Affect, risk perception and future optimism after the tsunami disaster

Judgment and Decision Making, Vol. 3, No. 1, January 2008, pp. 64–72.
Affect, risk perception and future optimism after the tsunami
disaster
Daniel Västfjäll1,2? Ellen Peters1,3 and Paul Slovic1,3
1 Decision Research
2 Göteborg University
3 University of Oregon
Abstract
Environmental events such as natural disasters may in?uence the public’s affective reactions and decisions. Shortly
after the 2004 Tsunami disaster we assessed how affect elicited by thinking about this disaster in?uenced risk perceptions
and future time perspective in Swedish undergraduates not directly affected by the disaster. An experimental manipu-
lation was used to increase the salience of affect associated with the disaster. In Study 1 we found that participants
reminded about the tsunami had a sense that their life was more ?nite and included fewer opportunities than participants
in the control condition (not reminded about the tsunami). In Study 2 we found similar effects for risk perceptions. In
addition, we showed that manipulations of ease-of-thought in?uenced the extent to which affect in?uenced these risk
perceptions, with greater ease of thoughts being associated with greater perceived risks.
Keywords: affect, risk perception, disaster.
1 Introduction
are constructed on the basis of various contextual factors
(such as incidental affect or mood; Lichtenstein & Slovic,
Major societal events such as natural disasters and terror-
2006; Peters, 2006; Johnson & Tversky, 1983) and peo-
ist attacks in?uence our thoughts and feelings. In the face
ple tend to rely on their affective reactions when making
of a major environmental event, many people tend to react
decisions (Slovic et al., 2002; see also P?ster & Böhm,
with emotion and emotion-laden decisions (Lerner et al.,
this issue).
2003). The 2004 East Asian tsunami disaster had a pro-
Affect is de?ned here as the speci?c quality of good-
found psychological impact on many countries, not only
ness or badness experienced as a feeling state (with or
those that were directly hit by the tsunami waves. Sweden
without awareness) and demarcating a positive or nega-
(pop. 9 million) had an unusually high number of tourists
tive quality of a stimulus.1 Reliance on such feelings in
visiting the area at the time of the disaster, resulting in
judgment and decision making has earlier been described
over six hundred Swedes being killed or missing. The
as an affect heuristic (Finucane et al., 2000; Slovic et
Tsunami disaster was therefore considered a major na-
al., 2002). Most previous research on affect and decision
tional tragedy in Sweden (Grandien, Nord, & Strömbäck,
making has focused on integral affect (affect attached to
2005). A consequence of this tragedy, and the media at-
mental representations of objects; Slovic et al., 2002).
tention it received (Mann, 2007), was that many Swedes
However, in many judgments other sources of affect are
felt deeply involved and saddened (Grandien et al., 2005).
also present. A large number of studies shows that affec-
The feelings elicited by such an event may also have an
tive states that are unrelated to the judgmental target in-
impact on everyday decisions. Previous research in judg-
?uence judgments and decisions nonetheless (Isen, 1997;
ment and decision making has shown that preferences
1Affective responses can occur rapidly and automatically, and may
be elicited by stimulus properties, physical stimulation, perception of
?The research reported in this paper was supported by National
one´s immediate environment, thoughts and memories, or propriocep-
Science Foundation; contract/grant numbers: SES–0517770; SES–
tive cues (Schwarz & Clore, 2007). Mood, one form of affect, is a
0526020; SES–0339204; SES–0241313.
We thank Gisela Böhm,
relatively stable and mild affective state that does not have a speci?c fo-
Wibecke Brun, Jon Baron, Rolf Reber and three anonymous review-
cal object (Morris, 1999), whereas emotions, another form of affect, are
ers for helpful comments on an earlier version of this manuscript.
more intense and are of shorter duration. Incidental affect is an affective
Address: Daniel Västfjäll, Department of Psychology, Göteborg Uni-
state, such as a mood state, brought about by environmental or intrinsic
versity, PO Box 500, SE 40530 Gothenburg, Sweden.
Email:
stimulation. Integral affect, on the other hand, is elicited by perceiving
daniel.vastfjall@psy.gu.se.
the target or a mental representation of the target.
64

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
65
Schwarz & Clore, 1983). In a famous example, Johnson
may further modulate the impact of affect on judgments.
and Tversky (1983) found that incidental affect (i.e., a
Supporting this, a study by Lerner and Gonzalez (2005)
mood state) induced by reading a newspaper article in?u-
showed that ?uency manipulations in?uenced the effect
enced subsequent risk judgments.
of speci?c emotions on risk perception.
However, the effect of incidental affect on judgments is
The conceptual model guiding this research can be de-
not a stable, unchangeable or unavoidable fact; it should
scribed in the following way: 1) We expect that major
rather be seen as a constructive process where the indi-
environmental events such as a natural disaster may in-
vidual tries to determine if their affective reactions to a
?uence experienced affect even among individuals not
target are a reliable and relevant source of information
directly affected by the disaster. 2) The experienced af-
(Clore & Huntsinger, 2007). At the core of this argu-
fect will, in turn, impact various affective and cogni-
ment lies the notion that when asked to make an eval-
tive judgments. 3) The effects of affect incidental to the
uative judgment, individuals seek information to deter-
judgment task can be diminished by introducing infor-
mine how they should make this judgment. People tend
mation (such as ?uency manipulations) that questions the
to use whatever information is available to them at the
diagnostic value of experienced feelings for judgments.
time of making a decision (Clore & Huntsinger, 2007;
Speci?cally, we tested the prediction that affect elicited
Schwarz, 2004). In the absence of other relevant or more
by thinking about a recent major natural disaster would
salient information, people use their affective reactions
in?uence judgments of well-being (Schwarz & Clore,
to the target to evaluate the object (Pham, 1998; see also
1983) and future pessimistic/optimistic thinking (Wright
deVries, Witteman & Holland, this issue). People in
& Bower, 1992) in a mood-congruent manner (Schwarz
positive moods tend to evaluate objects more favorably
& Clore, 2007). Previous research has documented the
than participants in a negative mood (mood-congruence;
effects of laboratory-induced mood using standardized
Schwarz & Clore, 1983; but see Andrade, 2005 for a dis-
mood induction procedures (autobiographical recall or
cussion about mood-incongruent effects). One important
affect-inducing scenarios). We extended this research by
point here is that people incorrectly attribute their inci-
inducing affect through a procedure in which participants
dental moods as a reaction to the target. This misattri-
were asked to think about a recent and relevant major en-
bution can be corrected or changed by introducing infor-
vironmental disaster.
mation that questions the diagnostic value of the affec-
tive reaction for the judgment. For instance, in Schwarz
Our research strategy compared ratings of affect and
and Clore’s (1983) study participants were given a sim-
ratings of future personal and societal events in two
ple reminder about the cause (sunny vs. cloudy weather)
groups of participants, one reminded about the tsunami
of their moods which resulted in mood no longer in?u-
and a control group. We hypothesized that reminding
encing judgments of well-being. Importantly though, it
participants about the tsunami would elicit negative af-
was the diagnostic value of the affective reaction for the
fect associated with the event. We further anticipated that
judgment task, not the affective reaction itself, that was
this affect would spill over to judgments of well-being as
affected by this manipulation (Schwarz, 2004).
well as optimistic/pessimistic thinking.
Incidental mood is only one of many sources of expe-
In Study 1, participants in both conditions completed
riential information that can be used in judgments. The
a measure of future pessimism (future time perspective
meta-cognitive experience of the ease or ?uency of in-
(FTP) scale, Lang & Carstensen, 2002) and rated their
formation processing has been shown to be an important
well-being (Pavot & Diener, 1993). The FTP scale was
experiential factor informing judgments and decisions
originally developed as an individual difference mea-
(Schwarz & Clore, 2007). In a study on the effect of ?u-
sure of the perceived time remaining in life (Carstensen,
ency on decision making, participants were more likely to
2006). In Study 1, we used the FTP scale as a depen-
defer choice when they generated more reasons for mak-
dent variable and we expected to ?nd that individuals re-
ing the choice (thus decreasing ?uency; Novemsky, Dhar,
minded about the tsunami perceived life as more ?nite
Schwarz & Simonson, 2007). Studies on perceptual ?u-
and limited than participants in a control condition. In
ency (the subjective ease of perceptual processing) have
Study 2, half of the participants in the tsunami-remind
found that if the color in which a statement is printed
condition were given an additional experimental manipu-
makes it easy to read, this can impact the perceived truth-
lation (ease-of-thought-generation; Schwarz, 2004). Par-
fulness of the statement (with an easier-to-read font lead-
ticipants then made risk estimates of various future posi-
ing to a higher probability of endorsing a statement as
tive and negative events (Lerner & Gonzalez, 2005). We
true; Reber & Schwarz, 1999). Thus, as Schwarz (2004,
expected that this manipulation of the ease with which ex-
p. 341) notes, it seems that “the subjective experiences
amples of other disasters comes to mind would in?uence
that accompany our thought processes are informative in
the diagnostic value of feelings for judgments of future
their own right.” Consequently, meta-cognitive feelings
risk, but it does not change the feelings themselves.

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
66
2 Study 1. Future time perspective
Table 1: Means and inferential statistics (df = 102) for
As people grow older they experience time as more lim-
mood and speci?c emotion ratings obtained in the exper-
ited, closed, and ?nite (Carstensen, 2006). However,
imental and control conditions.
chronological age is not the only determinant of how the
Measure Tsunami-remind Control
t
p <
future is perceived. For instance, Fung and Carstensen
(2006) showed that younger adults prioritized emotional
Mood scales
goals, a behavior indicative of a limited time perspective,
Valence
?1.11
0.67
13.20
.01
when facing a major environmental event (i.e., the SARS
Activation
1.33
0.20
9.16
.01
epidemic). Building on these ?ndings, we hypothesized
Emotion scales
that participants reminded about the tsunami would ex-
perience a more limited future time perspective than par-
Sad
2.71
0.99
14.42
.001
ticipants in a control condition. In addition, we expected
Depressed
2.14
1.08
8.11
.001
that participants reminded about the tsunami would ex-
Anxious
2.34
1.02
6.92
.01
perience stronger negative affect and lessened well-being
Afraid
1.99
0.53
9.40
.001
compared to the control group and that this difference
Worried
2.25
0.87
10.65
.001
would account for the hypothesized difference between
conditions.
Angry
2.04
0.46
15.19
.001
2.1 Method and measures
responded to the question “How well does each question
Twenty-eight men and 77 women with a mean age of 25.3
describe you” by circling a number between 1 (not at all)
(SD = 4.1) participated. Data were collected in Sweden
and 7 (very well). The measure was reverse-scored where
during the Spring of 2005, roughly 3–5 months after the
appropriate and averaged across the eight items (Cron-
tsunami disaster. The tsunami was still very actively cov-
bach’s alpha = .77) into a single index; higher values
ered by the media in Sweden during this time (Mann,
indicated a more closed or limited future time perspec-
2007).
tive. In addition, the Pavot and Diener (1993) subjective
To manipulate access to feelings, we used an experi-
well-being scale as well as mood and speci?c-emotion
mental approach resembling the techniques developed by
scales (Västfjäll et al., 2002) were administered. The
Lerner et al. (2003) and studies on affective imagery and
mood scale consisted of six adjective pairs found in pre-
decision making (Slovic, 1995). In a between-groups de-
vious research (Västfjäll et al., 2002) to tap valence and
sign, half of the participants were asked to write down
activation, respectively. Sleepy-awake, dull-peppy, and
the ?rst three images that came to mind when hearing
passive-active were used to de?ne the activation scale,
the word “tsunami”. The other half of the participants
displeased-pleased, sad-glad, and depressed-happy were
(the control condition) were asked to produce images to
used to de?ne the valence scale. Participants were asked
a neutral word (“round”). Pre-testing showed that this
to circle a number (range = ?4 to 0 to +4) that best cor-
priming manipulation made affect2 associated with the
responded with their current feeling. The three adjectives
tsunami disaster salient (Siemer & Reisenzein, 1998) and
tapping each dimension were averaged into two index
provided us with the opportunity to study the relative im-
variables corresponding to valence and activation, respec-
pact of affect in the two conditions.
tively.
After the experimental manipulation, participants re-
The speci?c emotion scale consisted of the adjectives
sponded to a series of questions. To measure future
sad, depressed, anxious, worried, afraid, and angry. Par-
time perspective, a version of the FTP scale (Lang &
ticipants were asked to rate how intensely they felt each
Carstensen, 2002) was used. The measure contains eight
emotion by circling a number on a unipolar scale an-
items: 1. Many opportunities await me in the future. 2.
chored by 0 (not at all) to 6 (very much).
My future is ?lled with possibilities. 3. Most of my life
lies ahead of me. 4. My future seems in?nite to me. 5.
2.2 Results and discussion
There is plenty of time left in my life to make new plans.
6. I have the sense that time is running out. 7. There are
To show that affect indeed was more negative in the
only limited possibilities in my future. 8. As I get older,
tsunami-remind condition than in the control condition,
I begin to experience time as more limited. Participants
the mood and emotion ratings were submitted to a series
of independent t-tests. Negative affect ratings were sig-
2Pre-testing showed that this procedure in general heightened neg-
ative feelings such as sadness, depression, and anxiety and no speci?c
ni?cantly higher in the experimental condition suggesting
emotion was more salient.
that the manipulation was successful (see Table 1).

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
67
Figure 1. Subjective well-being mediated the in?uence
of condition on FTP. Speci?cally: (a) condition predicted
SWL (F(1,103) = 8.59, regression weight = .27, p < .05);
(b) condition predicted FTP (F(1,103) = 82.69, regres-
sion weight = .67, p < .01); (c) Subjective well-being
predicted FTP (F(1,103) = 25.91, regression weight =
Figure 1: Mediation analysis of Condition and Subjective
.45, p < .01); and (d) the condition variable dropped
well-being (SWL) on Future Time Perspective (FTP).
signi?cantly when controlling for Subjective well-being
(F(2,102) = 6.37, regression weight = ?.08, ns.).
Taken together, these ?ndings suggest that thinking
In order to test the primary research hypotheses, that
about a major environmental event such as the Tsunami
well-being should be lower and FTP more limited in the
disaster elicits negative feelings (as indexed by speci?c
tsunami-remind condition than in the control condition,
emotion/mood ratings as well as well-being ratings).
two contrasts were performed. As expected, participants
These feelings in turn in?uence how people think about
reminded about the tsunami rated their overall well-being
and view their future possibilities. Negative feelings lead
as lower (M = 3.81) than participants in the control con-
individuals to be more pessimistic, viewing time as more
dition (M = 4.50), t(103) = 9.09, p < .001. Similarly, FTP
limited and holding fewer possibilities. Although the
was more limited (M = 4.80) in the tsunami-remind con-
FTP scale was originally developed as an individual dif-
dition than in the control condition (M =3.22), t(103) =
ference measure (Lang & Carstensen, 2002), the ?nd-
2.92, p < .001.
ings here suggest that it also can be used as a depen-
The ?nding that the experimental manipulation in-
dent variable measuring group differences. In contrast
creased negative affect which, in turn, decreased over-
to the present ?ndings, previous research has shown that
all subjective well-being is a replication of Schwarz and
the FTP measure is relatively uncorrelated with current
Clore’s ?ndings (1983). The result that future time per-
mood (Lang & Carstensen, 2002). However, these stud-
spective changed with experienced affect is, however, a
ies have not experimentally manipulated moods which
novel demonstration. To further show that affect in?u-
may be why mood effects were not obtained. The FTP
enced FTP judgments, we conducted a mediation anal-
scale is used here as a measure of pessimism and the
ysis. Because both affect and well-being ratings were
link between mood and optimism-pessimism is well doc-
more negative in the experimental condition, we decided
umented in the mood literature (Isen, 1997; Wright &
to test whether well-being (as an overall proxy of affect)
Bower, 1992). Although the ?nding that participants in
mediated the effect of condition on FTP. The choice of
the experimental condition experienced a more limited
well-being ratings as an overall measure of affect was
FTP may be predicted from previous research without
motivated by literature on happiness that suggested that
involving experienced affect as an explanatory variable
current feelings are integrated into more global assess-
(Fung & Carstensen, 2006), we extend this research by
ments of affective well-being (Schwarz & Clore, 1983;
showing that well-being mediates the effect.
Schwarz & Strack, 1999). The subjective well-being
A limited future time perspective may have many detri-
measure however includes aspects other than experienced
mental consequences for different individual behaviors
feelings (e.g., life circumstances; Diener, 1984; Pavot &
such as preference for immediate consumption of food
Diener, 1993) and when used as a variable in a media-
and money at the cost of long-term health behavior and
tion analysis it may be more conservative than ratings of
well-being (Shiv et al., 2005). In Study 2, we study risk
current affect.3 To test mediation, a series of regression
perception in different domains of one’s future life and
models were estimated (Baron & Kenny, 1986). To ex-
also test potential measures to counteract the negative im-
amine the degree of mediation, we ?rst regressed subjec-
pact of feelings on judgments.
tive well-being on the condition variable, then regressed
FTP on the condition variable, and ?nally regressed FTP
on both the condition variable and subjective-well being.
3 Study 2. Future life expectations
The degree to which the in?uence of condition on FTP is
reduced when accounting for the in?uence of subjective
and ease-of-thought
well-being expresses the degree of mediation.
The results of these regression analyses are depicted in
In Study 2, future pessimism was assessed by obtaining
risk estimates of future events across different decision
3Analyses with either the valence mood index or the composite spe-
domains. As in Study 1, a between-groups comparison
ci?c emotion index as mediating variables yielded comparable results
to the well-being ratings. Subjective well being had strong correlates
was used (remind about the tsunami vs. a control group).
with these two indices (r = .71 and .68, respectively, p < .01).
In addition, we assessed whether the ease with which

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
68
thoughts about other natural disasters comes to mind may
ters that occurred anywhere in the world during the last
modulate the impact of feelings on risk perception. Pre-
100 years.4 After the experimental manipulation, partici-
vious research has demonstrated that the ease or ?uency
pants responded to a series of questions. To measure risk
of thoughts and feelings determines the impact of those
perception, we used a modi?ed version of the scale devel-
thoughts and feelings on judgment (Schwarz, 2004). For
oped by Lerner and Gonzalez (2005). Participants indi-
example, participants asked to generate eight examples of
cated from 1 (extremely unlikely) to 7 (extremely likely)
behaviors that increased the risk of heart disease (a rela-
the likelihood that each of 15 events would happen to
tively dif?cult task) were more likely to report that they
them at any point in their future life. This measure is
were invulnerable to heart-disease problems than partic-
thus similar to the FTP scale in that it taps future pes-
ipants asked to generate three examples (an easy task;
simism. However, rather than asking speci?c questions
Rothman & Schwarz, 1998). The explanation for this
about the remaining time in life and the possibility of
result is that participants in the generate-eight condition
changing one’s circumstances, the risk perception scale
noticed how dif?cult it was to think of examples and, on
asks participants to judge the likelihood of various posi-
the basis of that dif?culty, thought that they must be rela-
tive and negative events in different life domains (social,
tively invulnerable to heart disease.
health, ?nancial, recreational). The 15 items were: 1. I
Building on this logic, we asked participants to list a
enjoyed my job. 2. I had a heart attack before age 50. 3.
more dif?cult six (versus an easier two) examples of ma-
My achievements were written up in a newspaper. 4. I
jor natural disasters during the last one hundred years. We
chose the wrong career. 5. I married someone wealthy.
expected that participants in the dif?cult condition should
6. I received recognition in my profession. 7. I could not
notice that such events are very rare and thus experience
?nd a job for 6 months. 8. My income doubled within
less con?dence in their feelings about the tsunami disas-
10 years after my ?rst job. 9. I developed gum problems
ter as a basis for risk judgments. In other words, the dif?-
in my mouth. 10. I did something in a job interview that
cult ease-of-thought generation should render incidental
made me embarrassed. 11. I said something idiotic in
affect from the tsunami disaster relatively less diagnostic
front of my class mates. 12. I got lost at night for more
for judgments (Pham, 1998).
than 15 minutes. 13. I was on an airplane that encoun-
The design of Study 2 closely resembles Lerner and
tered severe turbulence. 14. I received favorable medical
Gonzalez’s (2005) Study 1, but with one important dif-
tests at age 60. 15. I encountered a dangerous snake while
ference: Rather than studying the effects of speci?c emo-
on vacation.
tions we focus on the effects of generalized moods on
The measure was averaged (with reverse scoring for
risk perception. We predict (in line with Lerner and Gon-
the appropriate items) across the 15 items into an overall
zalez’s ?ndings for speci?c emotions) that an ease-of-
pessimistic future risk index (Cronbach’s alpha = .89).
thought manipulation will interact with the effects of inci-
In addition, the Pavot and Diener (1993) subjective well-
dental affect on judgments, effectively debiasing risk esti-
being scale used in Study 1 was administered.5
mates in the hard (list many) condition but not in the easy
Overall, we expected that participants in the tsunami-
(list few). However, since the affect induced by think-
remind condition who were asked to list few natu-
ing about the tsunami does not directly rely on the ease
ral disasters (easy condition) would give more pes-
with which one can list six vs. two other major disas-
simistic risk estimates than participants in the control and
ters, it may be predicted that well-being ratings will be
tsunami-remind dif?cult (list many natural disasters) con-
relatively untouched by this manipulation. Thus, we ex-
dition. Furthermore, we expected that participants in the
pect that one experimental manipulation (tsunami-remind
tsunami-remind conditions would report an overall lower
vs. control) will in?uence both well-being and risk per-
well-being than participants in the control condition, in-
ception, whereas the second experimental manipulation
dependent of the ease-of-thought manipulation.
(ease-of-thought) will only in?uence risk perception.
3.2 Results and discussion
3.1 Method and measures
To test the hypotheses, three contrasts were performed for
Fifty men and 75 women with a mean age of 27.1 (SD
both the subjective well-being ratings and the risk esti-
= 6.2) participated. The study was run in Sweden 3–4
4The ease-of-thought manipulation was only used in the tsunami-
months after the tsunami disaster.
remind conditions since our hypothesis pertains to the debiasing effect
In addition to the tsunami-reminder manipulation used
of this manipulation in this experimental condition.
in Study 1, half of the participants in the remind condi-
5Mood and discrete emotion scales were however not included in
tion were given an “ease-of-thought-generation” manip-
this study. We chose to only include well-being as a measure of affect
since both pre-studies and Study 1 consistently showed that negative
ulation (Schwarz, 2004) in which they were asked to list
affect ratings discriminated between the two conditions and, further,
either six (hard) or two (easy) other major natural disas-
that these ratings co-varied with well-being ratings (see footnote 3).

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
69
mates. As expected, participants in the control condition
than participants in the hard condition, this feeling did
reported signi?cantly higher well-being (M = 4.86) than
not impact the judgment task. If the positive affect asso-
participants in the tsunami-remind dif?cult (M = 3.95,
ciated with ?uency had spilled over to the risk estimates,
t(74) = 3:30, p < .05) and the tsunami-remind easy (M
we would expect that participants in the easy condition
=3.82, t(82) = 3.79, p < .05) conditions. The dif?cult and
would have rated the risk of future negative outcomes
easy conditions did not differ (t(88) = 0.43, ns).
lower than participants in the hard condition. However,
For the risk estimates, participants in the control condi-
the reverse pattern was found, suggesting that the pos-
tion reported less pessimistic estimates (M = ?.19) than
itive affect associated with ?uency experiences neither
participants in the tsunami-remind easy condition (M =
changed the negative affect elicited by the experimental
?.96, t(74) =5.18, p < .01), but similar estimates to the
manipulation, nor did it in?uence the judgment task.
participants in the tsunami-remind dif?cult (M = ?.22,
t(73) = - 0.13, ns). Further, risk estimates in the easy con-
dition were signi?cantly higher than those in the dif?cult
4 General Discussion
condition (t(87) = 4.75, p < .01).
Taken together, these ?ndings suggest that thinking
The results of these studies suggest that the negative af-
about the tsunami decreased perceived well-being and
fect elicited by thinking about a recent major natural
systematically biased risk estimates. In addition, the ef-
disaster leads to a more pessimistic view of the future.
fect on risk estimates was modulated by the ease with
Participants reminded about the recent tsunami disaster
which participants could list few versus many natural
felt that their life had fewer possibilities and that time
disasters. Combining the logic of research on ?uency
was limited (Study 1) and that the risk of future self-
(Schwarz, 2004) and mood effects (Schwarz & Clore,
relevant negative events was high and the likelihood of
2007), generating many disasters (hard condition) likely
positive events was low (Study 2). This ?nding is consis-
produced the experience of dif?culty/low ?uency, which
tent with other research documenting the effects of emo-
then caused the participants in the hard condition to ques-
tions elicited by major events on judgment and decision-
tion the diagnostic value of their feelings for estimat-
making (Lerner et al., 2003). The implication of this in-
ing risk. In the easy condition, the subjective ease-of-
fusion of affect in everyday judgment is vast. Not only
thought/?uency should explain why these participants
may judgments be affected when the affect is considered
were less likely to question the validity of their feel-
relevant, such as the perceived risk of travelling to areas
ings for the judgment task. These ?ndings are consis-
affected by the disaster, but also perhaps affected are ev-
tent with previous research on ?uency and decision mak-
eryday decisions concerning consumption, health, social
ing (Lerner & Gonzalez, 2005; Novemsky et al., 2007).
and ?nancial domains.
In addition to this replication, we found that the ease-
Emotions and moods are usually determined in a
of-thought manipulation did not substantially in?uence
highly idiosyncratic manner (Morris, 1999) suggesting
ratings of well-being. Schwarz (2004) suggested that
that the overall effect of feelings on individual every-
the subjective experience of ?uency is a form of meta-
day decisions will vary considerably across individu-
cognitive experience that helps inform judgments. While
als. Therefore, the net effect on a societal level will
the validity and relevance of feelings as a proxy of risk
vary depending on the mean mood of the population
estimates may have been called into question by the ?u-
(Hirschliefer & Shumway, 2003). However, in the case
ency manipulation, there is little reason why the validity
of affect elicited by an event that is important or relevant
of the feelings per se should be questioned by this ma-
for a whole society or country, the impact of affect on in-
nipulation. However, previous research has noted that
dividual decisions as well societal decisions may be much
the experience of ?uency may generate positive affect
more homogenous and far-reaching. The current research
which potentially could in?uence the obtained difference
does not speak directly to this issue since we did not di-
between the easy and hard conditions (Reber, Schwarz &
rectly assess this type of national mood change in a whole
Winkielman, 2004). Several suggestions regarding why
population. Instead, we studied a sample of people who
?uency is marked with positive affect have been offered,
recently experienced the aftermath of natural disaster. Al-
ranging from perceptual harmony to the adaptive value
though it is dif?cult to conclude with certainty that the
of processing information with ease (Schwarz & Clore,
experimental approach used here is representative of the
2007). While we cannot completely refute the possibil-
effects on a whole population, a comparison with other
ity that the ?uency manipulation induced positive affect
data suggests that the reactions of our participants resem-
in the present research, the ?nding that well-being ratings
bled that of the larger population. In other studies con-
were comparable in the hard vs. easy conditions suggests
ducted in Sweden using nationally representative samples
that this effect was minor. Most importantly, even if par-
immediately following the tsunami, and six months later,
ticipants in the easy condition had more positive affect
we have found effects on judgment tasks similar to those

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
70
found in the present study, suggesting that the experimen-
disasters are relatively rare phenomena and tried to cor-
tal manipulation used here is a reasonably valid approach
rect for this by minimizing reliance on their affective re-
to investigate the effects of feelings associated with the
actions. However, it may be argued that this assumes that
tsunami disaster (Västfjäll, Peters & Slovic, 2007).
participants were aware of their reactions, what caused
Does this suggest that affect elicited by a major natu-
them, and had a naive theory about how feelings may spill
ral disaster, and its effect on judgments, is different from
over to judgments. The present results cannot be used
other type of mood effects? Previous research studying
to infer the exact psychological mechanisms involved in
the emotional impact of natural disasters/major events
these correction effects and future research is needed to
has found that generalized anxiety and depression (Lau
resolve this issue. However, the mere fact that correction
et al., 2006), negative well-being (Grandien et al., 2005)
processes could be relatively easily elicited using sim-
and negative speci?c emotions (Lerner et al., 2003) tend
ple manipulations has important implications for every-
to increase compared to normal times. The results from
day judgments. Many of the negative effects of the feel-
the manipulation check in Study 1 seem consistent with
ings associated with a major environmental event such as
this increased negative response in that the experimen-
risk aversion concerning travel after a disaster (Grandien
tal manipulation resulted in a general increase in speci?c
et al., 2005) and over-insurance when affective images of
negative emotions and negatively valenced mood. Again,
terrorism are made salient (Hsee & Kunreuther, 2000),
this ?nding is in line with the results of the nationally
as well as the impact of those feelings on everyday deci-
representative sample; immediately after the tsunami,
sions, could be mitigated by simple reminders of the low
a generalized negative response was found. When the
probability of the event. However, rather than simply stat-
same negative emotions were measured six months later,
ing that the event is of low probability (experience by de-
all ratings signi?cantly decreased (Västfjäll, Peters &
scription; Hertwig et al., 2004), ?uency manipulations let
Slovic, 2007). Other studies using nationally represen-
individuals experience (through meta-cognitive feelings)
tative Swedish samples have found similar effects (Gran-
that the event is unlikely. Such manipulations may prove
dien et al., 2005). As pointed out earlier, it seems that
to be more effective than traditional means of providing
the main difference between a normal mood change and
debiasing information (Schwarz et al., 2007). Although
a change brought about by a natural disaster lies in the
beyond the scope of the current article, future research
fact that national moods are large-scale reactions that may
could further address this issue by contrasting debiasing
be quite homogenous across individuals. Further, this
techniques that rely on description against techniques re-
type of affect is continually bolstered by new information
lying on experiential information. However, the ?nding
(from media, other people, etc.; Mann, 2007) that is as-
that experienced feelings were relatively unaffected by
sociated with uncertainty (How many people were killed?
the ?uency manipulations suggests that the potential im-
Did I know someone who went on vacation in Thailand?),
pact of affect may persist over time. The implication of
and may therefore not dissipate as quickly as a normal
this ?nding is that debiasing strategies would need to be
change in mood (Russell, 2003). This type of change in
used repeatedly for each new judgment for the duration
affective tone or affective background may therefore be
of the mood.
more similar to chronic impairments of the mood system
Another force counteracting the effects of ?uency
(Morris, 1999). However, research has demonstrated that
is motivated information processing.
Rothman and
chronic affect in?uences risk perceptions and judgments
Schwarz (1998) found that ?uency effects could be re-
in similar ways to normal incidental mood (Gasper &
versed when participants were motivated to think about
Clore, 1998). For that reason, we expect that the present
an issue carefully. It is thus possible that strategies rely-
results will be informative for any type of study on affect
ing on low ?uency to debias affect-laden risk perception
and its impact on judgments.
could back?re and individuals still would exhibit biased
Another major ?nding was that the potentially large
judgments similar to that of individuals not using these
impact of affect on behavior was mitigated by very subtle
strategies.
manipulations of the ease with which examples of other
Overall, this research suggests that major environmen-
disasters came to mind. The result that participants who
tal events may send psychological ripples globally, with
were asked to list many (in contrast to few) other natu-
the consequence that individuals and societies remote
ral disasters corrected for the effect of their feelings on
from the actual disaster may change their everyday deci-
risk estimates is consistent with the notion that meta-
sion behavior. The ?ndings reported here may be used to
cognitive processes have important biasing/debiasing ef-
better understand public risk perception and decision be-
fects on judgments (Schwarz et al., 2007). The extent of
havior in the aftermath of natural disasters. Further, the
correction on judgments may depend on the naive theory
present research is a ?rst step towards developing means
used by the participant (Schwarz, 2004). In the present
to counteract the sometimes negative impact of feelings
case, it seems likely that participants realized that natural
on judgments.

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
71
References
Personality and Social Psychology Bulletin, 31, 454–
466.
Andrade, E. B. (2005).
Behavioral consequences of
Lerner, J. S., Gonzalez, R. M., Small, D. A., and Fis-
affect: Combining evaluative and regulatory mecha-
chhoff, B. (2003). Effects of fear and anger on per-
nisms. Journal of Consumer Research, 32, 355–362
ceived risks of terrorism: A national ?eld experiment.
Carstensen, L. L. (2006). The in?uence of sense of time
Psychological Science, 14, 144–150.
on development. Science, 312, 1913–1915.
Lichtenstein, S. & Slovic, P. (2006), The construction of
Clore, G. L. & Huntsinger, J.R. (2007). How emotions
preference. New York: Cambridge University Press.
inform judgment and regulate thought. Trends in Cog-
Mann, H. (2007). Upplevd nyhetsrapportering och as-
nitive Science, 11, 393–399.
socierade känslor i samband med katastrofer: En
Diener, E. (1984). Subjective well-being. Psychological
explorativ studie.
[News coverage and experi-
Bulletin, 95, 542–575.
enced feelings in connection to disasters: and ex-
Finucane, M.L., Alhakami, A., Slovic, P., & Johnson, S.
plorative study]. Reports from Stockholm Univer-
M. (2000). The affect heuristic in judgments of risks
sity. http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-
and bene?ts. Journal of Behavioral Decision Making,
6777 (retrieved 2007–01–19).
13, 1–17.
Morris, W. N. (1999). The mood system. In D. Kahne-
Fung, H. H. & Carstensen, L. L. (2006). Goals change
man, E. Diener, & N. Schwarz (Eds.), Well-being: The
when life’s fragility is primed: Lessons learned from
foundations of hedonic psychology (pp. 169–189).
older adults, the September 11th attacks and SARS.
Novemsky, N., Dhar, R., Schwarz, N., & Simonson, I.
Social Cognition, 24, 248–278.
(2007). Preference ?uency in choice. Journal of Mar-
Gasper, K., & Clore, G. L. (1998). The persistent use of
keting Research, 44, 347–356.
negative affect by anxious individuals to estimate risk.
Pavot, W., & Diener, E. (1993). Review of the satisfaction
Journal of Personality & Social Psychology, 74(5),
with life scale. Personality Assessment, 5, 164–172.
1350–1363.
Peters, E. (2006). The functions of affect in the con-
Grandien, C., Nord, L., & Strömbäck, J. (2005). Efter
struction of preferences. In S. Lichtenstein & P. Slovic
?odvågskatastrofen (after the Tsunami disaster). Kris-
(Eds.), The construction of preference, New York:
beredningsmyndighetens temasserie: Stockholm.
Cambridge University Press.
Hertwig, R., Barron, G., Weber, E. U., & Erev, I. (2004).
Pham, M. T. (1998). Representativeness, relevance, and
Decisions from experience and the effect of rare events
the use of feelings in decision making. Journal of Con-
in risky choices. Psychological Science, 15, 534–539.
sumer Research, 25, 144–159.
Hirschleifer, D., & Shumway, T. (2003). Good day sun-
Reber, R., & Schwarz, N. (1999). Effects of percep-
shine: Stock returns and the weather. The Journal of
tual ?uency on judgments of truth. Consciousness and
Finance, 58, 1009–1032.
Cognition, 8, 338–342.
Hsee, C. K. & Kunreuther H. C. (2000). The affection
Reber, R., Schwarz, N., & Winkielman, P. (2004). Pro-
effect in insurance decisions. Journal of Risk and Un-
cessing ?uency and aesthetic pleasure: Is beauty in
certainty, 20, 141–159.
the perceiver’s processing experience? Personality and
Isen, A. M. (1997). Positive affect and decision making.
Social Psychology Review, 8, 364–382.
In W. M. Goldstein & R. M. Hogarth (Eds.), Research
Rothman, A. J., & Schwarz, N. (1998). Constructing per-
on judgment and decision making: Currents, connec-
ceptions of vulnerability: Personal relevance and the
tions, and controversies (pp. 509–534). New York:
use of experiential information in health judgments.
Cambridge University.
Personality and Social Psychology Bulletin, 24, 1053–
Johnson, E. J., & Tversky, A. (1983). Affect, generaliza-
1064.
tion, and the perception of risk. Journal of Personality
Russell, J. A. (2003). Core affect and the psychological
and Social Psychology, 45, 20–31.
construction of emotion. Psychological Review, 110,
Lang, F. R. & Carstensen, L. L. (2002). Time counts:
145–172.
Future time perspective, goals and social relationships.
Schwarz, N. (2004). Meta-cognitive experiences in con-
Psychology and Aging. 17, 125–139.
sumer judgment and decision making. Journal of Con-
Lau, J., Lau, M., Kim, J. H. & Tsui, H. I. (2006). Im-
sumer Psychology, 14, 332–348.
pacts of media coverage on the community stress level
Schwarz, N., & Clore, G. L. (1983). Mood, misattribu-
in Hong Kong after the tsunami on 26 December 2004.
tion, and judgments of well-being: Informative and di-
Journal of Epidemiology & Community Health, 60,
rective functions of affective states. Journal of Person-
675–682.
ality & Social Psychology, 45, 513–523.
Lerner, J. S., & Gonzalez, R. M. (2005). Forecasting
Schwarz, N., & Clore, G. L. (2007). Feelings and phe-
one’s future based on ?eeting subjective experiences.
nomenal experiences. In A. Kruglanski & E. T. Hig-

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Affect and future optimism
72
gins (eds.), Social psychology. Handbook of basic
Slovic, P. (1995). The construction of preference. Amer-
principles (2nd ed., pp. 385–407). New York: Guil-
ican Psychologist, 50, 364–371.
ford.
Slovic, P., Finucane, M., Peters, E., & MacGregor, D.
Schwarz, N., & Strack, F. (1999). Reports of subjective
(2002). The affect heuristic. In T. Gilovich, D. Grif-
well-being: Judgmental processes and their method-
?n, & D. Kahneman (Eds.), Heuristics and biases: The
ological implications. In D. Kahneman, E. Diener, &
psychology of intuitive judgment (pp. 397–420). New
N. Schwarz (Eds.), Well-being: The foundations of he-
York, NY: Cambridge University Press.
donic psychology (pp. 61–84). New York: Russell-
Västfjäll, D., Friman, M., Gärling, T & Kleiner, M.
Sage.
(2002). The measurement of core affect: A Swedish
Schwarz, N., Sanna, L., Skurnik, I., & Yoon, C. (2007).
self-report measure. Scandinavian Journal of Psychol-
Metacognitive experiences and the intricacies of set-
ogy, 43, 19–31.
ting people straight: Implications for debiasing and
Västfjäll, D, Peters, E., & Slovic, P. (2007). Affect and
public information campaigns. Advances in Experi-
risk-bene?t judgments in the aftermath of tsunami dis-
mental Social Psychology, 39, 127–161.
aster. Manuscript submitted for publication.
Siemer, M., & Reisenzein, R. (1998). Effects of mood on
Wright, W. F., & Bower, G. H. (1992). Mood effect on
evaluative judgments: In?uence of reduced processing
subjective probability assessment. Organizational Be-
capacity and mood salience. Cognition and Emotion,
havior and Human Decision Processes, 22, 276–291.
12, 783–806.
Shiv, B., Loewenstein, G., Bechara, A., Damasio, H., and
Damasio, A. (2005). Investment behavior and the dark
side of emotion. Psychological Science, 16, 435–439.

---