Factors that Affect Time Perception
Several factors have been shown to affect perceived
durations. Three variables are particularly relevant to the
present study. These are: (a) attention and amount of information
processing (b) arousal, and (c) affective valence.
Attention and amount of information processing.
Attentional models of time perception have proposed that,
during a time judgment task, attentional resources allocated
to the stimulus are subtracted from the attention that individuals
devote to the processing of time. As a consequence,
when an interesting stimulus or a stimulus that requires more
attentional resources is presented during the interval to be
Copyright 1997 Psychonomic Society, Inc 972
TIME PERCEPTION AND EMOTIONS 973
estimated, fewer time units are processed and the individual
tends to underestimate the temporal intervals (see, e.g.,
Thomas & Weaver, 1975; Treisman, 1963; Zakay,1992,
1993a, 1993b; Zakay & Yehoshua, 1989).
In studies designed to verify the influence of attention
on time perception, attentional variables have often been
confused with arousal or emotional valence. For instance,
Hawkins and Tedford (1976) had subjects listen to different
prose passages recorded on tape. Subjects judged interesting
tapes to be shorter in duration than non-interesting
ones, in agreement with the predictions of attentional models
of time perception. However, “interesting” tapes had a
sexual content and therefore also involved higher levels of
arousal as well as a positive valence.
Several measures of information processing of the
stimulus have been studied in relation to time perception,
such as interest, processing times, originality, complexity,
and attention. It is always important to distinguish between
the difficulty of a task and the complexity of a stimulus,
because these factors refer to different cognitive sets. The
former are related to active-productive processing whereas
the latter are related to passive-perceptual processing. The
amount of cognitive processing during passive tasks cannot
be directly measured (Zakay, 1993a; Wallace & Rabin,
1960) since it depends on internal mental activity that does
not require an overt response. A classic measure of the
amount of attentional resources allocated to a stimulus (the
primary stimulus) is the delay to react to a secondary probe
stimulus (dual task paradigm). This is an indirect but precise
index, based on the assumption that the amount of
attentional resources devoted to the secondary stimulus is
inversely proportional to the amount devoted to the primary
stimulus. However, performance in the primary task
is affected by the presence of a secondary task. If the task
of interest is the evaluation of a time interval, then the introduction
of an interfering task within the same period will
create a strong confounding factor.
An alternative approach is to examining attnetional
attentional variables is to measure non invasive indices of
attention, such as the heart rate changes, which do not require
the introduction of an additional task.
Several studies have shown that activation (as operationally
defined by the authors) can affect time evaluation.
Activation levels have been manipulated in a number of
ways, such as increase or decrease of body temperature (Fox,
Bradbury & Hampton, 1967), manipulation of circadian
rhythms and administration of drugs (Gupta & Cummings,
1986). Incrementing physical activation levels has been
found to cause an overestimation of perceived time. However,
Curton and Lordahl (1974) found that different activating
methods (physical activity vs. threat of shock) have
different effects on time perception.
In several time estimation experiments, valence manipulation
was performed in waiting conditions (Curton &
Lordahl, 1974; Edmonds, Cahoon, & Bridges, 1981; Hare,
1963; Schiff & Thayer, 1968). Block, George, and Reed
(1980) showed that waiting condition per se affects time
judgments. Fraisse (1963) proposed that time estimates in
waiting conditions might be independent of the valence of
the stimulus and be associated uniquely to an enhanced attention
to time. Therefore, it seems that these studies were
more concerned with the expectancy condition than with
valence manipulation. Indeed, during waiting it is not possible
to control subjects’ cognitive and emotional activity
because they are not engaged in any task.
Langer, Wapner and Werner (1961) found shorter time
estimates for subjects experiencing fear of danger than for
subjects in a neutral condition. However, their study is seriously
flawed by the fact that the pressure on the button
that was used to estimate the period of time also caused the
danger to disappear. Thayer and Schiff (1975) asked subjects
to estimate the time interval spent in an eye-contact
social task and manipulated the facial expression of the
subject’s partner, a critical factor that can influence the affective
valence of an interpersonal contact. Time estimates
where longer when combined with a negative-unpleasant
(scowling-angry) than with a positive-pleasant (smilingfriendly)
facial expression.
Watts and Sharrock (1984) showed that spider-phobic
subjects gave longer estimates of a short interval spent observing
a spider than did nonphobic controls. In this study
as well, no attempt was made to control for arousal levels.
Gupta and Cummings (1989) found that events that seem
to occur quickly are perceived as more pleasant (positive)
than those that seems to occur more slowly. In this study,
time perception was the independent variable and was manipulated
through changes in either physical activation (by
controlling circadian rhythms and the use of caffeine) or
the presentation of an external ticker.
In summary, the literature offers some evidence that
negative stimuli cause an overestimation of the time spent
attending to them, and positive conditions cause an underestimation.
However, studies manipulating more systematically
emotional factors are still lacking.