66 Mood, Cognition, and Memory
and thus they facilitate affect infusion.Heuristic processingis
most likely when the task is simple, familiar, of little personal
relevance, and cognitive capacity is limited and there are no
motivational or situational pressures for more detailed pro-
cessing. This is the kind of superficial, quick processing style
people are likely adopt when they are asked to respond to un-
expected questions in a telephone survey (Schwarz & Clore,
1983) or are asked to reply to a street survey (Forgas &
Moylan, 1987). Heuristic processing can lead to affect infu-
sion as long as people rely on affect as a simple inferential cue
and depend on the “how do I feel about it” heuristic to produce
a response (Clore et al., 2001; Schwarz & Clore, 1988).
When simpler strategies such as direct access or motivated
processing prove inadequate, people need to engage in sub-
stantive processingto satisfy the demands of the task at hand.
Substantive processing requires individuals to select and in-
terpret novel information and relate this information to their
preexisting, memory-based knowledge structures in order to
compute and produce a response. This is the kind of strategy
an individual might apply when thinking about interpersonal
conflicts or when deciding how to make a problematic re-
quest (Forgas, 1994, 1999a, 1999b).
Substantive processing should be adopted when (a) the
task is in some ways demanding, atypical, complex, novel, or
personally relevant; (b) there are no direct-access responses
available; (c) there are no clear motivational goals to guide
processing; and (d) adequate time and other processing re-
sources are available. Substantive processing is an inherently
open and constructive strategy, and affect may selectively
prime or enhance the accessibility of related thoughts, mem-
ories, and interpretations. The AIM makes the interesting and
counterintuitive prediction that affect infusion—and hence
mood congruence—should be increased when extensive and
elaborate processing is required to deal with a more complex,
demanding, or novel task. This prediction has been borne out
by several studies that we will soon review.
The AIM also specifies a range of contextual variables
related to the task,theperson,and the situationthat jointly
influence processing choices. For example, greater task fa-
miliarity, complexity, and typicality should recruit more sub-
stantive processing. Personal characteristics that influence
processing style include motivation, cognitive capacity, and
personality traits such as self-esteem (Rusting, 2001; Smith &
Petty, 1995). Situational factors that influence processing
style include social norms, public scrutiny, and social influ-
ence by others (Forgas, 1990).
An important feature of the AIM is that it recognizes that
affect itself can also influence processing choices. As noted
earlier, both Bless (2000) and Fiedler (2000) have proposed
that positive affect typically generates a more top-down,
schema-driven processing style whereby new information is
assimilated into what is already known. In contrast, negative
affect often promotes a more piecemeal, bottom-up process-
ing strategy in which attention to external events dominates
over existing stored knowledge.
The key prediction of the AIM is the absenceof affect in-
fusion when direct access or motivated processing is used,
and the presenceof affect infusion during heuristic and sub-
stantive processing. The implications of this model have now
been supported in a number of the experiments considered in
following sections.Evidence Relating Processing Strategies
to Mood CongruenceThis section will review a number of empirical studies that il-
lustrate the multiple roles of affect in cognition, focusing on
several substantive areas in which mood congruence has
been demonstrated, including affective influences on learn-
ing, memory, perceptions, judgments, and inferences.Mood Congruence in Attention and LearningMany everyday cognitive tasks are performed under condi-
tions of considerable information overload, when people
need to select a small sample of information for further pro-
cessing. Affect may have a significant influence on what peo-
ple will pay attention to and learn (Niedenthal & Setterlund,
1994). Due to the selective activation of an affect-related
associative base, mood-congruent information may receive
greater attention and be processed more extensively than af-
fectively neutral or incongruent information (Bower, 1981).
Several experiments have demonstrated that people spend
longer reading mood-congruent material, linking it into a
richer network of primed associations, and as a result, they
are better able to remember such information (see Bower &
Forgas, 2000).
These effects occur because “concepts, words, themes,
and rules of inference that are associated with that emotion
will become primed and highly available for use...[in]...
top-down or expectation-driven processing...[acting]...
as interpretive filters of reality” (Bower, 1983, p. 395). Thus,
there is a tendency for people to process mood-congruent
material more deeply, with greater associative elaboration, and
thus learn it better. Consistent with this notion, depressed psy-
chiatric patients tend to show better learning and memory for
depressive words (Watkins, Mathews, Williamson, & Fuller,
1992), a bias that disappears once the depressive episode is
over (Bradley & Mathews, 1983). However, mood-congruent
learning is seldom seen in patients suffering from anxiety