Psychology of Being Right
Psychology of Being Right
Psychology of Being Right
396
ARIE W. KRUGLANSKI
JUDGMENTAL ACCURACY
dard, notably a consensus between the subject's and the standard setter's judgments. Admittedly, the consensus definition
does not convey a given judgment's superiority over others,
whereas the correspondence definition does convey a standard
setter's superiority. These considerations imply an asymmetrical consensus definition of accuracy. A rather different interpretation is implied by the pragmatic definition of accuracy considered next.
Accuracy as pragmatic utility. Recently, Swann (1984) and
McArthur and Baron (1983) proposed pragmatic utility as a
standard of accuracy. Interestingly, the pragmatist standard is
open to different interpretations with correspondingly varied
implications for accuracy assessment. The following examines
the issues involved.
According to the pragmatist doctrine, a judgment is deemed
accurate if it is useful. But what exactly is "useful"? It seems
reasonable to define as useful any action or belief that facilitates
goal attainment, that is, brings about a desired reward. Such a
definition allows two separate senses of utility, a subjective sense
and an objective sense. In the subjective sense a reward need
not be an external object. Instead, it could be an internal experience. For example, one might choose to believe in the existence
of God because such a belief put one's mind at rest and in that
way was useful. Thus, usefulness could be highly subjective and
reflect a motivational bias toward beliefs whose contents are
comforting, pleasing, or otherwise desirable.
Beyond its subjective sense, utility could be defined objectively, for example, in terms of successful task accomplishment
by the actor. For example, i f a student believes that effort mediates academic success, and she or he exerts effort and attains
success, the accuracy of the original belief will be supported.
Indeed, it is such an objective sense of utility that social psychologists generally have in mind when discussing the pragmatic
standard of accuracy. For instance, Baron (1988) noted that the
organism's perception of a crack's width is accurate if it can
crawl through it. Similarly, Swann (1984) suggested that a perception is (pragmatically) accurate if the target upholds the perceiver's expectations.
Strictly speaking, however, what defines accuracy in the foregoing examples is a correspondence between a judgment and a
criterion rather than utility per se. The fact that an organism
succeeds in crawling through the crack is criterial evidence that
corroborates the perception of the crack's sufficient width.
Moreover, the judgment that the crack was wide enough is not
restricted to the (crawling) organism as such but presumably
would be reached by any observer exposed to the same information, even if such a judgment was largely devoid of utility to the
observer. Similarly, the target's behavior in accordance with the
perceiver's expectations establishes that those expectations were
accurate by the correspondence definition of accuracy rather
than because of their pragmatic usefulness.
Note that the subjective sense of utility defines accuracy from
the subject's perspective, whereas the objective sense of utility
adopts the perspective of the standard setter. The two perspectives need not coincide. For instance, motivational biases could
be idiosyncratic to the subject and not necessarily shared by the
standard setter. Thus, a parent may be motivationally disposed
to judge that a child is able and likable, whereas the teacher,
397
398
ARIE W. KRUGLANSKI
JUDGMENTAL ACCURACY
Taft, 1986) For the most part, the negative view held by a person's antagonist is unlikely to be shared by the target. In turn,
an absence of shared reality may result in predictive inaccuracies. For instance, if B's antagonist believed B to be cruel and
callous, whereas B believed instead that she or he was warm and
kind, the antagonist may err in predicting B's behavior toward
a neutral C.
It thus appears that several motivational factors may determine whether interacting persons subscribe to a commonly
shared reality and hence whether they are accurate in predicting
each other's behavior. The diversity of possible motivations
highlights the difficulty of estimating people's general accuracy
in natural settings.
A final problem with comparing the prevalence of accuracy
in natural and artificial settings is determining just what constitutes a natural environment for a perceiver. This problem is
particularly acute with human perceivers in today's rapidly
changing world; in these circumstances, change itself(from old
to new environments) may represent a natural state of affairs
(Einhorn & Hogarth, 1981; Slovic, Fischoff, & Lichtenstein,
1977). An absence of stringent criteria for what is natural poses
a danger of conceptual circularity: Thus, one may argue after
the fact that a situation was artificial or natural, given that an
erroneous or an accurate judgment has occurred. For instance,
succumbing to a car salesman's tricks may be excused on the
grounds of the artificial nature of the car lot situation, whereas
accurately seeing through such legerdemains may be explained
by the considerable naturalness of such settings to members of
the "car generation." In conclusion, the notion that persons are
more accurate in natural settings than in artificial settings is
open to various exceptions and on close examination seems less
than totally compelling.
Circumscribed versus global accuracies. Recently, Swann
(1984) argued that the circumscribed, or situation-specific,
form of accuracy is generally easier to attain than global or
transsituational accuracy. Specifically,
forms of circumscribed accuracy usually require only that perceivers predict target behaviors in a limited range of social situations
. . in which perceivers themselves may often be present. . . . This
not only reduces the number of potential influences on target behavior that perceivers must consider, it also raises the probability
that perceivers will be familiar with these influences and with how
they are apt to affect the behavior of targets. (Swarm, 1984, p. 462)
Note, however, that from a social cognitive, constructionist
perspective (Fiske & Taylor, 1984; Kruglanski, 1989), the
"number of potential influences" that perceivers may consider
is not necessarily constrained by the specificity or generality
of the relevant judgment. Such potential influences represent
perceiver hypotheses about the determinants of target behavior.
In turn, the number of hypotheses generated by the perceiver
may depend on a variety of cognitive and motivational factors
internal to the individual; it is unlikely to be determined exclusively by external factors such as the topic of the hypotheses or
their generality versus specificity.
Thus, a person could entertain numerous interpretative
hypotheses about a specific situation and only a few hypotheses
about a general situation, all depending on (a) the number of
relevant constructs momentarily accessible to this individual
399
(Higgins & King, 1981), (b) his or her stable pool of available
constructs on the topic (Higgins, King, & Mavin, 1982), (c) his
or her (epistemic) motivation to generate or search for further
relevant notions (Kruglanski, in press); and (d) his or her creativity or constructive ability (cf. Amabile, 1983). In summary,
the notion that some stimulus situations (e.g., circumscribed
ones) generally invite fewer interpretative notions than other situations, and hence that they afford a greater likelihood of accuracy, may be debated from the constructivist perspective.
The argument that accuracy in circumscribed settings benefits from perceivers' familiarity with the situation also needs to
be qualified Specifically, familiarity may not invariably improve accuracy and occasionally may even undermine accuracy. In this vein, Cantor and Kihlstrom (1987, pp. 115-116)
commented on the "double-edged sword" of expertise. According to their argument, the processing of ample information on
a topic may lead to the development of elaborate, firmly entrenched schemas that are rather refractory to inconsistent information. To the extent that such schemas deviate from a veridicality criterion, the substantial familiarity on which they are
based may detract from the judge's readiness to modify them
and hence to ultimately lower accuracy.
Of unique social psychological interest is Swann's (1984) assertion that circumscribed accuracy is often higher (or easier
to attain) than global accuracy because of identity negotiation
processes in social perception. Two separate such processes are
distinguished: behavioral confirmation and self-verification.
Behavioral confirmation refers to targets' tendency to behave in
ways that confirm the expectancies of perceivers (cf. M. Snyder
et al., 1977). Self-verification refers to targets' tendency to behave in ways that confirm their self-concepts. According to
Swann (1984), "if target self-conceptions and perceiver expectancies differ, both global and circumscribed forms of accuracy
are likely to be higher if self-verification occurs" because in such
circumstances "perceivers are induced to bring their actions
and beliefs into harmony with target self-conceptions [hence]
they will be equipped to predict how targets will behave in the
future because targets will theoretically continue to behave in
accordance with their self-conceptions" (p. 466). By contrast,
processes of behavioral confirmation could contribute to circumscribed accuracy while detracting from global accuracy because "targets may behave quite differently once they escape the
constraining influence of the perceiver" (Swann, 1984, p. 466).
Although Swann's conclusions may often hold, behavioral
confirmation processes need not invariably lower global accuracy. First, the target could occasionally use his or her own behavior as a cue to his or her properties (attitudes, traits) and
come to internalize perceiver expectancies via dissonance reduction or self-perception processes (Bem, 1972) If that occurs, a perceiver who expected the target to behave similarly
across situations could be right. Second, at least on some occasions the perceiver could take into account the situational constraints (including his or her own influence) on target actions
and not expect such behaviors to generalize to alternative contexts For instance, a psychoanalyst could recognize that a client's aggressiveness represents a transference process uniquely
engendered by the therapeutic setting. Similarly, the powerful
boss may interpret an employee's friendliness as ingratiation
400
ARIE W. KRUGLANSKI
Accuracy-Mediating Process
As has been seen, the debate about persons' general accuracy
has led to the consideration of putative boundary conditions
for accuracy (i.e., of naturalistic vs. artificial conditions and of
circumscribed vs. global conditions). In turn, discussion of
boundary conditions typically hints at the general process assumed to underlie accurate judgments. Aspects of such an implied process are now examined in some detail.
Amount of information considered. Several authors suggest
that the more information a perceiver takes into account, the
greater the likelihood of his or her being accurate. McArthur
and Baron's (1983) ~irgument that active perceivers are typically
more accurate than passive perceivers could be interpreted in
terms of the greater amount of information that active exploration may afford. Similarly, their notion of "sins of omission"
refers to cases in which the perceiver misses part of what is
afforded because of attentional selectivity or because the stimulus array is impoverished. In both cases, therefore, errors are
traced to informational deficiencies.
Einhorn and Hogarth's (1981) suggestion that feedback from
the environment enhances accuracy implies that the ability to
receive sufficient amounts of relevant information improves accuracy. Finally, Swann's (1984, p. 462) suggestion that circumscribed accuracy is easier to attain than global accuracy is
premised in part on the assumption that in the former case,
perceivers have more relevant information at their disposal.
On close examination, however, the relation between the
amount of information and accuracy seems complex. First, the
terms informational sufficiency and informational insufficiency,
which are often used to suggest that more information means
better accuracy, are circular. Such terms typically fail to specify
in advance a definite informational quantity and are used after
the fact, that is, after an accurate or inaccurate judgment has
occurred.
Second, as several authors have stressed (Campbell, 1969;
Weimer, 1979), any amount of information is compatible with
multiple alternate hypotheses. Thus, one may continue to hold
on to an inaccurate judgment despite considerable amounts of
information that though consistent with the "correct" alternative (which the individual may have failed to entertain) are
equally consistent with the "incorrect" hypothesis. In other
words, i~onsiderable amounts of information could be nondiagnostic ~(tI'rope & Bassok, 1983) with regard to the correct and
incorrect hypotheses.
Third, as was noted earlier (Cantor & Kihlstrom, 1987), ex-
JUDGMENTAL ACCURACY
rect; thus, the tendency to abandon it on the basis of further,
possibly invalid, information could increase the likelihood of
error. The foregoing arguments suggest that increasing incentives (for correct judgments) need not reduce the incidence of
error. Reviews of the relevant empirical literature (e.g., Einhorn
& Hogarth, 1981) corroborate this conclusion.
Normative models ofjudgment. It has been implied that the
process whereby accurate judgments are formed is represented
by the normative models of inference (Kahneman et al., 1982;
Nisbett & Ross, 1980). However, normative models also can be
conceived of as constructed, hence potentially fallible, representations of reality (Kruglanski & Ajzen, 1983). As Einhorn
and Hogarth (1981) noted, a normative model may be inapplicable to given circumstances (e.g., the model's assumptions may
not hold). When this occurs, judgments derived from the model
will be inaccurate (i.e., will deviate from the criterion). Thus,
normative models do not appear to safeguard accurate judgments, at least not in an absolute sense.
In summary, the accuracy literature has identified several
process variables that are implied to generally improve accuracy. However, on close scrutiny, possible exceptions to those
arguments become apparent. In particular, high degrees of motivation, large amounts of information, or the use of normative
models may not reliably improve accuracy. It seems fair to conclude that as of now, no compelling analysis exists concerning
the general process whereby accuracy is obtained.
401
Specifically, if the criterion itself is tentative and potentially revocable, the same judgment that was previously discrepant
from the criterion, and hence considered incorrect, could now
coincide with the (revised) criterion, and hence be considered
correct. Similarly, a previously correct judgment would be incorrect in reference to the revised criterion.
Consider that any judgment is determined by a combination
of the relevant variables at appropriate levels (e.g., an appropriate extent of pertinent information, an appropriate degree of
motivational involvement). If the accuracy criterion shifts, the
very same combination of variables could give rise to gorrect
and incorrect judgments in different situations. As Funder
(1987) aptly noted, the process that promotes errors in one setting may mediate accurate judgments in another setting (see
also Kruglanski & Ajzen, 1983). Thus, it does not seem feasible
to hope that a general accuracy-mediating process will be uncovered. Nor does it seem feasible to hope for a meaningful answer to the question of whether persons are generally accurate
or inaccurate in their judgments: Admitting the possibility of
criterial shifts, different answers to this question could be warranted at different times.
The discussion so far may seem to paint a gloomy picture
for the future of accuracy research. Actually, such a negative
outlook befits only general questions about human accuracy. In
contrast, situationally specific questions about accuracy appear
more tractable. Indeed, the remainder of this article explores
possible directions for accuracy research within situationally
specific paradigms. It is suggested that in such cases, accuracy
issues are profitably approached from the perspective of the
process that governs the formation of all judgments. Thus, although historically the study of the judgmental process supplants that of accuracy (Funder, 1987; Jones, 1985), I propose
that the two perspectives may be fruitfully combined.
A Process-Based A n a l y s i s o f S i t u a t i o n a l A c c u r a c y
The foregoing conclusions serve as a basis for a reanalysis
of the accuracy problem in social perception and cognition. A
general framework for conceptualizing accuracy is outlined and
applied to two separate paradigms for possible accuracy research. The realist paradigm approaches the study of accuracy
from the external standpoint of the standard setter. The phenomenal paradigm approaches accuracy from the internal
standpoint of the judge. These notions are now considered in
greater detail.
Accuracy Framework
Consider the definition of accuracy as a correspondence between a judgment and a criterion that constitutes, in turn, some
standard setter's judgment. Such a conception of accuracy implies a compound judgment (or metajudgment) that consists of
the following simple judgments: (a) the target judgment, (b)
(judgment of) the criterion, and (c) (judgment of) correspondence between judgment and criterion.
When the standard setter is someone other than the subject,
accuracy may be investigated from a realist perspective. In other
words, the standard is assumed to represent the external reality,
402
ARIE W. KRUGLANSK1
and the key question is what may cause the subject's judgment
to mirror or deviate from such reality. When the standard setter
happens to be the subject him- or herself, accuracy may be investigated from a phenomenal perspective. In such a case the
question centers on the perception of accuracy, that is, on the
degree to which the subject regards two of his or her own judgments (the target judgment and the criterion) as correspondent.
In both the realist and phenomenal paradigms, accuracy issues are assumed to be closely tied to the judgmental process at
large. Thus, before pursuing the discussion of the accuracy
topic, a brief outline of such a process is given.
Realist Paradigm
The realistic study of accuracy stresses the psychological factors that lead subjects' judgments to coincide with a given situational criterion. In such a framework the standard setter's judgmental process is neglected. Rather, the focus is on the subject's
judgment and its degree of correspondence with an externally
defined reality. Most social cognitive research on error and bias
falls into this category, for example, the study of cognitive heu-
JUDGMENTAL ACCURACY
ristics (for reviews, see Kahneman et al., 1982) in which lay
judgments are compared with a priori assumed normative criteria. Application of the judgmental perspective may afford a
systematization of such research and may suggest new ways of
thinking about issues of realistic accuracy.
In particular, the notion of the judgmental process suggests
several categories of factors that may lead a judgment's content
to coincide with or deviate from a specific criterion. Those factors are considered now in some detail.
Availability of inferential rules. A person's tendency to render
(situationally) accurate judgments should depend on the relevant inferential rules that she or he has available in memory. As
was noted earlier, such inferential rules are if-then statements
that link a given category of evidence with a given judgment.
Thus, the same information (or evidence) could serve as a basis
for correct or incorrect judgments, depending on the specific
inferential rule being used.
The research literature in social and cognitive psychology attests that persons have in their mental repertories a variety of
inferential rules on diverse topics and that those can detract
from or contribute to judgmental accuracy in specific contexts.
For instance, some people may have acquired various judgmental heuristics (e.g., the representativeness heuristic discussed by Kahneman et al., 1982) that imply judgments at variance with normative statistical predictions. To the extent that
the latter are assumed to represent the criterion in some situations, the use of heuristics may be said to propagate errors (Nisbert & Ross, 1980).
Some inferential rules may relate to persons' conceptions of
their own selves or to their self-schemata (Markus, 1977). Depending on the situational criterion, such self-schematic rules
may facilitate or hinder accuracy. Thus, some individuals may
use negative self-schemata in reference to given classes of judgment, whereas others may use more positive schemata. According to Beck's theory of depression (Beck, Rush, Shaw, & Emery,
1979), for example, negative self-schemata may systematically
influence the judgments of depressives. "If it was my performance, it must have been tow" is a hypothetical example of a
self-schematic rule that may lead a depressed person to underestimate the level of his or her performance. On the other hand,
if the objective performance was in fact low, the very same rule
may promote accuracy.
In this connection, Dykman, Abramson, Alloy, and Hartlage
(1987) found that depressives (in comparison with nondepressives) erred more in rendering inappropriately negative selfjudgments when their negative schemas were relevant to the experimental task. Both depressives and nondepressives exhibited
positive errors, negative errors, and accurate responses, depending on the way in which their preexisting schemata related to
the situational criteria of correct responding. The Dykman et
al. (1987) research furnishes important evidence that "depressive realism" (Abramson & Alloy, 1981; Alloy & Abramson, in
press) is situationally specific rather than general. It may depend
on whether depressed or nondepressed subjects' judgmental
processes yield outcomes that correspond with the situational
criteria of accuracy.
Although some inferential rules may derive from subjects'
self-schemata, others may derive from their conceptions of ex-
403
404
ARIE W. KRUGLANSKI
JUDGMENTAL ACCURACY
tors that induce the perception of accuracy. This topic is addressed in the following section.
Phenomenal Paradigm
In the phenomenal approach to accuracy, the comparison of
a judgment with a criterion is carried out by the subject. A subject's perception of accuracy is likely to be affected not only by
(a) what she or he perceives is the case (subject's criterion) but
also by (b) what she or he perceives is a given target judgment
(e.g., someone else's opinion) and by (c) his or her perception of
the degree of correspondence between the two.
In other words, the accuracy judgment differs from other possible judgments in its tripartite structure that includes the target
judgment, the criterion, and their interrelation. In this sense,
accuracy perception may be thought of as a compound judgment or metajudgment. However, judgment qua accuracy perception should be affected by the same factors that govern the
judgmental process at large. If so, one's conception of the judgmental process should contain useful guidelines for the study of
phenomenal accuracy.
Rule for accuracy inference. The simple rule most people
probably use to determine the accuracy of a judgment is this:
A target judgment that corresponds to the criterion is deemed
correct, and one that deviates from it to some appreciable degree is deemed incorrect. Thus, in order for the accuracy inference to be made, it is first necessary that the target and criterion
judgments be formed and be available in memory. Second, both
judgments should be jointly accessed and compared with each
other. Probably, such a comparison would not be carried out
without at least some degree of interest in its outcome, that is,
a given degree of motivation to make an accuracy assessment.
Finally, in order to make an inference of accuracy or inaccuracy, it is essential that the evidence (for accuracy or inaccuracy) be appropriately interpreted. Those issues are considered
here in turn.
Target and criterial judgments. Both target and criterial
judgments are assumed to form in accordance with the general
judgmental process described earlier, although the contents of
those judgments may often differ. For instance, the target judgment may be formed in a different context from that of the criterion. In turn, contextual differences may activate divergent
cognitive categories, ultimately resulting in judgments of
different contents (Higgins & Stangor, 1988). Thus, a target
judgment might involve ascribing a political opinion to oneself
as a young student in the 1960s. If the context associated with
such an opinion was encoded as "a militant university campus,"
the constructs this might activate could prompt an ascription
to oneself of a strong antiestablishment sentiment. This may or
may not correspond to one's current views of the 1960s establishment (i.e., one's criterion) and may or may not contribute
to the assessment of one's prior opinions as correct or incorrect.
Alternatively, the target judgment may involve an opinion ascribed to another person. To the extent that the other person
was identified by some category label (e.g., a woman, a physicist,
an Englishman, or a professor), this could activate stereotypic
constructs that may influence the target judgment. For example, a professor label might contribute to the perception that the
405
406
ARIE W. KRUGLANSKI
the likelihood that the two will be compared. It seems less probable that the reverse process will occur. Thus, thinking about
one's own opinion on an issue (the criterion) is less likely to
bring to mind alternative target judgments on this problem
(e.g., other people's opinions): Such judgments may not be
strongly connected in the individual's mind to the topic as such;
hence they are less likely to be primed by its mere consideration.
Motivational bases of comparison. Occurrence of a comparison between judgment and criterion may be facilitated or inhibited by the same epistemic motivations that affect the formation
of target or criterion judgments. Thus, an individual may have
a high need for closure regarding a judgment's accuracy, and
this may increase the likelihood of the comparison. In contrast,
a person motivated to avoid closure on the accuracy issue would
be less likely to carry out the comparison. Finally, persons may
have needs for specific closure pertaining to accuracy assessment, and those may appropriately affect the comparison tendency: An individual may wish for a specific comparison outcome, for example, proving oneself correct; expectation of such
an outcome may enhance the likelihood of a comparison,
whereas expectation of the opposite outcome may reduce this
likelihood.
Perception of correspondence or discrepancy. The comparison outcome as such, that is, the perception of correspondence
or the discrepancy between target and criterion, may also be
influenced by various cognitive and motivational determinants
of judgments. For instance, an individual may strongly expect
to agree with another person, for example, a longtime friend
known to generally share his or her attitudes and opinions. Similarly, a person's self-concept may suggest that his or her judgments would be consistent over time. Both expectations may
raise the individual's perceptual threshold for discrepancies and
hence lower his or her tendency to judge that an error has occurred. Opposite expectancies, notably of a disagreement with
another person or with one's own prior judgment, may lower
the threshold for perceived discrepancies and increase the tendency to recognize errors.
The perceptual threshold for discrepancies may also be influenced by motivational factors. In some situations, one's
wishes might dispose a perception of discrepancy between one's
own views and those of another person. Possibly, perceptions of
media bias by subjects in the Vallone et al. (1985) research were
motivated by a desire to sway the media to their own side on
future occasions. Similarly, the perception of one's own past
errors could be motivated by the desire to view oneself as having
grown wiser or as having "seen the light" (Ross & McFarland,
1988). The various cognitive and motivational determinants of
perceived discrepancies could be profitably explored in future
research on phenomenal accuracy.
Interpreting the comparison outcome." Determination of accuracy or error. Although inconsistency between target and criterion judgments is compatible with the notion that the former is
erroneous, alternative interpretations are possible and may be
adopted under the appropriate conditions. One such alternative
interpretation is that the criterial (rather than the target) judgment is mistaken; another is that both judgments are accurate,
albeit in different situations. I examine these two alternatives in
turn.
Consequences of phenomenal accuracy for interpersonal attitudes and actions. Phenomenal accuracy could be highly relevant to social attitudes and behavior. Common experience suggests that the conviction of being right (self-righteousness) may
often foster a feeling of superiority over others who happen to
disagree with one's judgments. Occasionally, this may encourage a derogation of those persons and may legitimize punitive
acts designed to correct their distorted views. For instance, in
Milgram's (1974) obedience research, subjects' compliance
with the request to administer painful shocks to learners could
at least partially be due to the cover story that the punishment
is meant to improve the accuracy of learners' performance.
Of interest is what factors may mediate persons' interpersonal
reactions to perceived accuracy of self and others. Two general
classes of such factors are possibly involved: (a) subjects' confidence in their accuracy assessments and (b) the degree to which
they regard accuracy as an important value. Both factors may
interactively determine persons' interpersonal reactions to accurate or inaccurate others. Presumably, persons who both attach high value to accuracy and have high confidence in their
accuracy assessments would react more extremely as a function
of other persons' perceived accuracy than would persons who
either place low value on accuracy or have low confidence in
JUDGMENTAL ACCURACY
their accuracy assessment. In other words, highly confident persons who value accuracy might be less tolerant of persons whose
views differ from their own and more respectful of persons
whose views they accept.
Research on the authoritarian personality (Adorno, FrenkelBrunswick, Levinson, & Sanford, 1950) is generally consistent
with those predictions. Specifically, high authoritarians have
been distinguished by their strong views and convictions (indicative of high confidence), negative attitudes toward minorities
(suggesting intolerance of divergent perspectives), adulation of
high-status persons (presumably considered right), and disrespect of low-status persons (suspect of being wrong). Further
research could study reactions to putatively correct or incorrect
others as a function of subjects' situationally induced confidence and the situational value of accuracy.
Summary: Accuracy research in phenomenal paradigm. The
phenomenal paradigm focuses on factors affecting the perception of accuracy. Perceived accuracy is conceptualized as a
metajudgment involving (a) the target judgment, (b) the criterial judgment, and (c) the judgment of correspondence between target and criterion judgments. Beyond its unique tripartite structure, the accuracy metajudgment is assumed to be governed by cognitive and motivational determinants of judgments
at large. Specifically, for accuracy to be assessed, target and criterion judgments need to be available in the individual's memory and be jointly accessible. Furthermore, target and criterion
judgments need to be compared with each other, and the comparison outcome needs to be appropriately perceived and interpreted. All of those events depend on the accessibility of appropriate cognitive categories and on the appropriate motivational
conditions. In general, then, cognitive and motivational determinants of the judgmental process at large seem highly relevant
to the study of accuracy in the phenomenal paradigm.
Recapitulation and Conclusion
This article considers recent conceptions of accuracy in social perception and cognition. From a review of several theoretical treatments of the topic, I have concluded that several major
questions typically posed about human accuracy have not been
satisfactorily resolved, nor do they seem likely to be resolvable
in the foreseeable future. It has not been possible to meaningfully answer whether people are generally accurate or inaccurate
or to reliably estimate the proportion of times at which they
may be. Furthermore, it has not been possible to delineate general classes of boundary conditions for accurate and inaccurate
judgments or to elucidate the process (or method) for reaching
accuracy.
Instead of coping with general questions about human accuracy, a situationally specific approach to the issue has been outlined. Such an approach adopts the correspondence definition
of accuracy and assumes that the criterion for accuracy represents the standard setter's judgment as to the true state of affairs.
It follows that the criterion is relative to the situation and to the
standard setter's perceptions rather than being absolute in any
strong sense of the term.
The present interpretation suggests that, despite historical
separation, a study of the judgmental process is highly relevant
407
to the study of accuracy. According to this analysis, determination of accuracy constitutes a metajudgment comprising three
components: the target judgment, the criterion, and the correspondence between them. Each of those components is itself a
judgment governed by the cognitive and motivational factors
that affect judgments in general.
This conception suggests two general paradigms for accuracy-related research: In the realist approach the situational criterion (the reality) is assumed to be known, and the research addresses the conditions under which subjects' judgments approximate the criterion. In the phenomenal approach, the focus is
on subjects' compound perception comprising the judgment,
the criterion, and their correspondence. Both research paradigms contain numerous possibilities for further accuracy research. By systematically outlining their judgmental underpinnings, this analysis may help extend the study of accuracy in
previously unexplored directions.
References
Abelson, R. P. (1959). Models of resolution of beliefdilemmas. Journal
of Conflict Resolution, 3, 343-352.
Abramson, L. Y., & Alloy, L. B. (1981). Depression, nondepression,
and cognitive illusion: Reply to Schwartz. Journal of Experimental
Psychology: General, 110, 436-447.
Adorno, T. W., Frenkel-Brunswick; E., Levinson, D. J., & Sanfordp
R. N. i1950). The authoritarian personality: Studies in prejudice.
New York: Harper & Row.
Ajzen, I. (1977). Intuitive theories of events and the effects of base rate
information on prediction. Journal of Personality and Social Psychology, 35, 303-314.
Alloy, L. B., & Abramson, L. Y. (in press). Depressive realism: Four
theoretical perspectives. In L. B. Alloy (Ed.), Cognitive processes in
depression. New York: Guilford Press.
Amabile, T. M. (1983). The social psychology of creativity. New York:
Springer-Verlag.
Baron, R. (1988). An ecological framework for establishing a dual mode
theory of social knowing. In D. Bar-Tal & A. W. Kruglanski (Eds.),
The social psychology of knowledge. Cambridge, England: Cambridge University Press.
Bar-Tal, D. (1986). The Masada syndrome: A case of central belief. In
N. Milgram (Ed.), Stress and coping in time of war. New York: Brunner/Mazel.
Bechtold, J., Zanna, M. P., & Nacarrato, M. (1985). Unpublished data.
Waterloo University, Waterloo, Ontario, Canada.
Beck, A. T., Rush, A. J., Shaw, B. E, & Emery, G. (1979). Cognitive
therapy of depression. New York: Guilford Press.
Bern, D. J. (1972). Self perception theory. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 6). New York: Academic Press.
Bruner, J. S. (1973). Beyond the information given: Studies in the psychology of knowing. New York: Norton.
Campbell, D. T. (1969). Prospective: Artifact and control. In R. Rosenthai & R. L. Rosnow (Eds.), Artifact in behavioral research. New
York: Academic Press.
Cantor, N., & Kihlstrom, J. F. (1987). Personality and social intelligence. Englewood Cliffs, NJ: Prentice-Hall.
Chaiken, S., & Stangor, C. (1987). Attitudes and attitude change. Annual
Review of Psychology, 38, 575-630.
Cronbach, L. J. (1955). Processes affecting scores on "understanding of
others" and "assumed similarity.'" Psychological Bulletin, 52, 177193.
408
ARIE W. KRUGLANSKI
409
JUDGMENTAL ACCURACY