Title: Risk As Analysis and Risk As Feelings
1Risk As Analysis and Risk As Feelings An
Examination of Affect, Reason, Risk, and
Rationality Paul Slovic Decision Research 1201
Oak Street Eugene, OR 97401 Conference on
Social Contexts and Responses to Risk University
of Kent January 28, 20005
2(No Transcript)
3Neural Economics P. Read Montague (Baylor
College of Medicine) Survival is about
economic evaluation The brain is an
economic evaluation engine The core of
neural economics is the recognition that rapid,
ongoing economic evaluation is a central
function carried out by the nervous systems of
mobile creatures. Without some kind of
internal currency in the nervous system, a
creature would be unable to assess the relative
value of events like drinking water, searching
for predators, or chasing prey. The nervous
system must estimate the value of each of these
actions and convert it to a common scale. Recent
work has shown that fluctuations in the delivery
of the neurotransmitter dopamine may represent
one such common currency. The dopamine
system is well designed to handle the kinds of
decisions that an early human would have
encountered however it is grossly inadequate in
many situations that have arisen in modern
society e.g. drug abuse.
4At the psychological levelthis common
currencymay be affect
- A valenced quality (e.g., goodness
- or badness) associated
- with a stimulus
5(No Transcript)
6Information
Affect
Meaning
- Affect conveys meaning upon information
- Without affect, information lacks meaning and
will not be used in judgment and decision making - Affect is a key ingredient of rational behavior
- Affect sometimes leads to poor decision making
7There are 1,198,500,000 people alive now in
China. To get a feel for what this means, simply
take yourself in all your singularity,
importance, complexity, and love and multiply
by 1,198,500,000. See? Nothing to it. -Annie
Dillard, For the Time Being (1999)
8Affect Response Probability Frequency Rel
ative Frequency Risk Benefit Valuation
(buying and selling prices) Choice/Preference
9Phineas Gage
10Phineas Gage
11(No Transcript)
12The strategies of human reason probably did not
develop . . . without the guiding force of the
mechanisms of biological regulation, of which
emotion and feeling are notable expressions.
Moreover, even after reasoning strategies become
established their effective deployment probably
depends . . . on a continued ability to
experience feelings. Antonio Damasio 1994,
p.xii
13- The emotions are of quite extraordinary
importance in the total economy of living
organisms and do not deserve being put into
opposition with intelligence. The emotions
are, it seems, themselves a high order of
intelligence. - O. Hobart Mowrer
- Learning Theory and Behavior
- (1960, p. 308)
14Images ?Affect ?Behavior
- Converging Views
- Psycholinguistics (Osgood 1952)
- Connotative meaning of words dominated
- by evaluative (good/bad) dimension.
- Learning Theory (Mowrer 1960)
- Behavior is guided and controlled by
- conditioned emotional reactions to images.
- Neurology (Damasio 1994)
- Thought is made largely from images,
- broadly construed to include perceptual and
- symbolic representations. Through learning,
- images become marked by positive and
- negative feeling states (Somatic Markers).
- A negative SM linked to image alarm
- A positive SM is a beacon of incentive
15There is no dearth of evidence in everyday life
that people apprehend reality in two
fundamentally different ways, one variously
labeled intuitive, automatic, natural,
non-verbal, narrative, and experiential, and the
other analytical, deliberative, verbal, and
rational. Seymour Epstein 1994, p. 710
16(No Transcript)
17Image Word Sound Smell Tactile sensation Memory Im
agination Visual Image
18(No Transcript)
19(No Transcript)
20HATE
21prime ideograph liking judgment advance slides
smile or frown degree of liking 1-6 scale
degree of liking 1-6 scale
neutral polygon or blank slide degree of liking 1-6 scale
Subliminal 4 milliseconds Supraliminal 2 seconds 8 seconds
Time Time Time Time
FIG 1. Sequence of events on one trial during liking judgment task. Source Winkielman, Zajonc, Schwarz (1997) FIG 1. Sequence of events on one trial during liking judgment task. Source Winkielman, Zajonc, Schwarz (1997) FIG 1. Sequence of events on one trial during liking judgment task. Source Winkielman, Zajonc, Schwarz (1997) FIG 1. Sequence of events on one trial during liking judgment task. Source Winkielman, Zajonc, Schwarz (1997)
22Heart and Mind in ConflictShiv Fedorikhin
(1999)
- Room 1
- memorize a number
- 2 digits or 7 digits
choose a snack chocolate cake or fruit salad
Room 2 Report the number that was memorized
7-digit group 63 chose cake
2-digit group 41 chose cake
23Valuation By Feelings Valuation By Calculation
- Reliance on Feelings Increases With
- Innumeracy
- Cognitive Load
- complexity of task information
- amount of information
- memory demands
- Stress
- time pressure
- pain
- poor health
- Older age
- Affect rich outcomes images
- evaluable information displays
24- Some Elements of the Affect Story
- Evaluating gambles
- Destination preferences (vacations, jobs,
retirement) - Risk perception/communication (dangerousness)
- Marketing/advertising/promotion
- names, labels, images, packaging
- Protective measures/insurance/life saving
- Stigmatization of places, products, technologies
- Investment judgments and decisions
- Punitive damage awards
- Youth smoking
25Evaluability Chris Hsee (University of
Chicago) To say that an attribute is hard to
evaluate . . . means that people do not know
whether a given value on the attribute is good or
bad . . .
26Attributes of Two Dictionaries in Hsees Study
Year of publication Number of entries Any defects?
Dictionary A 1993 10,000 No, its like new
Dictionary B 1993 20,000 Yes, the cover is torn otherwise its like new
Source Adapted from Hsee (1998)
27Affect, Evaluability and Attractiveness of Simple
Gambles
- Mean attractiveness
- (0-20 scale)
- 9.4
- 14.3
- Gamble 1. 7/36 chance to win 9
- Otherwise win nothing
- Gamble 2. 7/36 chance to win 9
- 29/36 chance to lose 5
- Attractiveness is determined by probability in
Gamble 1. The 9 payoff is not evaluable and
carries little weight. - In gamble 2, 9 comes alive with feeling as a
very attractive outcome compared to the small
loss. It then carries weight in the judgment.
28Results April 2003
- 9 is most positive when paired with the 5
loss (Condition B) - the 5 loss is rated more positively than 5
and 0 - Affect for 9 predicts attractiveness ratings
but highest predictability is in Condition B - Affect for the other outcome predicts
attractiveness only in B (-5) liking for 5
predicts in positive direction
29(No Transcript)
30Numeracy (from the cancer literature)
- The ability to understand and use basic
probability and mathematical concepts (Lipkus et
al., 2001) - New studies with Ellen Peters
31Numeracy in UO undergraduate sample
Table 1. Eleven items in the Numeracy Scale (Lipkus et al., 2001). Mean (median) score in UO undergraduate sample 8.2 (8) out of 11 possible Table 1. Eleven items in the Numeracy Scale (Lipkus et al., 2001). Mean (median) score in UO undergraduate sample 8.2 (8) out of 11 possible Table 1. Eleven items in the Numeracy Scale (Lipkus et al., 2001). Mean (median) score in UO undergraduate sample 8.2 (8) out of 11 possible Table 1. Eleven items in the Numeracy Scale (Lipkus et al., 2001). Mean (median) score in UO undergraduate sample 8.2 (8) out of 11 possible
Items correct UO undergrads correct older adults
1. Imagine that we roll a fair, six-sided die 1,000 times. Out of 1,000 roles, how many times do you think the die would come up even (2, 4, or 6)? 61 --
2. In the BIG BUCKS LOTTERY, the chances of winning a 10.00 prize are 1. What is your best guess about how many people would win a 10.00 prize if 1,000 people each buy a single ticket from BIG BUCKS? 65 --
3. In the ACME PUBLISHING SWEEPSTAKES, the chance of winning a car is 1 in 1,000. What percent of tickets of ACME PUBLISHING SWEEPSTAKES win a car? 42 --
4. Which of the following numbers represents the biggest risk of getting a disease? 1 in 100, 1 in 1000, 1 in 10 93 76
5. Which of the following represents the biggest risk of getting a disease? 1, 10, 5 95 79
6. If Person As risk of getting a disease is 1 in ten years, and Person Bs risk is double that of As, what is Bs risk? 78 66
7. If Person As chance of getting a disease is 1 in 100 in ten years, and person Bs risk is double that of A, what is Bs risk? 70 57
8/9. If the chance of getting a disease is 10, how many people would be expected to get the disease out of 100?, out of 1000? 90, 86 72, 67
10. If the chance of getting a disease is 20 out of 100, this would be the same as having a ____ chance of getting the disease. 85 --
11. The chance of getting a viral infection is .0005. Out of 10,000 people, about how many of them are expected to get infected? 51 --
32The ability to understand numbers will influence
- The precision of feelings about probabilistic
information. - The extent to which irrelevant affect influences
choices. - The accessibility and influence of nonsalient
frames of the same number. - The extent of affect elicited through a
comparison of numbers.
33High (and not low) numerate drive the surprising
effect and find the objectively worse bet more
attractive
34High numerate feel more positive about 9 when 5
cent loss is present
Interaction F(1,40)4.4, plt.05 main effect of
numeracy is also significant
35Risk As Analysis vs. Risk as Feelings
Analytic/ Deliberative
Experiential/ Affective
36Street Calculus
By Garry Trudeau
37In the world, risk and benefit are positively
correlated. In peoples minds, they are
negatively correlated.
38Relationship between risk and benefit in peoples
minds
39The strength of the inverse (negative)
relationship between risk and benefit judgments
for a particular hazard (e.g. nuclear power)
depends on the degree to which that hazardous
activity is judged to begood or bad.
40Radiation
5 4 3 2 1
Benefit
Risk
Risk
Benefit
Nuclear Power
X-rays
Chemicals
5 4 3 2 1
Benefit
Risk
Risk
Benefit
Pesticides
Prescription Drugs
Figure 3. Mean perceived risk and perceived
benefit for medical and nonmedical sources of
exposure to radiation and chemicals. Each item
was rated on a scale of perceived risk ranging
from 1 (very low risk) to 7 (very high risk) and
a scale of perceived benefit ranging from 1 (very
low benefit) to 7 (very high benefit). Note that
medical sources of exposure have more favorable
benefit/risk ratings than do the nonmedical
sources. Data are from a national survey in
Canada by Slovic et al., 1991.
41The Affect Heuristic
A model of the affect heuristic explaining the
risk/benefit confounding observed by Alhakami and
Slovic (1994). Judgments of risk and benefit are
assumed to be derived by reference to an overall
affective evaluation of the stimulus item.
42- Study 1 Risk and Benefit
- Judgments under Time Pressure
- Time pressure reduces opportunity
- for analytic deliberation, and
- increases reliance on affect
- Prediction Under time pressure
- people are more likely to use the
- affect heuristic to make
- judgments.
-
43Study 2 Manipulating Affect by Providing Risk
and Benefit Information
Technique provide information to change overall
impression, e.g., create a more favorable
affective evaluation of nuclear power with info
that it has high benefit. Perceived risk should
then decrease.
Nuclear Power
Affect
Information benefit is high
Inference risk is low
44Model showing how information about benefit (A)
or information about risk (B) could increase the
overall affective evaluation of nuclear power and
lead to inferences about risk and benefit that
coincide affectively with the information given.
Similarly, information could decrease the overall
affective evaluation of nuclear power as in C and
D. Source Finucane et al. (2000).
45Probability and Relative Frequency
- Are they the same or different in communicating
risk? - e.g., 1 chance
- vs.
- 1 out of 100
46RISK COMMUNICATION A patient Mr. James Jones
has been evaluated for discharge from an acute
civil mental health facility where he has been
treated for the past several weeks. A
psychologist whose professional opinion you
respect has done a state-of-the-art assessment of
Mr. Jones. Among the conclusions reached in the
psychologists assessment is the
following EITHER Patients similar to Mr.
Jones are estimated to have a 20 probability of
committing an act of violence to others during
the first several months after discharge.
OR Of every 100 patients similar to Mr.
Jones, 20 are estimated to commit an act of
violence to others during the first several
months after discharge.
47 Question
- If you were working as a supervisor at this
mental health facility and received the
psychologists report, would you recommend that
Mr. Jones be discharged from the hospital at the
present time?
48Patient Evaluation
- A. 10
- Very few people are violent
- 10 1/10
- Probably wont hurt anyone, though
- C. 10 out of 100
- Visual of 100 people and 10 who commit crimes
- Mr. Jones committing an act of violence
- People being harmed by the 10/100 patients
- B. 1 out of 10
- He could be the 1 out of 10
- Some guy going crazy and killing people
- The patient attacking someone
- An act of violence
- There has to be at least 1 in 10. Mr. Jones could
very well be that 1
49 Question
- If you were working as a supervisor at this
mental health facility and received the
psychologists report, would you recommend that
Mr. Jones be discharged from the hospital at the
present time?
50(No Transcript)
51Intuitive Toxicology Main Result
Many people lack dose-response sensitivity for
exposure to chemicals that can produce effects
that are dreaded, such as cancer (high
affect). If large exposures are bad, small
exposures are also bad.
52(No Transcript)
53Terrorism and Probability NeglectCass R.
SunsteinThe Journal of Risk and Uncertainty,
26(2/3) 121-136, 2003
- People are prone to . . . probability neglect,
especially when their emotions are intensely
engaged. Probability neglect is highly likely in
the aftermath of terrorism. People fall victim
to probability neglect if and to the extent that
the intensity of their reaction does not greatly
vary even with large differences in the
likelihood of harm. When probability neglect is
at work, peoples attention is focused on the bad
outcome itself, and they are inattentive to the
fact that it is unlikely to occur.
54From Sunstein, 2003
- In the context of the terrorist attacks of
September 11, 2001, . . ., public fear led to
private and public costs that were orders of
magnitude higher than the costs of the attacks
themselves, and that are best explained in part
by reference to probability neglect. The same
might be said about the extraordinary public fear
produced by the sniper attacks in the
Washington, D.C. area in October 2002 the extent
of the fear is hard to understand without an
appreciation of probability neglect.
55From Sunstein, 2003
- Fear, whether rational or not, is itself a cost,
and it is likely to lead to a range of other
costs, in the form of countless ripple effects, .
. . . If government is able to reduce the level
of fear produced by probability neglect, it
should do so . . .
56Insensitivity to the Value of Human Life
- I am deeply moved if I see one man suffering
and would risk my life for him. Then I talk
impersonally about the possible pulverization of
our big cities, with a hundred million dead. I
am unable to multiple one mens suffering by a
hundred million. - Albert Szent Gyorgi
57Figure 1. The value function from Kahneman and
Tverskys prospect theory. According to this
function, a fixed reduction in number of lives
lost (?) has more subjective value (?) when the
starting point is a small number (X1) than when
the starting point is a large number (Y1).
58Psychophysical Numbing Page 4B THE
REGISTER-GUARD, Eugene, Oregon, Sunday, January
30, 1994 By JEFF COHEN And NORMAN SOLOMON
COMMENTARY
Media ignore worlds dying kids (Excerpts)
excruciating detail. Later, President Clinton
mentioned her by name in his State of the Union
speech.
THE DEATH OF ONE child can be big news. But the
deaths of a million children are rarely news at
all.
seem far less tragic than the well-publicized
deaths of a few children. We need to wake
up, says Sam Harris, the executive director of
RESULTS, a grassroots anti-hunger organization.
Sometimes, when a child suffers, the news media
convey painful realities in human terms. We keep
seeing pictures, and hearing from loved ones. And
the entire nation seems to grieve. Thats
what happened when the hunt for 12-year-old Polly
Klaas ended with the awful discovery of her
lifeless body. The front pages and network news
told of her ordeal in
Every day, around the world, about 35,000
young children die not after being kidnapped,
but after being held in a different kind of
bondage poverty.
Children are dying due to lack of attention, our
attention, governments attention. They die not
because they have to, but because saving their
lives is not a priority.
Despite all the high-tech global
communications systems, we find out little about
those children. Scarcely reported, the
preventable deaths of several hundred thousand
kids each month
59(No Transcript)
60The proportion of lives saved carries more
affective meaning than the number of lived
saved For example, in separate evaluations there
will likely be more support for saving 80 of 100
lives at risk than saving 20 of 1000 lives at
risk.
61A Study of Proportion Dominance
Airport Safety Imagine you are a member of
the emergency response committee of the Eugene
Airport. There is a proposal before your
committee to purchase some expensive new
equipment for use in the event of a crash landing
of an airliner. The circumstances that might
require such equipment to be used are rare but
important. It is estimated that, over a 10
year period, there is about 1 chance in 1,000
that the equipment would be needed one time and
that it would save the 150 lives that would be in
jeopardy in such an event. Thus, the benefit
of this equipment in saving lives could be
portrayed by the gamble 1 chance in 1,000 to
save the 150 lives in jeopardy. Critics
argue that the money spent on this system could
be better spent enhancing other aspects of
airport safety. How strongly would you
support this proposed measure to purchase the new
equipment?
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Would not support at all Would not support at all Would not support at all Moderate support Moderate support Moderate support Very strong support Very strong support Very strong support
62Airport Safety Imagine you are a member of
the emergency response committee of the Eugene
Airport. There is a proposal before your
committee to purchase some expensive new
equipment for use in the event of a crash landing
of an airliner. The circumstances that might
require such equipment to be used are rare but
important. It is estimated that, over a 10
year period, there is about 1 chance in 1,000
that the equipment would be needed one time and
that it would save 98 of the 150 lives that
would be in jeopardy in such an event. Thus,
the benefit of this equipment in saving lives
could be portrayed by the gamble 1 chance in
1,000 to save 98 of the 150 lives in jeopardy.
Critics argue that the money spent on this
system could be better spent enhancing other
aspects of airport safety. How strongly would you
support this proposed measure to purchase the new
equipment?
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Would not support at all Would not support at all Would not support at all Moderate support Moderate support Moderate support Very strong support Very strong support Very strong support
63Proportion Dominance and Airport Safety
Saving a percentage of 150 lives receives higher
support ratings than does saving 150 lives.
Save 150 lives Save 98 of 150 lives Save 95 Save 90 Save 85
N 75 28 45 40 16
Mean support 10.4 13.6 12.9 11.7 10.9
64(No Transcript)
65Statistics are human beings with the tears
dried off How do we put the tears back on?
66(No Transcript)
67- If I look at the mass, I will never act.
- If I look at one, I will.
- Mother Teresa
68(No Transcript)
69(No Transcript)
70(No Transcript)
71(No Transcript)
72(No Transcript)
73Strong Affect Distorts Valuation
Value
Gain
Value functions based on calculations (dotted
line) and based on feeling (solid line). The
x-axis of the function is the scope of a
stimulus, and the y-axis is subjective value.
Source Hsee Rottenstreich.
74Valuation Insensitivity to number of lives at
risk Desvousges et al. (1992) asked separate
groups of participants how much they would donate
to save 2,000, 20,000, or 200,000 migrating birds
from drowning in uncovered oil ponds. The mean
responses were 80, 78, and 88 respectively.
Kahneman et al. (1999) argue that these results
could be explained by the notion that these
questions evoke a mental representation of a
prototypical incident, perhaps an image of an
exhausted oil-soaked bird and that respondents
decide how much to donate based on their
affective reactions to this image.
75- Do we lack a vocabulary to describe (and think
about) large losses of life? - havoc
- disaster
- holocaust
- catastrophe
- tragedy
- calamity
?
Hersey describes Hiroshima as havoc, tragedy,
disaster But Anne Franks story can move us.
76A Question to Think About
- Why do packages of food and tobacco products
carry all those little blurbs, such as new,
natural, or 98 fat free?
77(No Transcript)
78(No Transcript)
79(No Transcript)
801
2
3
5
4
6
7
8
9
10
11
12
13
81Risk Perception, Risk Knowledge, and Cigarette
Smoking
82(No Transcript)
83(No Transcript)
84Do Adolescent Smokers Know the Risks?
- Viscusi (1992) Smoking Making the Risky
Decision - Of 100 cigarette smokers, how many do you think
will get lung cancer because they smoke? - Mean 43/100 Actuarially
10/100 - Ages 16-21 gave higher estimates
- Conclusions
- Young people know (overestimate) the risks.
- No justification for information programs to
produce awareness. - Education about real risk will lead to more
smoking. - Smoking decisions are informed and rational.
85Critique of Viscusi
- Viscusi neglects
- Consequences
- Optimism bias
- Short-term perspective
- Misperception of addiction
- Experiential/affect based thinking
- Viscusis perspective is solely analytical,
treating the beginning smoker as a young
economist. In addition, his elicitation of the
young economists risk perceptions is invalid.
86Street Calculus
By Garry Trudeau
87Smoking The Dominance of Experiential
(Affective) Thinking Over Analytic Thinking 1.
Evidence Smokers give little or no
deliberate thought to risk, probabilities of
disease, severity of disease, the prospect of
addiction, or the long-term course of their
smoking behavior. 2. As a result
Smokers are not cognizant of the
frequencies/probabilities of disease in any
precise quantitative way (hence quantitative
estimates elicited by Viscusi are unreliable).
Short-term perspective immediate
affective factors dominate long-term thinking.
Smokers fail to anticipate how their perceptions
and values will change. Decision utility does
not equal experience utility. Young
smokers fail to appreciate the risk and meaning
of addiction (e.g., the visceral forces). 3.
Irrationality Although experiential
thinking is rational in many circumstances, it
produces decisions about smoking that the
individual later views as mistakes.
88(No Transcript)
89(No Transcript)
90(No Transcript)
91(No Transcript)
92(No Transcript)
93National Survey (Fall, 1999 Winter 2000 N
4000) Would You Start Smoking Again? (current
smoker) Yes No Adults 11.9
85.5 Teens 17.0 80.1 saying No
increases according to 1. Length of time
(years) smoked 2. Amount smoked/day
3. Self-rated level of addiction 4.
Number of quit attempts
94(No Transcript)
95(No Transcript)
96- The Power of Cigarette Advertising
- Sophisticated manipulation of affect and
preferences through words and images. - The mere exposure effect demonstrates that
repeated exposure to a stimulus (as through ads)
increases positive affect associated with that
stimulus. - According to the affect heuristic, increases in
positive affect through advertising will likely
depress perception of risk.
97(No Transcript)
98- Tobacco Companies Have Known This All Along
- Companies Internal Documents Show
- Initiation of smoking typically takes place
among children and adolescents whose decision
making capabilities are immature weighing
imagery, affect, emotion, and social
relationships more than logic, reason, or
analysis of risk - Young smokers are lured into the behavior by
prospects of fun, pleasure, and excitement - Time horizons are short (most starters
intended to smoke just a few, just for a lark
Bates 566627751) - Cigarette advertising and promotions have been
designed to exploit what companies have learned
from smoker psychology research (1000273741)
99There is more to the affect story
- Variants of Utility
- - decision utility
- - predicted utility
- - experienced utility
- Duration neglect
- Forecasting future affective states
(durability biases) - Temporal construal
100Temporal Construal(Y. Trope)Images near and
far
- Near images more emotionally concrete
- Psychologically complex
- Less optimistic
- Difficult to integrate and combine
- Far images more abstract
- Idealized
- Schematic/minimalist
- More optimistic
- More readily integrated and combined (similarity)
101Conclusions (1)
- The analytic and experiential (affective) systems
of thought are exquisitely sophisticated and
embody the essence of human rationality. - Both systems, however, can lead us astray.
- Each system needs the other for guidance.
102Conclusions (2)
- We cannot assume that an intelligent person can
understand the meaning of, and properly act upon,
even the simplest of numbers such as amounts of
money or numbers of lives at risk, not to mention
less familiar measures or statistics, unless
these numbers are infused with affect.
103(No Transcript)
104Conclusions (3)
- The scientific study of affective rationality is
in its infancy. It is exciting to contemplate
what might be accomplished by future research
designed to help humans understand the affect
heuristic and employ it beneficially in risk
analysis and other worthy endeavors.