Affective Forecasting:
How Happy or Sad Will You Be?
Lectures 11 & 12:
Affective Forecasting
Wilson, T.D., & Gilbert, D.T. (2003). Affective forecasting.
Advances in Experimental Social Psychology, 35, 345-411.
Straight From the Horse’s Mouth
The Power of a Crystal Ball:
Predicting the Future (fat wallets & hot dates)
• Self-Projection
where will I be?
what will I be doing?
who will I be doing it with?
• Decision Making
uncertainty (Kahneman & Tversky, 2000)
accuracy (Osberg & Shrauger, 1986)
temporal perspective (Trope & Liberman, 2004)
biases in prediction (Armor & Taylor, 1998)
• What’s missing?
anticipating events vs. anticipating feelings
Forecasting Feelings:
How Happy Will I Be?
How much happiness will events bring?
How long will happiness last?
Will owning a sportscar make me happy?
Will a cool drink be more refreshing than an
ice-cream?
Marrying which twin will bring
happiness in the long run?
more
AFFECTIVE FORECASTING (measuring
predicted and experienced emotional responses - accuracy)
Predicting One’s Emotions:
Types of Affective Forecasts and Errors
• Affective Forecasts - 4 components
(1) predictions about the valence of one’s
feelings
(2) specific emotions that will be experienced
(3) intensity of the emotions
(4) duration of emotions
Predicting Valence
•
Are people mistaken about the valence of future
events?
Wilson et al. (2002) - staged a dating game in which
students competed with a same-sex student for a
hypothetical date with an opposite-sex student.
Experiences were randomly assigned to win or lose
the date, after which they rated their mood.
Forecasters estimated what their mood would be if
they won or lost the date. All forecasters estimated
that they would be in a better mood if they won than
lost (which mirrored the actual judgments of the
experiencers). However, forecasters overestimated
how positive or negative they would feel.
Predicting Specific Emotions
•
While people are quite accurate at forecasting what
they will feel - when exactly an event will take place
shapes their reactions.
Liberman et al. (2002) - noted that people have overly
simplistic reactions to emotional events when
thinking about the distant (compared to the near)
future.
good day tomorrow (positive events with a few
negative occurrences)
good day in a year (people only report positive
events)
Thus, people’s forecasts may be more realistic for events
that will happen soon, but onverly simplistic for
events far in the future.
Predicting Intensity and Duration
I’ve won the lottery.
How happy will I feel?
For how long will I be happy?
I’ve dumped my partner.
How sad will I feel?
How long will the pain last?
Some Initial Observations:
Duration, Duration, Duration
Durability Bias
People have a tendency to overestimate the duration of
their future emotional reactions (Gilbert et al., 1998)
How happy will you feel the day after your favourite
team wins an important game?
people overestimate their happiness (Wilson et al., 2000)
- but what is biased (i) estimation of happiness, (ii)
reduction in happiness over time, or (iii) both?
Durability Reconsidered
Impact Bias
People have a tendency to overestimate the enduring
impact that future events will have on their emotional
reactions (Gilbert, Driver-Linn, & Wilson, 2002).
overestimate intensity
underestimate rate of dissipation
Exploring the Impact Bias
impact bias = most prevalent forecasting error
people overestimate the impact of future events on
their emotional reactions
People
(college students, professors, sports fans, dieters,
holiday makers, snake phobics, medical test takers)
Events
(romantic breakups, personal insults, sports victories,
electoral defeats, failure to lose weight, results of
pregnancy tests)
So why does the ‘impact bias’ occur? Why does affective
forecasting go awry?
Process of Affective Forecasting
Wilson & Gilbert (2002)
Affective Forecasting:
Sources of Error
• Construal - bringing the event to mind!
traffic jam vs. getting married (having a baby)
The problem of misconstrual - people mistakenly
imagine the wrong event.
How will your wedding day unfold?
romantic bliss vs. stressful, fight with in-laws, food
poisoning.
The future doesn’t always match our expectancies?
1. Misconstrual
•
Woodzicka and LaFrance (2001) asked women to
predict how they would react if asked sexually
harassing questions (3) during a job interview (and
compared these reactions with the behaviour of
women who experienced such questions).
68% of forecasters said they would refuse to answer at
least 1 of the 3 questions; 28% said they would
confront the interviewer
This was completely at odds with the behaviour of the
experiencers. Why? The forecasters imagined a
different situation than the one confronted by the
experiencers (i.e., not always easy to approach and
confront a harasser).
Misconstruals of future situations provide the greatest latitude of affective forecasting
errors (as there is no limit to how inaccurate people’s construals can be)
2. Framing Effects
Representations of events also depend on the way in
which people frame them, such as the particular
aspects of events that capture their attention.
Isolation Effect - people disregard components that
alternatives share and focus on components that
distinguish between them (Kahneman & Tversky,
1979)
What will make me most happy…a holiday by the ocean
or a weekend break in a large city?
sometimes shared features are important!
Location, Location, Location
Dunn, Wilson & Gilbert (2003)
asked college students to forecast what their overall
level of happiness would be the following year if they
lived in various campus dorms (random assignment to
dorms)- people’s forecasts were much more a
function of their ratings of the physical features
(which varied considerably) than the common social
features (good relationships with other members of
the dorm).
this resulted in a strong impact bias, by focusing too
much on a variable that distinguished the dorms,
people overestimated the effect of dorm assignment
on happiness.
predicted
undesirable dorm 3.43
actual
5.37
desirable dorm
5.45
5.96
Consequences of Framing
Framing effects would produce an impact bias if people
focus their attention on features that they think will
influence their emotional states but that actually will be
of little importance.
Dunn, Wilson & Gilbert (2003)
So, construal and framing. Any other important factors?
3. Recall and Affective Theories
Imagine that misconstrual and framing effects have been
avoided, the next important step is bringing to mind an
accurate representation of the future event.
previously experienced events (e.g., root canal work) = how
did it make me feel in the past?
So how accurate is memory for past emotional experiences?
What do We Remember?
Memories are reconstructed not replayed. People can
remember that root canal treatment is painful, but the
pain itself is not stored in memory in a form that can
be retrieved later.
Instead of replaying past emotions, people recall details
of an experience and have emotional reactions to
these memories (e.g., the sound of the drill, the pizza
in Venice).
As such, there is no guarantee that the feelings evoked by
these memories are the same as the feelings they
originally experienced.
Episodic details disappear, theories take over (Robinson
& Clore, 2002)
Mood and Menstruation
McFarland, Ross and DeCourville (1989) noted that many
women hold the theory that they are in worse moods
during menstruation and recall being in bad moods
during their periods.
however, when asked to rate their mood on a daily basis
for several weeks, these women were in a no worse
mood when they were menstruating than when they were
not.
Thus, people’s recall of their emotional experiences is biased
in systematic ways, prompting errors in affective
forecasting (intensity & duration).
4. Correction for Unique Influences
A basic problem with assessing our affective reaction to an
event (e.g., getting divorced) and deciding how likely we
are to have the same reaction in the future is that the
circumstances under which people make affective
forecasts are almost always different from the
circumstances under which they will actually experience
an event.
As such, people must subtract out several potential sources
of bias on their current assessments of their feelings (i.e.,
correction for unique influences).
Projection Bias
Imagine you have flu and are trying to decide whether to go
to a party in 2 weeks time. Typically people’s current
emotional state taints their predictions about how they
will feel in the future.
Projection Bias – tendency for people to under-appreciate
the effects of changes in their states, hence falsely
project their current preferences (and feelings) onto their
future preferences (and feelings).
Loewenstein et al. (1999)
this bias is an instance of mental contamination – people’s
judgments, emotions and behaviours are influenced in
unwanted ways (Wilson & Brekke, 1994).
Mental Decontamination is Tricky
why is decontamination tricky (Wilson & Brekke, 1994)?
awareness of bias
knowledge of direction of bias
motivation to correct
ability to correct
shoppers who have not eaten fail to take this into account
when in the supermarket (Gilbert et al., 2002).
how much would you enjoy eating spaghetti tomorrow
morning or evening? Under cognitive load, hungry
participants fail to adjust their judgments and report that
spaghetti for breakfast would be very enjoyable (Gilbert
et al., 2002)
Thus, inadequate correction can lead to a range of
forecasting errors.
5. Expectation Effects (Assimilation & Contrast)
Expectation effects occur when people’s affective
forecasts change their actual emotional experience.
Imagine going to a movie believing it will be the best
comedy of the year or just another film.
assimilation = people who expect the movie to be
good like it more.
contrast = people who expect to like the movie enjoy
it less.
Triggering Assimilation
Assimilation occurs when people’s expectations are not
too discrepant from their experiences (i.e., the
movie is not quite as good as people expected –
Wilson & Klaaren,1994)
Wilson et al. (1989) showed participants 6 cartoons, 3
of which were relatively funny and 3 of which were
not. When people viewed the cartoons with no
expectations about how funny they would be they
noticed the discrepancy and rated the first 3 as
funnier than the last 3. When people were told that
previous participants had found all 6 cartoons to be
funny they showed evidence of assimilation (i.e.,
the found the last 3 cartoons to be as funny as the
first 3).
6. Unique Influences on Actual Emotional
Experience – Hot/Cold Mind Sets
Just as there are unique influences on people’s
assessments of their emotional reactions to events
when making an affective forecast, so too there are
unique influences on their actual emotional
experiences – unique in the sense that they influence
people’s emotions but not their forecasts, leading to a
discrepancy between the two.
Hot vs. Cold Emotional States (Loewenstein et al., 1998)
Fancy a Twix?
Loewenstein et al. (1998) approached visitors to a museum, gave
them an 11-item trivia quiz, and asked them to choose either a
candy bar or the answers to the questions as compensation.
‘Cold’ State (choose before taking the test) – 79% candy bar
‘Hot’ State (choose after test) – 40% candy bar
‘Forecasters’ (what would be choose after test) – 62% candy bar
People in cold states have difficulty imagining what they will
prefer in hot states.
7. Unique Influences on Actual Emotional
Experience –Focalism
In addition to failing to anticipate unique influences on
their emotional reactions to an event, people often fail
to anticipate the extent to which unrelated events will
influence their thoughts and emotions, a tendency that
has been termed focalism (Wilson et al., 2000).
Events do not occur in a vacuum, but rather in the context
of other happenings that shape people’s reactions. By
neglecting to consider how much these other events
will capture their attention and influence their
emotions, people overestimate the impact of the focal
event.
Exploring Focalism
Wilson et al. (2000) required college football fans (2
months before a game) to report how happy they
would feel (immediately after the game) if their team
won. The day after the game (which the team won)
they rated their happiness. Participants displayed a
strong impact bias, such that they predicted that they
would be above their baseline level of happiness. In
reality, there were no happier than usual.
before making predictions, a second group of
participants were asked to imagine a day in the future
and report the number of hours they would spend
performing 10 different activities (e.g., going to class,
socializing). It was anticipated that when these people
rated how happy they would be after the game, the
salience of other events would impact their judgments
of happiness. The results revealed exactly this effect,
such that participants predicted that the game would
have less impact on their levels of happiness (i.e.,
correction for focalism)
8. Sense Making Processes
Arguably the most important source of affective
forecasting errors involve psychological processes
that moderate people’s emotional reactions.
birth of child
lottery win
job loss
People’s emotional reactions to events become less
intense with time (Wilson et al., 2002).
emotional evanescence
A major source of impact bias is that people fail to
anticipate the extent to which psychological factors
will ameliorate the impact of significant events.
The Emotional Implications of Sense Making
People make senses of their worlds in a way that speeds
recovery from emotional events and sense making is
largely automatic and non-conscious.
Sense Making and Emotional Evanescence – 4 steps
1. people orient to unexpected but relevant
information (i.e., goal-relevant).
2. people have more intense emotional reactions to
unexpected, relevant information than to other events.
3. once an unexpected event occurs and people have
an intense emotional reaction, they attempt to make
sense of the event (i.e., predictable world).
4. when people make sense of an event, it no longer
seems unexpected and surprising (i.e., less emotional)
Failure to Anticipate Sense-Making Processes:
Positive Events
People fail to anticipate how much they will transform
events psychologically in ways that reduce their
emotional power (extraordinary becomes ordinary).
ordinization neglect
Does a positive event produce lasting happiness?
Gilbert et al. (1998) noted that a positive tenure decision
did not cause the lasting happiness that untenured
professors had anticipated.
Psychological Immune System:
Making Sense of Bad Stuff
People possess powerful psychological defenses that
serve to ameliorate the impact of negative emotional
experiences.
They are especially motivated to interpret negative events
in ways that minimize their impact.
Like other sense making processes, the psychological
immune system occurs largely outside of awareness.
Failure to Anticipate Sense-Making Processes:
Negative Events
A major form of impact bias in response to negative
events is that people fail to anticipate how much their
psychological immune system will hasten their
recovery.
immune neglect
Does a negative life event produce lasting misery?
Gilbert et al. (1998) noted that people predicted they
would be equally unhappy if they received negative
personality feedback from two clinical psychologists
who had examined their test results or from a
computer program. In reality, people who received
feedback from the computer were less unhappy (i.e.,
rationalization…”what does a computer know about
people?”).
Things Worth Knowing
1.
What is affective forecasting?
2.
How and why does affective forecasting go awry?