314 Attitudes in Social Behavior
accessibility. Specifically, attitudes tended to interfere with
participants’ ability to perceive change in an attitude target,
and this effect was stronger for highly accessible attitudes than
for less accessible attitudes. In another set of studies, Stewart,
Vassar, Sanchez, and David (2000) showed that participants’
attitudes toward women’s and men’s societal roles influenced
whether they individuated male or female targets more: Indi-
viduals with traditional sex-role attitudes individuated male
targets more than they did female targets, whereas individuals
with nontraditional sex-role attitudes individuated female tar-
gets more than they did male targets.
Selective Memory
Attitudes have long been thought to influence memory and
learning of attitude-related information. A variety of processes
could contribute to selective memory, including paying more
attention to attitudinally consistent information (but see
Roberts, 1985), finding it easier to store attitudinally consistent
information, and finding it easier to retrieve attitudinally con-
sistent information from memory. Early studies (e.g., Levine &
Murphy, 1943) indicated that individuals learned and recalled
information that was consistent with their attitudes better
than they did information that was inconsistent with their atti-
tudes. Subsequent researchers, however, had difficulty obtain-
ing significant selective memory effects and questioned
the reliability of the phenomenon (e.g., Greenwald &
Sakumura, 1967).
In a comprehensive and detailed review and meta-analysis
of research on attitude-memory effects, Eagly, Chen,
Chaiken, and Shaw-Barnes (1999) concluded that the hypoth-
esized attitude congeniality effect (i.e., information congenial
with one’s attitudes is more memorable than is uncongenial
information) has been small in magnitude and inconsistent
across studies. Especially worrisome was evidence that the
effect has grown weaker in more recent experiments (com-
pared to earlier experiments), because the recent studies have
generally used more rigorous methods. It appears that selec-
tive memory may be a phenomenon weaker than selective
attention and selective perception.
Perhaps the clearest evidence of selective memory has
been obtained in studies testing whether individuals use their
attitudes as clues for searching memory (i.e., studies specifi-
cally testing selective search and retrieval effects, as opposed
to selective learning and memory in general). Ross (1989)
reviewed a number of studies showing that people used their
attitudes as clues for searching memory, reconstructing past
events, or both. For example, Ross, McFarland, and Fletcher
(1981) exposed respondents to one of two messages that had
previously been shown to have reliable persuasive effects in
opposite directions. In an apparently separate study, respon-
dents exposed to the persuasive message provided reports of
the frequency with which they had performed a number of
behaviors in the past month, including some behaviors
related to the target of the persuasive message. Respondents
reported more frequent behaviors consistent with the attitude
promoted in their message than with the attitude promoted in
the opposing message. Presumably, respondents used their
newly formed attitudes to search their memories and to re-
construct their behaviors in the previous month.Attitude PolarizationAttitudes guide information processing in another way—
namely, they guide spontaneous thinking about the attitude
object. Tesser (1978) showed that simply thinking about an
attitude object tended to polarize the evaluation even in the
absence of any new information. For example, simply think-
ing about a person who was either likable or unlikable led to
stronger evaluations (positive for the likable target, negative
for the unlikable partner) than did a control condition in which
participants performed a distracting task. Presumably, the ex-
isting attitude led participants to generate thoughts that were
consistent with it. This interpretation is supported by findings
that polarization effects are stronger when the individual is
knowledgeable about the attitude object and when the exist-
ing attitude is high in evaluative-cognitive consistency (see
Chaiken & Yates, 1985).ATTITUDES AND BEHAVIORWe discussed earlier how attitudes fulfill various functions
for individuals, including the rapid appraisal of attitude ob-
jects (object-appraisal function), the approach of rewarding
objects and the avoidance of punishing objects (utilitarian
function), the expression of underlying values and identity
(value-expressive function), and so on. All of these hypoth-
esized functions are predicated in part on the assumption
that individuals behave in ways that are consistent with their
attitudes—in other words, on the assumption that attitudes
influence action. In this final section, we review some of the
literature on attitude-behavior consistency.
The hypothesized strong relation between attitudes and be-
havior has sometimes proven difficult to document. For exam-
ple, Wicker (1969) reviewed 30 studies that examined
attitude-behavior consistency and concluded that there was
“little evidence to support the postulated existence of stable,
underlying attitudes within the individual which influence both
his verbal expressions and his actions” (p. 75). Fortunately,