Take a situation in which a student is denied
admittance to graduate school because of the appli-
cation of such empirical predictors as GPA and
Graduate Record Examination (GRE) test scores.
First, Dawes notes, some will argue with such pre-
dictors on technical grounds. They claim that the
indices are short-term and rather unprofound. The
plea“I just know I could succeed if they would
only give me a chance”is less an argument than
an expression of hope. The antistatistical argument
often claims that there are expert judges“out there
somewhere”who could do as well or even better
than formulas. But somehow, these experts never
seem to be produced!
Second, this approach may be rejected for psy-
chological reasons. Many persons easily remember
those instances in which their intuition was right
but conveniently forget those occasions when it
was wrong. To take another example, a clinician
may unconsciously work harder with a client for
whom the clinician has predicted success in therapy.
A positive outcome will then prove to the clinician
that his or her clinical hunch was right!
Third, there are ethical sources of resistance.
Some people have the idea that reducing an appli-
cant to a set of numbers is unfair or dehumanizing.
Dawes (1979) discusses this argument:
No matter how psychologically uncompel-
ling or distasteful we may find their results to
be, no matter how ethically uncomfortable
we may feel at“reducing people to mere
numbers,”the fact remains that our clients
arepeoplewhodeservetobetreatedinthe
best manner possible. If that means—as it
appears at present—that selection, diagnosis,
and prognosis should be based on nothing
more than the addition of a few numbers
representing values or important attributes,
sobeit.Todootherwiseischeatingthe
people we serve. (p. 581)The Case for a Clinical Approach
The difficulty with a statistical approach that relies
on prediction models or regression equations is that
clinical psychologists would need a multitude of
them to function as clinicians. The field currently
does not have well-established, cross-validated for-
mulas to predict therapy outcomes, make interpre-
tations during the course of a therapy session, or
recommend a special class rather than institutionali-
zation. Should the clinician suggest bibliotherapy, a
hobby, a marriage counselor, a trial separation, or
what? The busy, harried clinician does not have
available a regression equation for even important
decisions, let alone the pedestrian judgments that
must continually be made. It was Meehl (1957)
who long ago said, “Mostly we will use our
heads, because there just aren’t any formulas”
(p. 273). Unfortunately, the situation has not
changed much since Meehl made this observation.
Of course, when specific outcomes are to be
predicted and the clinician has enough time to
develop good formulas, the clinician can easily be
outperformed by those formulas. We will review
this evidence shortly. However, even here, the
clinician’s judgment can add something in some
instances, especially when the sample is relatively
homogeneous. Suppose, for instance, that the for-
mula for selecting students for graduate training
depends solely on Graduate Record Examination
scores and undergraduate grades. The formula
would probably do quite well in selecting from an
initial, heterogeneous sample of applicants those
who will do well. However, from that point on,
clinical judgments regarding motivation or person-
ality features may be quite helpful in further dis-
criminating among those selected. That is, the
final sample is so selective that previous grades and
test scores may not be very discriminating. Clinical
inferences may become useful after the initial
screening because they provide extra data that relate
to success in training. Holding large amounts of
data in our heads and integrating them are not
what we humans excel at (Dawes, 1979). Clinicians
should use computers and formulas for that and save
their own mental powers for what they do excel
at—selecting what to look at and deciding what
to do with the results.
Another important contribution involves the
clinician’s function as a data gatherer. For example,
it may turn out to be important to know aboutCLINICAL JUDGMENT 289