Risk As Analysis and Risk As Feelings

1
Risk 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)
3
Neural 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.
4
At the psychological levelthis common
currencymay be affect

• A valenced quality (e.g., goodness
• or badness) associated
• with a stimulus

5
(No Transcript)
6
Information
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

 
7
There 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)
8
Affect Response Probability   Frequency   Rel
ative Frequency   Risk   Benefit   Valuation
(buying and selling prices)   Choice/Preference
9
Phineas Gage
10
Phineas Gage
11
(No Transcript)
12
The 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)

14
Images ?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

15
There 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)
17
Image Word Sound Smell Tactile sensation Memory Im
agination Visual Image
18
(No Transcript)
19
(No Transcript)
20
HATE
21
prime 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)
22
Heart 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
23
Valuation 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

25
Evaluability 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 . . .
26
Attributes 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)
27
Affect, 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.

28
Results 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)
30
Numeracy (from the cancer literature)

• The ability to understand and use basic
  probability and mathematical concepts (Lipkus et
  al., 2001)
• New studies with Ellen Peters

31
Numeracy 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 --
32
The 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.

33
High (and not low) numerate drive the surprising
effect and find the objectively worse bet more
attractive
34
High 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
35
Risk As Analysis vs. Risk as Feelings
Analytic/ Deliberative
Experiential/ Affective
36
Street Calculus
By Garry Trudeau
37
In the world, risk and benefit are positively
correlated. In peoples minds, they are
negatively correlated.
38
Relationship between risk and benefit in peoples
minds
39
The 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.
40
Radiation
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.
41
The 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.

43
Study 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
44
Model 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).
45
Probability and Relative Frequency

• Are they the same or different in communicating
  risk?
• e.g., 1 chance
• vs.
• 1 out of 100

46
RISK 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?

48
Patient 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)
51
Intuitive 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)
53
Terrorism 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.

54
From 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.

55
From 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 . . .

56
Insensitivity 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

57
Figure 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).
58
Psychophysical 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)
60
The 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.
61
A 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
62
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 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
63
Proportion 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)
65
Statistics 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)
73
Strong 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.
74
Valuation 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.
76
A 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)
80
1
2
3
5
4
6
7
8
9
10
11
12
13
81
Risk Perception, Risk Knowledge, and Cigarette
Smoking
82
(No Transcript)
83
(No Transcript)
84
Do 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.

85
Critique 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.

86
Street Calculus
By Garry Trudeau
87
Smoking 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)
93
National 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)

99
There 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

100
Temporal 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)

101
Conclusions (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.

102
Conclusions (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)
104
Conclusions (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.