Illusory Correlation
The interpretation of projective test responses de-
pends heavily on the psychodiagnostician’sexperi-
ence. Observations are accumulated regarding the
presence of certain test responses that supposedly
occur in connection with certain personality charac-
teristics. As a result, the clinician“learns”to associate
these test responses with specific personality charac-
teristics. They become“signs”of these underlying
personality characteristics. However, a great deal of
research evidence fails to support the meaningfulness
of many of these diagnostic signs. Little and Shneid-
man (1959) found, for example, that eminent clini-
cians performed only slightly better than chance in
making valid statements about patients on the basis
of their test responses. Chapman and Chapman
(1969) believe that one reason for this poor perfor-
mance lies in the tendency to rely onillusory correla-
tions between test responses and personality
characteristics. Chapman and Chapman found that
inthecaseoftheRorschach,clinicianstendedto
focus on test responses that have a high associative
value with male homosexuality. Thus, when they
observed such test responses of males as“This looks
like a man bending over”or“This is an anal open-
ing,”they quickly but mistakenly assumed they had
evidence for the presence of homosexual tendencies.
At the same time, they overlooked valid signs that
had low associative strength, such as threatening ani-
mals or animals that are humanized (e.g., a headless
monster or a woman with butterfly wings). Although
the“diagnosis”of homosexuality is not relevant to
the practice of contemporary clinical psychology
because homosexuality is not considered a mental
disorder (as it was inDSM-II, 1968), the important
point is that illusory correlation based on associative
strength can introduce a powerful source of error.
Incremental Validity and Utility
As mentioned in the discussion of the MMPI-2,
incremental validityrefers to the degree to which a
procedure adds to the prediction obtainable from
other sources (Hunsley & Mash, 2007; Meehl &
Rosen, 1955; Sechrest, 1963).
For an assessment procedure to be of real value, it
must tell clinicians something of importance that they
cannot get from merely inspecting thebase rates(prev-
alence rates) for the population of interest. If a clinician
in a state facility for those with mental retardation
reports via the WISC-IV that a certain patient should
be given a diagnosis of mental retardation, this hardly
comes as a surprise because we already know that
almost all of the institutional populace has mental retar-
dation. If, however, the test can tell the clinician some-
thing about the patient’s patterning of abilities that
will assist in planning vocational training, then some
incremental validity may be ascribed to the testing
procedure. Discovering from a Rorschach that a 70-
year-old widower“seems to be grappling with intense
feelings of loneliness”hardly represents a breakthrough
in incremental validity, even though the statement
maybeentirelytrue.AssertingthattheTATthemes
of a patient with schizophrenia reveal adjustment diffi-
culties or that the Rorschach responses of a patient
known to be depressed are suggestive of dysphoria
and sadness adds little if anything to existing knowl-
edge, even though it may appear that something cor-
rect and profound has been said.
A recent review of studies that assessed the
incremental validity of popular psychological mea-
sures (Garb, 2003) revealed the following:- Several studies offer tentative support for the
incremental validity of select MMPI-2 scales in
the prediction of personality disorder and
aggression and in the differentiation between
depressed patients and substance abuse patients. - NEO-PI-R scores have been shown to provide
incremental predictive ability in the assessment
of personality disorder, maternal responsiveness
to infants, and violence. - Studies support the incremental validity of
Rorschach measures of thought disorder, a feature
of psychosis, whereas there is little support for the
incremental validity of other Rorschach scores. - The incremental validity of the TAT or pro-
jective drawings has not been adequately
investigated.
246 CHAPTER 8