Nunnally found that when
ratings are "reduced" to ranks, the =
correlations between the ratings and their
ranks is usually in +.90's. There may thus
be little gained by using ratings in most data,
anyway.
=0D
With regard to personal constructs, I think
the impact or rating scales may be worse,
in general, than most people realize. First,
the use of ranks is probably not so much
more primitive as they are basic. Basic
should not have the same connotations as
primitive, but it does for many people.
Consider a young child ranking Mom as
more like Dad than Sister is like Dad. The
young child may make these judgements
on the basis of whole figure constructions.
Very early the young child may
be able to see Mom as generally more like
Dad than is sister- in many little ways and
in a simply "just so" manner at the level of
wholes. Psychologists may be inclined to =
conclude that since the young child easily
ranks whole figure constructs but does not =
use ratings well, that both whole figure =
constructs and ranks are primitive. The just
so of whole figure perception seems primitive.
The whole must be reduced.
=0D
As the child gets older, he learns to reduce
people to traits- that is he uses abstraction
more than construction. Dad and Mom are
reduced to big, sister to small. In elementary =
school a child learns that he can converse
with more people by using traits. More people
understand the trait "big" than "like my dad". =
Traits then become socially more useful than
whole figure constructs- in direct proportion
to the decay of actual community. In smaller
social contexts, the child could say "like my =
dad" and his friends- all having met his dad-
would know in great detail and in just so
fashion what "like my dad" implied. But as
the number of people in social interactions =
grows, the superficiality of knowledge grows. =
Consequently, someone's dad becomes a
simplistic equation of traits in the minds of
others who have never met the dad and may
care less and less to meet the dad, as they
"mature". Is this progress?
=0D
As the child grows old there is an increased
reliance on abstraction. It can get to the
point that a colleague's mention of his
dead child becomes mere evidence of
pathology. No one cares to ask such =
questions as "What was her name?" Such =
personal discourse would be too close to
whole figure construction, too "primitive for
educated people", too distant from the trait
mentality that would allow psychologists
to dismiss a dead child who suffered by a =
statement like "Children die from time to time, =
so quit your moaning Bill."
=0D
The typical psychologist sees the
aging child reduce his peers to the
abstractions of their bandwagon (clique)
and concludes that development =
progresses by the use of abstractions.
The ostensible proof is in the fact that
the growing child uses abstactions more
logically and with greater flexibility than
does the younger child..
Unfortunately, few psychologists also
look to see if the child's use of whole
figure constructs also gains in logical
consistency and integrative complexity. =
In may own empirical studies,
kids do become more logical and
integratively complex with both whole figure
constructs and traits, as they mature.
Thus whole figure constructs are not just
primitive. =
=0D
Ranking similarities between whole figure
constructs, as in the coordinate grid, does
not force the person to reduce people to
traits. Traditional grids do. Using a rating
scale in a coordinate grid would tend to
force the person to reduce whole figures to
an abstract trait system. Comparisons
would not be made between one person =
and another, but between a person and an
abstraction of commonality between a group =
of persons. The use of ranks assumes that =
direct comparisons between people can still
"add up", even if people are directly compared
with one another. Why reduce people to
ratings/ traits when we can get the estimates
of logical consistency, integrative complexity
and factor structure from the more direct
rankings? =
=0D
In other words, we can get more from
rankings than from ratings. They are
more basic because they do not require
the interjection of so much abstraction
in the comparisons of people. They do
require more differentiation on the part of =
the person taking the grid. But this is good.
Such differentiation will not hurt the person
but it could take the busy psychologist
more time to administer the grid. Perhaps =
the doctor should spend more time getting =
know the clients "personal" constructs anyway,
and less time dreaming about his vacation to
the Kelly conference. If the person can not do
the ranking because of inability to differentiate,
then let them use tied ranks. The doctor should
know how to help the client do this. Of course,
if you want to study peoples reductions of =
people to trait/interval systems, use ratings. =
But know what you are doing.
=0D
What do you think ?
=0D
Bill
--PART.BOUNDARY.0.29563.emout07.mail.aol.com.826991619--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%