Looking Forward, Looking Back: Anticipation Is More Evocative Than Retrospection

Abstract

The results of 5 experiments indicate that people report more intense emotions during anticipation of, than during retrospection about, emotional events that were positive (Thanksgiving Day), negative (annoying noises, menstruation), routine (menstruation), and hypothetical (all-expenses-paid ski vacation). People's tendency to report more intense emotion during anticipation than during retrospection was associated with a slight, but only occasionally significant, tendency for people to expect future emotions to be more intense than they remembered past emotions having been. The greater evocativeness of anticipation than retrospection was also associated with and statistically mediated by participants' tendency to report mentally simulating future emotional events more extensively than they report mentally stimulating past emotional events. The conclusion that anticipation is more evocative than retrospection has implications for research methodology, clinical practice, decision making, and well-being.

Full text

Looking Forward, Looking Back: Anticipation Is More Evocative
Than Retrospection
Leaf Van Boven
University of Colorado at Boulder
Laurence Ashworth
Queen’s University
The results of 5 experiments indicate that people report more intense emotions during anticipation of,
than during retrospection about, emotional events that were positive (Thanksgiving Day), negative
(annoying noises, menstruation), routine (menstruation), and hypothetical (all-expenses-paid ski vaca-
tion). People’s tendency to report more intense emotion during anticipation than during retrospection was
associated with a slight, but only occasionally significant, tendency for people to expect future emotions
to be more intense than they remembered past emotions having been. The greater evocativeness of
anticipation than retrospection was also associated with and statistically mediated by participants’
tendency to report mentally simulating future emotional events more extensively than they report
mentally stimulating past emotional events. The conclusion that anticipation is more evocative than
retrospection has implications for research methodology, clinical practice, decision making, and well-
being.
Keywords: affective forecasting, anticipation, emotion, memory, simulation
The bias toward the future applies most clearly to events that are in
themselves pleasant or painful. The thought of such events affects us
more when they are in the future rather than the past. (Parfit, 1984, p.
160)
Despite the deep desire of many science fiction aficionados, the
physical ability to travel in time has eluded reality. Even though
one can purchase inexpensive time travel machines on the Internet,
the constraints of physical laws make such products persistently
disappointing. Mental reality is far less constrained than physical
reality, however; people have the ability to, and do routinely,
engage in mental “time travel,” reliving past events and imagining
future events (Johnson & Sherman, 1990; Tulving, 2002). Such
mental time travel matters because, among other things, it can
influence current cognitions, decisions, and emotions. In this arti-
cle, we examine the consequences of mental time travel for emo-
tional experiences in the here and now.
Emotions aroused by anticipation of and retrospection about
emotional events pervade everyday experience. For example, in-
dividuals may use the emotion aroused by anticipation of future
events or retrospection about past events as a basis for making
decisions (Mellers, Schwartz, Ho, & Ritov, 1997; Schwarz, 2002;
Soman, 2003; Wirtz, Kruger, Scollon, & Diener, 2003). Individ-
uals may measure the quality of their own lives based on the
emotions aroused during retrospection about past events (Chang,
2004; Strack, Schwarz, & Gschneidinger, 1985). Anxiety associ-
ated with the anticipation of future stressful events, such as public
speaking, can be debilitating (Eng, Coles, Heimberg, & Safren,
2005; Hambrick, Turk, Heimberg, Schneier, & Liebowitz, 2003;
Leary & Kowalski, 1995). And the emotions aroused by persistent,
unpleasant, unwanted retrospection can be diagnosed as depressive
rumination (Papageorgiou & Wells, 2005). Given the importance
of emotional reactions to anticipation and retrospection, it is im-
portant to ask whether people have different emotional reactions
when looking forward than they have when looking backward.
We wish to clarify at the outset the distinction between people’s
predicted and remembered emotions versus their emotional reac-
tions to anticipation and retrospection. As an illustration, vacation-
ers can predict how enjoyable a future beach holiday will be and
can then remember how enjoyable that holiday was. Vacationers
can also experience positive emotions in the here and now, during
the anticipation of a future beach holiday (imagining fine white
sand, crystal blue water, and well-tanned sunbathers), and during
retrospection about a past holiday (remembering beaches, blues,
and bodies).
That anticipation and retrospection can influence current emo-
tions has been well established (if inadvertently) with various
methodologies used to manipulate emotional arousal. Researchers
have sometimes used retrospection to manipulate emotions by
asking people to contemplate an emotional episode from their
personal past (e.g., Lerner & Keltner, 2001; Levenson, Carstensen,
& Gottman, 1994; Strack et al., 1985). At other times, researchers
have used anticipation to manipulate emotions by asking people to
imagine future emotional episodes, such as an upcoming stressful
task that makes people anxious (e.g., Folkman & Lazarus, 1985;
Mauss, Wilhelm, & Gross, 2004).
Leaf Van Boven, Department of Psychology, University of Colorado at
Boulder; Laurence Ashworth, School of Business, Queen’s University,
Kingston, Ontario, Canada.
This research was supported by National Science Foundation Grant
0552120. Peter McGraw and Tiffany Ito provided valuable comments on
an earlier version of this article. We thank Jennifer Donnelly, Erin Gilmer,
Alex Heath, Valerie Munger, and Sarah Scarbeck for helping us collect
data.
Correspondence concerning this article should be addressed to Leaf Van
Boven, Department of Psychology, University of Colorado at Boulder, 345
UCB, Boulder, CO 80309-0345. E-mail: vanboven@colorado.edu
Journal of Experimental Psychology: General Copyright 2007 by the American Psychological Association
2007, Vol. 136, No. 2, 289–300 0096-3445/07/$12.00 DOI: 10.1037/0096-3445.136.2.289
289

Our central question is whether people might experience and
report emotional reactions of a different intensity during anticipa-
tion of, than during retrospection about, emotional events. We
suggest, following Parfit’s (1984) observation, that anticipation
arouses more intense emotion in the here and now than does
retrospection. This hypothesis follows from three potential dis-
crepancies between the way people think about future emotional
events and the way they think about past emotional events.
Discrepancies Between Anticipation and Retrospection
Uncertain Futures
One potential discrepancy between anticipation and retrospec-
tion is that future events are usually less certain than are past
events. The greater uncertainty of future events can produce more
intense emotion during anticipation than during retrospection be-
cause uncertainty amplifies emotional reactions to the contempla-
tion of emotional events (Wilson, Ceterbar, Krermer, & Gilbert,
2005). Not only can uncertainty be inherently stressful and upset-
ting (Buhr & Dugas, 2002; Monat, Averill, & Lazarus, 1972; van
de Bos, 2001), people are also less able to normalize, or rational-
ize, uncertain events, making those events more evocative (Wilson
et al., 2005).
Because future events tend to be less certain than past events,
and because uncertainty amplifies emotion, anticipation of emo-
tional events may be more arousing than retrospection about those
events. Because there are more ways that a future beach holiday
might happen—more beaches one might visit, more sunsets one
might see, more books one might read—than ways it did happen,
people might experience more pleasure during anticipation of, than
during retrospection about, beach holidays.
Extreme Expectations
Another potential discrepancy between anticipation and retro-
spection is that people may predict that they will experience more
extreme emotional reactions to future events than the emotional
reactions they remember experiencing during past events. People
tend to expect the best of good events and the worst of bad events
(Schkade & Kahneman, 1998; Wilson, Wheatley, Meyers, Gilbert,
& Axsom, 2000), but they may end up with more moderate
memories of what actually transpired. This is partly because emo-
tional events often provoke rationalization processes that make
those events seem more mundane in hindsight than in foresight
(Wilson & Gilbert, 2003). Indeed, at least two studies suggested
that people expect to enjoy future vacations more than they re-
member enjoying past vacations, although the statistical signifi-
cance of such comparisons was not reported in the original articles
(Mitchell, Thompson, Peterson, & Cronk, 1997; Wirtz et al.,
2003).
To the extent that people expect to experience more extreme
emotions in the future than they remember having experienced in
the past, it follows that people might be more aroused during
anticipation than during retrospection (Elster & Loewenstein,
1992; Wirtz et al., 2003). People may expect future beach holidays
to be more enjoyable than they remember past holidays having
been, and these extreme expectations can make the anticipation of
future holidays more pleasant than the retrospection of past holi-
days.
Asymmetric Mental Simulation
A final potential discrepancy arises from the natural covariation
between temporal perspective, mental simulation, and emotional
arousal. A key function of emotion in everyday life is to place
people in states of action readiness, directing them to approach or
to avoid significant events (Frijda, 1988; Frijda, Kuipers, & ter
Schure, 1989; Lazarus, 1991; Neese, 1990). The functionality of
action readiness is obviously more relevant to future than to past
events and, therefore, may be associated with a more extensive
mental simulation of future events than past events. That is, people
might think more thoroughly and episodically about future events
than about past events because such mental simulation would
allow for more informed decisions about approaching or avoiding
future events. To the extent that people mentally simulate future
events more extensively than they simulate past events, they might
experience more intense emotion during anticipation than during
retrospection. This is because mental simulation can amplify emo-
tion, possibly because simulation engenders embodied states that
mirror the states that people would actually experience in emo-
tional situations (Niedenthal, Barsalou, Winkielman, Krauth-
Gruber, & Ric, 2005).
Consistent with the possibility that mental simulation amplifies
emotional arousal, people experience more intense emotion when
asked to contemplate the detailed what of distant emotional events
than when asked to contemplate the more abstract why of those
events (Kross, Ayduk, & Michel, 2005; Strack et al., 1985). Events
that people episodically remember experiencing are rated as more
emotional than are events that people remember indirectly (Lind-
say, Wade, Hunter, & Read, 2004). Asking people to generate
mental imagery while listening to descriptions of emotional events
arouses more intense emotion than asking people to attend to the
verbal meaning of the descriptions (Holmes & Mathews, 2005).
Thus, people may mentally simulate future beach holidays more
extensively than they simulate past beach holidays. They may
imagine future strolls on sandy beaches beneath bright suns more
than they imagine past sunny, sandy strolls, and this differential
simulation may arouse more pleasant emotions during anticipation
than during retrospection.
Overview of the Present Experiments
The primary purpose of the present experiments was to examine
whether people might, in fact, report more intense emotional
reactions to anticipation of, than to retrospection about, emotional
events. Participants in five experiments were asked to contemplate
future or past emotional events. We measured participants’ self-
reported emotional reactions to anticipation and retrospection. We
expected participants to report more intense emotion during antic-
ipation than during retrospection.
We also examined the viability of the three discrepancies dis-
cussed earlier. First, we sought to control for asymmetric uncer-
tainty by asking people to contemplate well-defined emotional
events that are routine (Experiment 2) and unambiguous (Experi-
ments 2, 3, and 5). In each case, there is little difference in the
uncertainty of future events and the uncertainty of past events, so
290 VAN BOVEN AND ASHWORTH

asymmetric uncertainty should be unlikely to produce more in-
tense emotion during anticipation than during retrospection. Sec-
ond, we tested the viability of extreme expectations by examining
whether predicted emotional reactions to future events were more
extreme than remembered emotional reactions to past events and
whether the difference between predicted and remembered emo-
tional extremity statistically mediated any difference between
emotional reactions to anticipation and emotional reactions to
retrospection.
Third, we tested the role of asymmetric mental simulation in two
ways. One is that we examined whether participants reported more
intense emotions during anticipation of, than during retrospection
about, purely hypothetical events (Experiment 4). A demonstration
that anticipation of hypothetical events is more evocative than
retrospection about those events, controlling for any extreme ex-
pectations, would imply that there is some difference in the way
people imagine future emotional events versus the way they imag-
ine past emotional events. We also examined whether participants
reported mentally simulating future events more extensively than
they reported mentally stimulating past events and whether this
differential mental simulation statistically mediated the difference
between reported emotion during anticipation and reported emo-
tion during retrospection (Experiment 5).
Experiment 1: Thanksgiving Day
In an initial study, we asked people to anticipate and then to
retrospect about a familiar, naturally occurring emotional event:
their annual Thanksgiving holiday. We asked them to report their
emotional reactions to the anticipation and retrospection and to
predict and remember their enjoyment of the holiday. We hypoth-
esized that people would report more intense emotional reactions
to anticipation than retrospection. We also examined whether any
difference between the extremities of the participants’ predicted
emotional experience and their remembered emotional experience
would be associated with the difference between the intensity of
emotional reactions to anticipation and the intensity of emotional
reactions to retrospection.
Method
Participants were 36 university students (15 males, 21 females)
who completed two questionnaires distributed via electronic mail
in exchange for extra credit in their psychology courses and for a
coffee mug emblazoned with the university logo.
1
Participants
completed the first questionnaire approximately 2 weeks prior to
Thanksgiving Day (M⫽13.83 days, SD ⫽1.23); they completed
the second questionnaire approximately 2 weeks after Thanksgiv-
ing Day (M⫽13.36 days, SD ⫽0.64).
Both before and after their holiday, participants were asked to
spend a few minutes contemplating their (future or past) Thanks-
giving Day. Participants were then asked (on a 9-point scale;
variables for future and past versions of each statement appear in
brackets), “When you think about what your Thanksgiving Day
[will be/was] like, how happy does it make you?” (1 ⫽not happy,
9⫽very happy). Participants also answered three questions about
how much they expected to enjoy, or remembered enjoying, their
Thanksgiving Day: “How happy [will/did] this [upcoming/past]
Thanksgiving Day make you?” (1 ⫽not happy, 9⫽very happy);
“How much [will/did] you enjoy this [upcoming/past] Thanksgiv-
ing Day?” (1 ⫽not enjoy,9⫽enjoy very much); and “How much
fun [will/was] this [upcoming/past] Thanksgiving Day [be?/?]”
(1 ⫽not fun,9⫽extremely fun). After returning the second
questionnaire, participants were thanked and debriefed via an
electronic mail message.
Results and Discussion
As expected, participants reported that anticipation of their
future Thanksgiving Day made them happier (M⫽6.86, SD ⫽
1.76) than retrospection about their past holiday (M⫽6.17, SD ⫽
1.86), t(35) ⫽2.91, SE ⫽0.24, p⫽.006, d⫽0.98 (see Table 1).
After averaging participants’ three predicted and remembered en-
joyment measures in two composite indices (␣s⫽.90 and .91, for
each index), there was a nonsignificant tendency for participants to
expect that they would enjoy their future Thanksgiving more (M⫽
6.75, SD ⫽1.47) than they remembered having enjoyed their past
Thanksgiving (M⫽6.45, SD ⫽1.33), t(35) ⫽1.67, SE ⫽0.18,
p⫽.105, d⫽0.56 (see Table 1).
We conducted a linear regression to examine whether the dif-
ference in current reported emotions might be partially explained
by the difference between predicted enjoyment and remembered
enjoyment. We calculated the difference between participants’
current reported happiness during anticipation of Thanksgiving
and their current reported happiness during their retrospection
about Thanksgiving (M⫽0.69, SD ⫽1.43) and regressed that
difference on the difference between participants’ predicted enjoy-
ment and their remembered enjoyment (M⫽0.30, SD ⫽1.08),
which was associated with a significant coefficient (b⫽.86, ␤⫽
.65), t(34) ⫽5.04, SE ⫽0.17, p⬍.001. Importantly, the constant
was significantly positive (b⫽.44), t(34) ⫽2.29, SE ⫽0.19, p⫽
.029, indicating that anticipation was associated with more intense
reported happiness than retrospection, when controlling for the
extremity of predicted and remembered enjoyment.
These results offer preliminary support for our central prediction
that people would report more intense emotional reactions to
anticipation of an emotional event than to retrospection about the
same emotional event. These results also suggested that there was
a nonsignificant tendency for individuals to expect that they would
enjoy their future Thanksgiving Day more than they remembered
enjoying their past Thanksgiving. The difference between pre-
dicted enjoyment and remembered enjoyment was correlated with
the difference in reported current happiness. However, the differ-
ence between predicted enjoyment and remembered enjoyment did
not fully account for the difference between current happiness
during anticipation and current happiness during retrospection.
Experiment 2: Menstruation
We next broadened our investigation to the anticipation of and
retrospection about routinely experienced and familiar negative
1
An additional 16 participants completed the anticipation, but not the
retrospection, questionnaire. There were no significant differences in pre-
dicted enjoyment, t(50) ⫽0.38, SE ⫽0.44, p⫽.705, observed power ⫽
.066, d⫽0.11, or in immediate emotions associated with anticipation
t(50) ⫽0.44, SE ⫽0.53, p⫽.660, observed power ⫽.072, d⫽0.13,
between participants who completed the entire study and those who com-
pleted only the anticipation questionnaire.
291
ANTICIPATION, RETROSPECTION, AND EMOTION

emotions. We asked women to contemplate either their upcoming
menstrual period or their most recent menstrual period, to report
their emotional reactions to the contemplation, and to predict or
remember how unpleasant their menstrual experience would be or
was. We assumed that women would expect their menstrual ex-
perience to be unpleasant and would remember it as having been
unpleasant—more unpleasant, perhaps, than women report while
actually experiencing their periods (McFarland, Ross, & DeCour-
ville, 1989). We predicted that women would report more negative
emotion during anticipation of their period than during retrospec-
tion about it.
Method
Participants were 51 female university students who were not
currently experiencing their menstrual period and who participated
in the study in exchange for course credit or a chocolate bar.
Participants randomly assigned to the anticipation condition (n⫽
25) were asked to “take a few minutes to think about what your
next period will be like . . . the physical and emotional symptoms
you will experience” and try “really to get a sense of what your
period will be like.” Participants in the retrospection condition
(n⫽26) were given nearly identical instructions, except that they
were asked to contemplate their last period rather than their next
period.
Participants were then asked, “When you think about your
period, how does it make you feel right now? That is, how does the
act of thinking about your period influence your current mood?”
(on a 9-point scale; 1 ⫽has no effect,9⫽substantially worsens
current mood). Participants were also asked to predict or to re-
member the “overall level of discomfort or unpleasantness that you
[expect to experience/remember experiencing] during your [next/
last] period” (on a 9-point scale; 1 ⫽no discomfort,9⫽extreme
discomfort). Finally, participants were asked whether their periods
were “regular (i.e., you get your period every month)” and whether
they were taking any medications, such as birth control pills, that
might influence their menstrual cycle. Participants were then
thanked and debriefed.
Results and Discussion
We excluded 2 participants from the analyses: one who did not
report her current emotion and one who did not report how
unpleasant her previous period was (resulting N⫽49). Most
women (80.43%) reported experiencing regular periods. Nearly
half (43.48%) reported taking medication (primarily birth control
pills) that influenced their menstrual cycle. Neither the regularity
of participants’ menstrual cycles nor their use of medication dif-
fered by condition.
As expected, women reported that their mood was worse during
anticipation of their next period (M⫽4.88, SD ⫽2.44) than it was
during retrospection about their last period (M⫽3.32, SD ⫽2.34),
t(47) ⫽2.28, SE ⫽0.68, p⫽.027, d⫽0.66 (see Table 2).
Participants did not predict that their next period would be signif-
icantly more unpleasant (M⫽5.63, SD ⫽2.14) than they remem-
bered their last period having been (M⫽5.16, SD ⫽1.82), t(47) ⫽
0.82, SE ⫽0.58, p⫽.416, observed power ⫽.127, d⫽0.24, (see
Table 2). A multiple linear regression indicated that women’s
reports of how much thinking about their period worsened their
mood was associated both with the extremity of their predicted or
remembered menstrual unpleasantness (b⫽.65, ␤⫽.52), t(46) ⫽
4.37, SE ⫽0.15, p⬍.001, d⫽1.29, and with their temporal
perspective (coded 1 and –1 for anticipation and retrospection,
respectively; b⫽.63, ␤⫽.25), t(46) ⫽2.14, SE ⫽0.29, p⫽.037,
d⫽0.63.
These results conceptually replicate the central findings of Ex-
periment 1 with an even more routinely experienced and familiar
negative emotional event. Women reported more intense emotion
during anticipation of, than during retrospection about, their peri-
ods, even after the extremity of predicted or remembered unpleas-
antness was controlled for. These findings cast some doubt on the
possibility that differential uncertainty was primarily responsible
for the greater evocativeness of anticipation than retrospection.
Because women routinely experience their periods, they are un-
likely to be substantially less certain about what their future
periods will be like compared with what their past periods were
like.
Experiment 3: Annoying Noise
The results presented thus far suggest that people report more
intense emotion during anticipation of, than during retrospection
about, naturally occurring emotional events. We next sought to
examine reported emotional reactions to anticipation and retro-
spection in a laboratory setting that would allow more control over
the nature of the emotional event and over the temporal distance
between participants and the event. We also wished to measure
people’s emotional reactions to anticipation of and retrospection
about a well-defined hedonic event that was unlikely to be asso-
ciated with differential uncertainty.
Table 1
Experiment 1: Participants’ Reactions to Anticipation and
Retrospection About Thanksgiving Day
Measure
Temporal perspective
Anticipation Retrospection
Reported current happiness 6.86 (1.76) 6.17 (1.86)
Predicted or remembered enjoyment 6.75 (1.47) 6.45 (1.33)
Note. The numbers in the table are means and, in parentheses, standard
deviations.
Table 2
Experiment 2: Female Participants’ Reactions to Anticipation
and Retrospection About Their Menstrual Periods
Measure
Temporal perspective
Next
period
Last
period
Reported worsening of current mood 4.88 (2.44) 3.32 (2.34)
Predicted or remembered discomfort 5.63 (2.14) 5.16 (1.82)
Note. The numbers in the table are means and, in parentheses, standard
deviations.
292 VAN BOVEN AND ASHWORTH

We asked participants to listen to an unambiguously annoying
noise: the sound of a dial-up computer modem connecting to an
Internet service provider. Participants in the anticipation condition
listened to a short sample sound and were told that in exactly 20
min they would listen to the same sound for a longer time period.
Participants in the retrospection condition listened to the longer
sound before listening, exactly 20 min later, to a reminder: the
same shorter sound heard by participants in the anticipation con-
dition. Thus, participants in both conditions listened to a shortened
version of an annoying noise that they either would listen to in 20
min or had listened to 20 min ago.
After listening to the sample or reminder sound, depending on
condition, we measured participants’ emotional reactions with a
less direct and more comprehensive measure than was used in
previous experiments. Participants in Experiment 3 were simply
asked to indicate how much they were experiencing various emo-
tions. We predicted that participants would report more intense
emotion following anticipation than retrospection.
Method
Participants were 61 university undergraduates (27 males, 34
females) who participated in exchange for course credit. Partici-
pants learned that they would be asked to listen to the sound of a
dial-up telephone modem connecting to an Internet service pro-
vider. Pretests and personal experience indicated that the screech-
ing, disharmonic sound is indeed rather noxious and would elicit
negative emotion.
Participants randomly assigned to the anticipation condition
(n⫽32) were first asked to listen to a 5-s sample of the noise
played loudly over external speakers connected to a desktop com-
puter. They were then told that in exactly 20 min they would listen
to the same sound, at the same volume, extended to 58 s. Partic-
ipants in the retrospection condition (n⫽29) were asked to listen
to 58 s of the sound before completing unrelated questionnaires for
exactly 20 min. They were then asked to listen to a reminder: the
same 5-s sound heard by participants in the anticipation condition.
After listening to the 5-s sound, participants in both conditions
were asked to contemplate listening to the full 58-s sound. For
participants in the anticipation condition, this longer sound was 20
min in the future. For participants in the retrospection condition,
the sound was 20 min in the past.
To measure emotional reactions to anticipation or retrospection,
we asked participants to report how much (on a 6-point scale; 1 ⫽
not at all,6⫽a great deal) they were currently experiencing each
of 15 emotions: agitated, angry, annoyed, comfortable, delighted,
glad, happy, irritated, negative, nervous, positive, sad, uncomfort-
able, unfortunate, and unhappy. To measure predicted (or remem-
bered) feelings while listening to the 58-s sound, we asked partic-
ipants to predict (or to remember) how much (on a 6-point scale;
1⫽not at all,6⫽a great deal) they would (or did) experience
4 emotions while listening to the longer noise: comfortable, irri-
tated, negative, and positive. Participants were thanked and de-
briefed after completing these measures.
Results and Discussion
After we reverse scored the ratings such that higher numbers
indicated more negative emotion, we averaged the 15 ratings of
current emotions into a composite measure (␣⫽.87). We simi-
larly reverse scored and then averaged participants’ four predic-
tions or memories of their emotions while listening to the longer
sound (␣⫽.66).
As expected, participants reported more intense negative emo-
tion after anticipation (M⫽3.42, SD ⫽0.78) than after retrospec-
tion (M⫽2.66, SD ⫽0.91), t(59) ⫽3.50, SE ⫽0.22, p⫽.001,
d⫽0.91 (see Table 3). Participants did not expect to feel signif-
icantly worse when they listened to the noise in 20 min (M⫽4.45,
SD ⫽1.03) than they remembered feeling when they listened to
the noise 20 min ago (M⫽4.17, SD ⫽0.89), t(59) ⫽1.14, SE ⫽
0.25, p⫽.261, observed power ⫽.201, d⫽0.30 (see Table 3). A
multiple linear regression indicated that participants’ reported cur-
rent emotion was correlated both with the extremity of their
predicted or remembered emotion (b⫽.56, ␤⫽.59), t(58) ⫽
6.31, SE ⫽0.09, p⬍.001, d⫽1.66, and with their temporal
perspective (coded 1 and –1 for anticipation and retrospection,
respectively; b⫽.30, ␤⫽.33), t(58) ⫽3.54, SE ⫽0.09, p⫽.001,
d⫽0.93.
These findings conceptually replicated the results of Experi-
ments 1 and 2 in a more controlled laboratory setting with a more
indirect and comprehensive measure of current emotional experi-
ence. Participants reported experiencing more intense negative
emotion during anticipation of, than during retrospection about, a
noxious sound, even after the extremity of their predicted or
remembered emotion was controlled for. This pattern provided
additional evidence suggesting that the extremity of predicted and
remembered emotions does not fully explain why people experi-
ence more intense emotions during anticipation than during retro-
spection. These results also cast further doubt on the role of
asymmetric uncertainty, given that the short sample or reminder
sound made clear to participants exactly what the longer sound
would be like.
Experiment 4: Hypothetical Ski Vacation
The evidence accumulated thus far indicates that people report
more intense emotion when they anticipate emotional events that
are in their future than when they retrospect about emotional
events that are in their past. In each of the preceding experiments,
however, people’s temporal perspective was confounded with an
actual temporal relation to emotional events. That is, participants
anticipated events that were actually in their future and they
retrospected about events that were actually in their past. It is
therefore difficult to ascertain whether the greater evocativeness of
anticipation, compared with that of retrospection, arose from some
Table 3
Experiment 3: Participants’ Reactions to Anticipation and
Retrospection About Listening to an Annoying Noise
Measure
Temporal perspective
Anticipation Retrospection
Reported current negative emotion 3.42 (0.78) 2.66 (0.91)
Predicted or remembered emotion 4.45 (1.03) 4.17 (0.89)
Note. The numbers in the table are means and, in parentheses, standard
deviations.
293
ANTICIPATION, RETROSPECTION, AND EMOTION

mental framing of future or past events or from participants actu-
ally being in a state of action readiness during anticipation. In
Experiment 4, we wished to disentangle temporal perspective from
actual temporal relation by asking people to look forward to or
backward on a purely hypothetical event.
The possibility that people mentally simulate future events more
extensively than they simulate past events implies that anticipation
of a hypothetical future event would produce more emotional
arousal than would retrospection about a hypothetical past event.
Because temporal perspective and mental simulation naturally
covary over time, their association may become overgeneralized,
such that merely adopting a future perspective may produce more
extensive simulation than adopting a past perspective, and this
mental simulation may arouse more intense emotion during antic-
ipation than during retrospection. To test this possibility, we asked
participants to report their emotional reactions to contemplating a
hypothetical all-expenses-paid ski vacation that was framed in
either the future or the past tense. We expected participants to
report more intense emotion during anticipation of the vacation
than during retrospection about the vacation, even though the
vacation was hypothetical in both cases.
Method
Participants were 95 university undergraduates (gender was not
recorded) who participated in exchange for course credit. Partici-
pants were asked to imagine that they had won a 5-day, all-
expenses-paid ski trip to the Whistler–Blackcomb ski resort in
British Columbia, Canada.
2
Participants randomly assigned to the
anticipation condition (n⫽45) were provided with a detailed map
of downhill ski trails at the resort and were given the following
instructions:
Imagine that you have won a holiday package from a local radio
station. Six months from now you will go on an all expenses paid ski
vacation at the Whistler–Blackcomb ski resort in British Columbia,
Canada. The trip will include a 5-night stay (Tuesday through Satur-
day) at the Chateau Whistler, five daily lift tickets for 2 (Wednesday
through Sunday), half-day private snowboard or skiing instructions
for 2, and snowboard or ski rentals (if needed). The hotel room will be
a suite in a 5-star hotel and will come with a Jacuzzi tub and a
complimentary breakfast each day of your stay. The hotel itself will be
virtually “ski to the door,” meaning that the ski chair will be a five
minute walk from the hotel’s front door. The hotel will also be
minutes from the Whistler nightlife and the other shops and restau-
rants.
Participants in the retrospection condition (n⫽50) read a nearly
identical scenario, except they were asked to imagine the trip took
place 6 months in the past. The second sentence, for instance,
began as follows: “Six months ago you went an on all expenses
paid ski vacation....”After reading the description, participants
were asked to spend 1 min imagining the vacation as though it
were actually going to (or did) happen. We conducted the exper-
iment during July and August so that participants in both condi-
tions were asked to imagine a vacation that took place in January
or February, when the snow is typically fresh and the powder is
fine.
After thinking about the future or past vacation, participants
answered three questions that measured how contemplating the
trip influenced their current emotions: “When you think about the
trip, how happy does it make you? That is, how happy do you feel
right now when you think about this trip?” (1 ⫽not happy,7⫽
extremely happy); “When you think about the trip how would you
say it affects your current mood? Does it put you in a better mood,
worse mood, or leave you unaffected?” (–3 ⫽worse mood, 0⫽no
change, ⫹3⫽better mood); and “How much would you say you
enjoy [anticipating/remembering] the Whistler trip?” (1 ⫽not at
all,7⫽a great deal). Finally, as an assessment of how much
participants predicted or remembered enjoying the vacation, they
indicated how much they thought they would actually enjoy (or
would have enjoyed) the vacation (1 ⫽not at all, 7⫽a great
deal). After completing these measures, participants were thanked
and debriefed.
Results and Discussion
After we recoded participants’ ratings of their mood change on
a scale ranging from 1 to 7, we averaged participants’ three reports
of their current emotions in a composite measure (␣⫽.76). As
expected, participants reported more intense positive emotion dur-
ing anticipation of the hypothetical future ski vacation (M⫽5.84,
SD ⫽0.92) than during retrospection about the hypothetical past
ski vacation (M⫽5.51, SD ⫽0.63), t(93) ⫽2.06, SE ⫽0.16, p⫽
.042, d⫽0.43 (see Table 4). Participants did not anticipate that
they would enjoy the vacation significantly more (M⫽6.16, SD ⫽
1.19) than participants “remembered” enjoying the vacation (M⫽
6.00, SD ⫽1.09, t(93) ⫽0.67, SE ⫽0.23, p⫽.507, observed
power ⫽.101, d⫽0.14 (see Table 4). A multiple linear regression
indicated that participants’ reported current emotion was associ-
ated with the extremity of their predicted or remembered emotion
(b⫽.471, ␤⫽.67), t(92) ⫽8.98, SE ⫽0.05, p⬍.001, d⫽1.87,
and with their temporal perspective (coded 1 and –1 for anticipa-
tion and retrospection, respectively (b⫽.129, ␤⫽.163), t(92) ⫽
2.18, SE ⫽0.16, p⫽.032, d⫽0.45.
These results demonstrate that simply adopting a future or past
temporal perspective, independent of people’s actual temporal
relation to an emotional event, is sufficient to produce more
intense emotion during anticipation than during retrospection.
These results also replicate people’s tendency to report more
intense reported emotion during anticipation than during retrospec-
tion, independent of any difference between the extremity of
predicted emotion and the extremity of remembered emotions.
This finding bolsters our suggestion that the association between
how people think about future events and how they think about
past events is overgeneralized to emotional events that are only
hypothetically in the past or future.
Experiment 5: Imagining Annoying Noises
In our final experiment, we sought to scrutinize more closely the
nature of the mental processes that might be associated with the
greater evocativeness of anticipation, compared with that of retro-
2
Because this experiment was conducted at a university in western
British Columbia, Canada, most participants (83%) had previously visited
the Whistler–Blackcomb ski resort, although none reported having won an
all-expenses-paid Whistler ski vacation. There were no main effects or
interactions that were contingent on whether participants had skied at
Whistler (largest F⫽0.40, smallest p⫽.531).
294 VAN BOVEN AND ASHWORTH

spection. In particular, we sought to gather more direct evidence
for the role of mental simulation by developing a self-report
measure of how extensively people mentally simulate an emo-
tional experience. We expected that participants would report more
extensive mental simulation during anticipation than during retro-
spection and that this differential mental simulation would statis-
tically mediate the greater evocativeness of anticipation, compared
with retrospection. A second purpose of this experiment was to
address a question begged by the results of Experiment 4; namely,
how does the emotion aroused by anticipating a hypothetical future
event compare with the emotion aroused by anticipating an actual
future event and the emotion aroused by retrospection about an
actual past event?
To pursue these two goals, we modified the procedure of Ex-
periment 3. Participants in the anticipation condition listened to a
short sample of an annoying noise and were told they would listen
to the same noise for a longer duration in 20 min. Participants in
the retrospection condition listened to the annoying noise for the
longer duration before listening to a short reminder sound exactly
20 min later. Participants in a third condition listened to the short
sample of the annoying noise, exactly as in the anticipation con-
dition, but were asked simply to imagine that they would listen to
a longer version of the noise in 20 min. It was made clear to these
participants that they would not actually listen to the noise.
The possibility that people’s different emotional reactions to
anticipation and retrospection arises from an overgeneralized as-
sociation between temporal perspective and mental simulation
suggests that participants who simply imagine listening to a noise
in the future should report more emotional arousal than those who
remember actually listening to the noise in the past. We therefore
predicted that participants in both the anticipation and imagination
conditions would report more intense emotion than participants in
the retrospection condition. We also predicted that participants in
both the anticipation and imagination conditions would report
mentally simulating the experience of actually listening to the
noise more than would participants in the retrospection condition
and that the difference in mental simulation would statistically
mediate the difference in emotion between conditions.
Method
Participants were 83 university undergraduates (gender was not
recorded) who participated in exchange for course credit. Partici-
pants were randomly assigned to either the anticipation condition
(n⫽27), the retrospection condition (n⫽29), or the imagination
condition (n⫽27). The procedure and measures in the anticipation
and retrospection conditions were identical to those of Experiment
3. In the imagination condition, participants listened to the 5-s
sample of the annoying sound of a dial-up telephone modem
connecting to the Internet. In contrast with the anticipation condi-
tion, participants in the imagination condition then read, “You will
NOT have to listen to the [58-s] noise, but we would like you to
IMAGINE that you were going to listen to the noise in 20 min
time.”
After listening to the 5-s sample or reminder sound, participants
in all three conditions were given exactly 30 s to contemplate
listening to the sound for 58 s. Participants completed the same
measures of current emotion as in Experiment 3 (␣⫽.89). In the
anticipation and imagination conditions, participants also made the
same predictions as were made in Experiment 3 about what their
emotions would be while listening to the noise; participants in the
retrospection condition made the same retrospective ratings of
what their emotions actually were while listening to the noise (␣⫽
.62).
To measure the extent to which participants mentally simulated
listening to the noise, we asked them to rate their agreement (0 ⫽
disagree,6⫽agree) with seven statements: “When I think about
listening to the noise [in 20 min/20 min ago], I imagine what the
noise [will be/was] like”; “Right now I can hear the noise in my
head”; “It feels as though I am actually listening to the noise right
now”; “I am thinking about what my feelings [will be/were] like
while listening to the noise”; “I am thinking about how long the
noise [will last/lasted]”; “I am thinking of how unpleasant it [will
be/was] listening to the noise”; and “I am thinking about the
auditory properties of the noise—its tone, volume, and so on.”
These measures were generated on the basis of debriefing discus-
sions with participants in previous experiments about what it was
like to mentally simulate an emotional event. We averaged these
measures into a composite measure of mental simulation (␣⫽
.72). After completing these items, participants in the imagination
and retrospection conditions were thanked and debriefed. Partici-
pants in the anticipation condition completed unrelated question-
naires for exactly 20 min, at which time they actually listened to
the 58-s noise, after which they were thanked and debriefed.
3
Results
Current, predicted, and remembered emotion. In a replication
and extension of the results of previous experiments, participants
reported more intense emotion during anticipation and imagination
than during retrospection (see Table 5). We submitted participants’
reported emotion to a multiple linear regression with two orthog-
3
To measure participants’ reported online emotions while they were
listening to the noise, we asked participants in the anticipation and retro-
spection conditions to rate, while they actually listened to the 58-s noise,
how much they felt each of the 15 emotions used to measure their reported
emotion during anticipation of and retrospection about the noise (␣⫽.89).
These ratings did not differ by condition (M⫽4.34, SD ⫽0.99, and M⫽
4.04, SD ⫽0.91, for the anticipation and retrospection conditions, respec-
tively), t(54) ⫽1.21, SE ⫽0.25, p⫽.232, observed power ⫽.221, d⫽
0.33. Anticipating that they would listen to the annoying noise did not
significantly influence participant’s reported emotions while they actually
listened to the noise, compared with the reported emotions of participants
who did not anticipate listening to the noise.
Table 4
Experiment 4: Participants Reactions to Anticipation and
Retrospection About a Hypothetical Ski Vacation
Measure
Temporal perspective
Anticipation Retrospection
Reported current positive emotion 5.84 (0.92) 5.51 (0.63)
Predicted or remembered enjoyment 6.16 (1.19) 6.00 (1.09)
Note. The numbers in the table are means and, in parentheses, standard
deviations.
295
ANTICIPATION, RETROSPECTION, AND EMOTION

onally coded contrast variables (R
2
⫽.17), F(2, 80) ⫽9.10,
MSE ⫽0.90, p⬍.001. The first contrast indicated that partici-
pants in the anticipation and imagination conditions reported sig-
nificantly more intense negative emotion (Ms⫽3.85 and 3.79, and
SDs⫽0.73 and 1.13, respectively; weights ⫽1 each) than
participants in the retrospection condition (M⫽2.89, SD ⫽0.95;
weight ⫽–2; b⫽.311, ␤⫽.43), t(80) ⫽4.26, SE ⫽0.44, p⬍
.001, d⫽0.95. The second contrast indicated that participants’
current emotions in the anticipation and imagination conditions
were not significantly different from each other (weights ⫽1 and
–1, respectively; b⫽.03, ␤⫽.03), t(80) ⫽0.27, SE ⫽0.26, p⫽
.79, d⫽0.06. Thus, simply imagining an emotional event in the
future produced reported emotional arousal of an intensity that was
comparable with anticipating an actual future event—an arousal
that was significantly more intense than the arousal during retro-
spection about the actually experienced event.
There was also a tendency for participants in the anticipation
and imagination conditions to expect to experience more negative
emotion than participants in the retrospection condition remem-
bered experiencing (see Table 5). We conducted a multiple linear
regression that estimated participants’ predicted or remembered
feelings from two orthogonally coded contrast variables (R
2
⫽
.07), F(2, 80) ⫽2.77, MSE ⫽0.70, p⫽.069. The first contrast
indicated that participants in the anticipation and imagination
conditions predicted that they would experience significantly more
intense negative emotion (Ms⫽4.93 and 4.94, and SDs⫽0.68
and 0.89, respectively; weights ⫽1 each) than participants remem-
bered experiencing (M⫽4.46, SD ⫽0.91; weight ⫽–2; b⫽.151,
␤⫽.25), t(80) ⫽2.35, SE ⫽0.34, p⫽.021, d⫽0.53. The second
contrast indicated that participants’ predicted feelings in the antic-
ipation and imagination conditions were not significantly different
from each other (weights ⫽1 and –1, respectively; b⫽–.01, ␤⫽
–.01), t(80) ⫽0.08, SE ⫽0.23, p⫽.935, d⫽0.02. Thus,
participants expected to feel worse while listening to the noise in
the future than they recalled feeling in the past, and this difference
was statistically significant.
Mental simulation. As expected, participants reported men-
tally simulating listening to the noise more extensively during
anticipation and imagination than during retrospection (see Table
5). We conducted a multiple linear regression estimating partici-
pants’ reported mental simulation from two orthogonally coded
contrast variables (R
2
⫽.17), F(2, 80) ⫽8.37, MSE ⫽0.86, p⬍
.001. The first contrast indicated that participants reported men-
tally simulating listening to the noise significantly more exten-
sively during anticipation and imagination (Ms⫽4.54 and 4.19,
and SDs⫽0.94 and 0.91, respectively; weights ⫽1 each) than
during retrospection (M⫽3.55, SD ⫽0.92; weight ⫽–2; b⫽.27,
␤⫽.39), t(80) ⫽3.85, SE ⫽0.43, p⬍.001. The second contrast
indicated that participants’ reported mental simulations in the
anticipation and imagination conditions were not significantly
different from each other (weights ⫽1 and –1, respectively; b⫽
.18, ␤⫽.14), t(80) ⫽1.40, SE ⫽0.25, p⫽.167.
Structural analyses. Our theoretical analysis implies a partic-
ular correlational structure among the variables measured in this
experiment. Specifically, the effect on current emotions of antic-
ipation and imagination versus the effect of retrospection should
be statistically mediated by two variables: the extremity of pre-
dicted or remembered emotion and the degree of reported mental
simulation. We tested both mediation paths simultaneously with a
structural equation model.
We included five variables in the structural equation model: (a)
a contrast-coded variable indicating whether participants were in
the anticipation condition or the imagination condition versus the
retrospection condition (coded 1, 1, and –2, respectively); (b) an
orthogonally coded contrast variable indicating whether partici-
pants were in the anticipation condition versus the imagination
condition (coded 1 and –1, respectively); (c) the index of partici-
pants’ reported mental simulation; (d) the index of participants’
predicted or remembered emotions; and (e) the index of partici-
pants’ current emotions. The full model is displayed in Figure 1.
The regression weights and the pattern of significant and non-
significant paths in the model support our two mediation predic-
tions. First, consider participants’ predicted or remembered emo-
tional intensity. The variable indicating whether participants were
in the anticipation condition or the imagination condition rather
than the retrospection condition (anticipation and imagination vs.
retrospection) was significantly associated with their predicted or
Table 5
Experiment 5: Participants Reactions to Imagination,
Anticipation and Retrospection About Listening to an Annoying
Noise
Measure
Temporal perspective
Anticipation Imagination Retrospection
Reported current negative
emotion 3.85 (0.73) 3.79 (1.13) 2.89 (0.95)
Reported mental simulation 4.54 (0.94) 4.19 (0.91) 3.55 (0.92)
Predicted or remembered
emotion 4.93 (0.68) 4.94 (0.89) 4.46 (0.91)
Note. The numbers in the table are means and, in parentheses, standard
deviations.
Figure 1. A structural equation model of participants’ average reported
current emotion during anticipation of, imagination of, or retrospection
about listening to an annoying noise (Experiment 5: Imagining annoying
noises). Solid arrows represent statistically significant correlations ( p⬍
.05); dashed arrows represent nonsignificant correlations. Numbers are
standardized regression weights.
296 VAN BOVEN AND ASHWORTH

remembered emotion (Z⫽2.38, p⫽.017), which, in turn, was
significantly associated with participants’ current reported emotion
(Z⫽4.81, p⬍.001). A Sobel test indicated that this combined
path was significant (Z⫽2.14, p⫽.032). The extremity of
predicted emotion versus that of remembered emotion thus par-
tially mediated—at least in this experiment—the effect of antici-
pation and imagination on current reported emotion versus that of
retrospection on current reported emotion.
Next, consider participants’ reported mental simulation. The
variable indicating whether participants were in the anticipation
condition or the imagination condition rather than the retrospection
condition was significantly associated with their reported mental
simulation (Z⫽3.90, p⬍.001), which, in turn, was significantly
associated with participants’ current emotion (Z⫽2.36, p⫽.018).
A Sobel test confirmed that this combined path was significant
(Z⫽2.02, p⫽.043). The effect of anticipation and imagination
versus the effect of retrospection on mental simulation thus par-
tially mediated the effect of temporal perspective on current re-
ported emotion.
Two additional observations of the structural equation model are
noteworthy. First, there is no independent correlation between
reported mental simulation and predicted or remembered emotion.
The influence of temporal perspective on the extremity of pre-
dicted or remembered emotion therefore is not mediated by the
influence of temporal perspective on mental simulation (or vice
versa). Second, there is a significant independent correlation be-
tween the anticipation and imagination of emotions and the cor-
relation of retrospection and current emotions (Z⫽2.49, p⫽
.012). This suggests that the greater evocativeness of anticipation,
compared with that of retrospection, is only partly explained by
extreme expectations and asymmetric mental simulation, as mea-
sured in this experiment.
Discussion
These results yield three important conclusions. First, the imag-
ination of a hypothetical future event produced current emotion of
approximately equivalent intensity to the anticipation of an actual
future event— emotion that was significantly more intense than
that produced by retrospection about an actually experienced
event. This result, in conjunction with the results of Experiment 4,
highlights the importance of how people think about emotional
events as a reason why anticipation arouses more intense emotion
than retrospection.
Second, individuals who anticipated or imagined an emotional
event expected to experience more intense emotion than partici-
pants remembered experiencing, and these extreme expectations
partially mediated the effect of temporal perspective on current
emotion. The correlation between the extremity of predicted or
remembered emotion and the current emotional experience is
consistent with the results of the preceding four experiments,
although those earlier differences were not significant. Neverthe-
less, as we discuss shortly, a meta-analysis of the five experiments
suggests that extreme expectations may play a slight role in pro-
ducing more intense emotions during anticipation than during
retrospection.
Finally, participants reported more extensive mental simulation
during anticipation and imagination than during retrospection,
consistent with the possibility of an overgeneralized association
among emotional arousal, mental simulation, and temporal per-
spective. Moreover, this differential mental simulation partly me-
diated the more intense emotional reactions to anticipation and
imagination versus emotional reactions to retrospection. Taken
together, then, the results of Experiment 5 provide evidence for
two mediational processes that produce more intense emotions
during anticipation than during retrospection: a tendency for peo-
ple to expect to experience more extreme emotions than they
remember having experienced and a tendency for people to men-
tally simulate future events more extensively than they mentally
simulate past events.
General Discussion
The results of five experiments provide consistent evidence that
people report more intense emotions when they anticipate emo-
tional events rather than when they retrospect about them. The
greater evocativeness of anticipation than that of retrospection
emerged in positive events (Thanksgiving holidays, ski vacations),
negative events (menstruation, annoying noises), uncommon
events (annual Thanksgiving holiday), commonplace events (an-
noying noises, menstruation), and events of varying temporal
distance (from annoying noises that are 20 min away to holidays
that are 2 weeks away to ski vacations that are 6 months away).
People also reported more intense emotion during anticipation of,
than during retrospection about, purely hypothetical events (all-
expenses-paid ski vacation). These results suggest that the ten-
dency to report more intense emotions during anticipation than
during retrospection is robust and pervasive in everyday life.
Tentative Conclusions
At the outset, we described three discrepancies in people’s
thoughts about future events versus their thoughts about past
events that could produce more intense emotion during anticipa-
tion than during retrospection. Although each possibility may
influence the evocativeness of anticipation and retrospection in
everyday life, not all processes were equally likely in our experi-
ments. Let us take stock of the three propositions, given the
cumulative evidence.
One proposition was that people might experience more intense
emotion during anticipation than during retrospection because
future events are usually more uncertain than past events and
because uncertainty arouses emotion. We tentatively reject this
possibility on the basis of two observations. First, participants
were highly familiar with and routinely experienced their men-
strual periods in Experiment 2, so it seems unlikely that they were
meaningfully more uncertain about their next menstrual experi-
ence than they were about their last menstrual experience. Second,
care was taken in Experiments 3 and 5 to ensure that participants
knew exactly what the future annoying noise would be like, so it
seems unlikely that they were substantially more uncertain about
future noises than about past noises.
Another proposition was that people experience more intense
emotion during anticipation than during retrospection because they
predict that they will experience more intense emotions than they
remember having experienced. Although only significant in one of
five experiments, predicted emotions were always more extreme
than remembered emotions, as reflected by a significant test of
297
ANTICIPATION, RETROSPECTION, AND EMOTION

pooled significance (Z⫽2.92, p⫽.004). Further, the extremity of
predicted or remembered emotion was correlated in every exper-
iment with the reported intensity of current emotions. And the
significant difference between predicted and remembered emotion
in Experiment 5 partially mediated the difference between reported
emotion during anticipation and retrospection. These results lead
to the tentative conclusion that extreme expectations may explain
a small part of people’s more intense reactions to anticipation than
to retrospection.
The final proposition was that emotions are more intense during
anticipation than during retrospection because of people’s over-
generalized tendency to mentally simulate future events more
extensively than they simulate past events and because mental
simulation tends to amplify current emotions. Consistent with this
proposition, simply asking people to contemplate a hypothetical
ski vacation as though it were in the future produced more intense
emotion than contemplating a hypothetical vacation as though it
had occurred in the past (Experiment 4). The emotion aroused by
imagining a hypothetical future event was more intense, even, than
the emotion aroused by remembering an actually experienced
emotional event (Experiment 5). More important, participants re-
ported more extensive mental simulation during anticipation and
imagination of an emotional event than during retrospection about
an emotional event, and this mental simulation partially mediated
the effect of temporal perspective on current emotions. We there-
fore tentatively conclude that asymmetric mental simulation is at
least partly responsible for the greater evocativeness of anticipa-
tion, compared with that of retrospection.
Questions and Limitations
One question raised by our experiments concerns participants’
self-reported emotion. Did participants report the emotions they
actually experienced during anticipation and retrospection or did
they report their intuitive beliefs about their emotional reactions to
anticipation and retrospection? It is possible that because people
have limited introspective access to psychological processes (Nis-
bett & Wilson, 1977; Wilson, 1985), participants in Experiments 1,
2, and 4 answered our explicit questions about how thinking about
an emotional event made them feel on the basis of their beliefs
about how thinking about that event should make them feel. It is
also possible that participants correctly guessed our hypotheses
and that experimental demand led them to report more intense
emotion during anticipation than during retrospection. In light of
such concerns, future research might seek, in addition to self-
report measures, convergent measures of emotional experience
during anticipation and retrospection.
Still, three observations should assuage these concerns about the
validity of participants’ self-reported emotions. First, like other
emotion researchers (e.g., Lazarus, 1991; Schachter & Singer,
1962; Smith & Ellsworth, 1985; Zillmann, 1988), we view beliefs
about affective experience as integral to emotion— emotions are
inherently subjective interpretations, appraisals, or beliefs about
affective states. Second, participants in Experiments 3 and 5 were
simply asked to report their current emotions rather than asked to
report how anticipation or retrospection made them feel, so their
responses are unlikely to arise solely from intuitions or experi-
mental demand.
Finally, we conducted a follow-up experiment in which we
made extra efforts to separate the emotion measurement and the
temporal perspective manipulation in the eyes of participants. The
procedure was nearly identical to Experiment 3. Participants (31
males, 21 females; N⫽52) reported their current emotions after
listening to a short sample or reminder of a longer sound they
would hear 20 min in the future or had heard 20 min in the past.
Unlike Experiment 3, however, the emotion measure was embed-
ded in an ostensibly unrelated background questionnaire that was
developed to measure various factors that may or may not influ-
ence responses to other questionnaires. Participants were asked to
report, on a 5-point scale, how much (1 ⫽very slightly/not at all,
5⫽very much) they currently felt agitated, angry, annoyed,
irritated, negative, nervous, uncomfortable, unfortunate, unhappy,
and upset (␣⫽.86). Replicating our earlier results, we found that,
after a natural log transformation was applied to restore normality,
participants reported more intense emotion following anticipation
(M⫽0.30, SD ⫽0.31, back-transformed M⫽1.33) than follow-
ing retrospection (M⫽0.14, SD ⫽0.18, back-transformed M⫽
1.14), t(50) ⫽2.05, SE ⫽0.07, p⫽.046, d⫽0.58. And the
fraction of participants (42.86%) who, during debriefing, correctly
estimated whether or not the emotion measure was related to the
annoying noises was not significantly different from the 50%
expected if participants were guessing (95% confidence interval ⫽
29.11%, 57.81%).
4
The results of this follow-up experiment bolster our assumption
that participants reported how they actually felt rather than how
they believed anticipation or retrospection would make them feel.
Reports of one’s emotions are simply that—reports of how one
feels at the moment. They are not explicitly reports of how
thinking about future or past events makes one feel, and they are,
therefore, less prone to concerns about biased responses that are
due to intuitive theories or experimental demand. And the fact that
participants did not identify, above chance levels, the relation
between the temporal perspective manipulation and the emotion
measure casts doubt on the potential role of experimental demand.
A potential limitation of the present experiments is the scope of
our conceptual analyses and stimuli, which are constrained to the
kinds of moderate emotional events that people experience in
everyday life. It is not clear whether our analyses and findings
would generalize to more extreme emotional events, such as win-
ning the lottery, being diagnosed with leukemia, or losing a loved
one. Many extreme emotional events are often unexpected, incom-
prehensible, and personally transformative. Few people expect to
experience windfalls, be diagnosed with a deadly disease, or lose
a child—and life as a lottery winner, a terminally ill person, or a
bereaved person is likely to be unimaginably different from what
came before. The difficulty of anticipating such extreme events
may render them qualitatively different from the spectrum of
emotional events analyzed in this article.
4
Participants’ indication of whether the emotion measure was related to
the annoying noises did not interact with the effect of temporal perspective
on current emotion, F(1, 48) ⫽0.96, MSE ⫽0.19, p⫽.33, observed
power ⫽.160.
298 VAN BOVEN AND ASHWORTH

Implications
The fact that current emotions are more intense during antici-
pation than during retrospection has at least four implications for
psychological science and practice. One is methodological. As
noted earlier, researchers sometimes manipulate emotional states
by asking people to retrospect about emotional events and some-
times manipulate emotions by asking people to anticipate emo-
tional events. The results of our experiments imply that manipu-
lating emotion through anticipation may be the more effective
method.
A second implication is clinical. Individuals often suffer from
debilitating anticipatory anxiety about future stressful experiences
(Hinrichsen & Clark, 2003; McCroskey, 1970; Vassilopoulos,
2004). Our results suggest that encouraging individuals to reframe
their future emotional experience in the past tense—looking back
on the future experience—might alleviate such anxiety.
A third implication concerns the valuation of future and past
experiences. People often use the emotions aroused by anticipating
future experiences as a basis for valuing those experiences (Hsee
& Rottenstreich, 2004; Slovic, Finucane, Peters, & MacGregor,
2002). For instance, people might deem a potential beach holiday
that arouses greater pleasure during anticipation as more valuable
than a holiday that arouses less pleasure during anticipation. Peo-
ple might also use the emotions aroused by retrospection about
past experiences as basis for valuing those experiences, as when
they deem a holiday that produces greater pleasure during retro-
spection as more valuable than a holiday that produces less plea-
sure during retrospection. Combined with the present findings,
such valuation by feelings (Hsee & Rottenstreich, 2004) suggests
that people might systematically value future outcomes more than
they value past outcomes. The same beach holiday may be more
valuable in foresight than in hindsight.
A final implication concerns subjective well-being. Several re-
searchers have suggested that the enjoyment people glean from
retrospection is an important component of life satisfaction (Ar-
gyle, 2002; Chang, 2004; Diener, Suh, Lucas, & Smith, 1999; Van
Boven & Gilovich, 2003). Our research suggests that the enjoy-
ment people glean from anticipation might also be an important
component of life satisfaction: one’s satisfaction with life is influ-
enced both by looking backward and by looking forward.
Conclusion
The importance of temporal distance has reappeared and is in
vogue among psychological scientists, owing largely to research
highlighting the importance of temporal distance for construal
(Trope & Liberman, 2003) and intertemporal choice (e.g., Loe-
wenstein, 1996; Metcalfe & Mischel, 1999). The present experi-
ments indicate that when it comes to emotion, people are sensitive
not only to temporal distance but also to temporal perspective. The
psychological distance between the self and a future emotional
event is almost always decreasing, whereas the psychological
distance between the self and a past emotional event is almost
always increasing. People’s cognitions and emotions therefore
critically depend on both the distance between and the direction
from the self and emotional events in psychological space. Emo-
tional experience, in other words, depends not only on whether an
emotional event is near or far but also on whether one is looking
forward or looking back.
References
Argyle, M. (2002). The psychology of happiness (2nd ed.). New York:
Routledge.
Buhr, K., & Dugas, M. J. (2002). The Intolerance of Uncertainty Scale:
Pscychometric properties of the English version. Behaviour Research
and Therapy, 40, 931–946.
Chang, E. C. (2004). Distinguishing between ruminative and distractive
responses in dysphoric college students: Does indication of past depres-
sion make a difference? Personality and Individual Differences, 36,
845– 855.
Diener, E., Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective
well-being: Three decades of process. Psychological Bulletin, 125, 276 –
302.
Elster, J., & Loewenstein, G. (1992). Utility from memory and anticipation.
In J. Elster & G. Loewenstein (Eds.), Choice over time (pp. 213–224).
New York: Russell Sage Foundation.
Eng, W., Coles, M. E., Heimberg, R. G., & Safren, S. A. (2005). Domains
of life satisfaction in social anxiety disorder: Relation to symptoms and
response to cognitive– behavioral therapy. Journal of Anxiety Disorders,
19, 143–156.
Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a process:
Study of emotion and coping during three stages of a college examina-
tion. Journal of Personality and Social Psychology, 48, 150 –170.
Frijda, N. H. (1988). The laws of emotion. American Psychologist, 43,
349 –358.
Frijda, N. H., Kuipers, P., & ter Schure, E. (1989). Relations among
emotion, appraisal, and emotional action readiness. Journal of Person-
ality and Social Psychology, 57, 212–228.
Hambrick, J. P., Turk, C. L., Heimberg, R. G., Schneier, F. R., & Liebo-
witz, M. R. (2003). The experience of disability and quality of life in
social anxiety disorder. Depression and Anxiety, 18, 46 –50.
Hinrichsen, H., & Clark, D. M. (2003). Anticipatory processing in social
anxiety: Two pilot studies. Journal of Behavior Therapy and Experi-
mental Psychiatry, 34, 205–218.
Holmes, E., & Mathews, A. (2005). Mental imagery and emotion: A
special relationship? Emotion, 5, 489 – 497.
Hsee, C. K., & Rottenstreich, Y. (2004). Music, pandas, and muggers: On
the affective psychology of value. Journal of Experimental Psychology:
General, 133, 23–30.
Johnson, M. K., & Sherman, S. J. (1990). Constructing and reconstructing
the past and the future in the present. In E. T. Higgins & R. M.
Sorrentino (Eds.), Handbook of motivation and cognition: Foundations
of social behavior (Vol. 2, pp. 482–526). New York: Guilford Press.
Kross, E., Ayduk, O., & Michel, W. (2005). When asking “why” does not
hurt. Psychological Science, 16, 709 –715.
Lazarus, R. S. (1991). Emotion and adaptation. New York: Oxford Uni-
versity Press.
Leary, M. R., & Kowalski, R. M. (1995). Social anxiety. New York:
Guilford Press.
Lerner, J., & Keltner, D. (2001). Fear, anger, and risk. Journal of Person-
ality and Social Psychology, 81, 146 –159.
Levenson, R. W., Carstensen, L. L., & Gottman, J. M. (1994). The
influence of age and gender on affect, physiology, and their interrela-
tions: A study of long-term marriages. Journal of Personality and Social
Psychology, 67, 56 – 68.
Lindsay, D. S., Wade, K. A., Hunter, M., & Read, J. D. (2004). Adults’
memories of childhood: Affect, knowing, and remembering. Memory,
12, 27– 43.
Loewenstein, G. (1996). Out of control: Visceral influences on behavior.
Organizational Behavior and Human Decision Processes, 65, 272–292.
Mauss, I. B., Wilhelm, F. H., & Gross, J. J. (2004). Is there less to social
anxiety than meets the eye? Emotion experience, expression, and bodily
responding. Cognition and Emotion, 18, 631– 662.
299
ANTICIPATION, RETROSPECTION, AND EMOTION

McCroskey, J. C. (1970). Measures of communication-bound anxiety.
Speech Monographs, 37, 269 –277.
McFarland, C., Ross, M., & DeCourville, N. (1989). Women’s theories of
menstruation and biases in recall of menstrual symptoms. Journal of
Personality and Social Psychology, 57, 522–531.
Mellers, B., Schwartz, A., Ho, K., & Ritov, I. (1997). Decision affect
theory: Emotional reactions to the outcomes of risky options. Psycho-
logical Science, 8, 423– 429.
Metcalfe, J., & Mischel, W. (1999). A hot/cool-system analysis of delay of
gratification: Dynamics of willpower. Psychological Review, 106, 3–19.
Mitchell, T. R., Thompson, L., Peterson, E., & Cronk, R. (1997). Temporal
adjustments in the evaluation of events: The “rosy view.” Journal of
Experimental Social Psychology, 33, 421– 448.
Monat, A., Averill, J. R., & Lazarus, R. S. (1972). Anticipatory stress and
coping reactions under various conditions of uncertainty. Journal of
Personality and Social Psychology, 24, 237–253.
Neese, R. M. (1990). Evolutionary explanations of emotions. Human
Nature, 1, 261–289.
Niedenthal, P. M., Barsalou, L. W., Winkielman, P., Krauth-Gruber, S., &
Ric, F. (2005). Embodiment in attitudes, social perception, and emotion.
Personality and Social Psychology Review, 9, 184 –211.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know:
Verbal reports on mental processes. Psychological Review, 84, 231–259.
Papageorgiou, C., & Wells, A. (2005). Depressive rumination: Nature,
theory and treatment. Chichester, England: Wiley.
Parfit, D. (1984). Reasons and persons. Oxford, England: Clarendon Press.
Schachter, S., & Singer, J. E. (1962). Cognitive, social, and physiological
determinants of emotional state. Psychological Review, 69, 379 –399.
Schkade, D. A., & Kahneman, D. (1998). Does living in California make
people happy? A focusing illusion in judgments of life satisfaction.
Psychological Science, 9, 340 –346.
Schwarz, N. (2002). Feelings as information: Moods influence judgments
and processing strategies. In T. Gilovich, D. Griffin, & D. Kahneman
(Eds.), Heuristics and biases: The psychology of intuitive judgment (pp.
534 –547). New York: Cambridge University Press.
Slovic, P., Finucane, M., Peters, E., & MacGregor, D. (2002). The affect
heuristic. In T. Gilovich, D. Griffin, & D. Kahneman (Eds.), Heuristics
and biases: The psychology of intuitive judgment (pp. 397– 420). New
York: Cambridge University Press.
Smith, C., & Ellsworth, P. (1985). Patterns of cognitive appraisal in
emotion. Journal of Personality and Social Psychology, 77, 548 –565.
Soman, D. (2003). Prospective and retrospective evaluations of experi-
ences: How you evaluate an experience depends on when you evaluate
it. Journal of Behavioral Decision Making, 16, 35–52.
Strack, F., Schwarz, N., & Gschneidinger, E. (1985). Happiness and
reminiscing: The role of time perspective, affect, and mode of thinking.
Journal of Personality and Social Psychology, 49, 1460 –1469.
Trope, Y., & Liberman, N. (2003). Temporal construal. Psychological
Review, 110, 403– 421.
Tulving, E. (2002). Episodic memory: From mind to brain. Annual Review
of Psychology, 53, 1–25.
Van Boven, L., & Gilovich, T. (2003). To do or to have? That is the
question. Journal of Personality and Social Psychology, 85, 1193–1202.
van de Bos, K. (2001). Uncertainty management: The influence of uncer-
tainty salience on reactions to perceived procedural fairness. Journal of
Personality and Social Psychology, 80, 931–941.
Vassilopoulos, S. (2004). Anticipatory processing in social anxiety. Be-
havioural and Cognitive Psychotherapy, 32, 303–311.
Wilson, T. D. (1985). Strangers to ourselves: The origins and accuracy of
beliefs about one’s own mental states. In J. H. Harvey & G. Weary
(Eds.), Attribution: Basic issues and applications (pp. 9 –36). Orlando,
FL: Academic Press.
Wilson, T. D., Ceterbar, D. B., Krermer, D. A., & Gilbert, D. T. (2005).
The pleasures of uncertainty: Prolonging positive moods in ways people
do not anticipate. Journal of Personality and Social Psychology, 88,
5–21.
Wilson, T. D., & Gilbert, D. T. (2003). Affective forecasting. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 35,
pp. 345– 411). San Diego, CA: Academic Press.
Wilson, T. D., Wheatley, T., Meyers, J. M., Gilbert, D. T., & Axsom, D.
(2000). Focalism: A source of the durability bias in affective forecasting.
Journal of Personality and Social Psychology, 78, 821– 836.
Wirtz, D., Kruger, J., Scollon, C. N., & Diener, E. (2003). What to do on
spring break? The role of predicted, on-line, and remembered experience
in future choice. Psychological Science, 14, 520 –524.
Zillmann, D. (1988). Cognitive excitation interdependencies in aggressive
behavior. Aggressive Behavior, 14, 51– 64.
Received June 26, 2006
Revision received October 2, 2006
Accepted October 3, 2006 䡲
300 VAN BOVEN AND ASHWORTH