The Risk Game
Paul Slovic
Decision Research
1201 Oak Street
Eugene, Oregon 97401 USA
phone: 541-485-2400
fax: 541-485-2403
email: pslovic@oregon.uoregon.edu
June 5, 1997
The Risk Game
page 3
Abstract
In the context of health, safety, and environmental decisions, the concept of risk
involves value judgments thatreflect much more than just the probability and
consequences of the occurrence ofanevent. This article conceptualizes the act ofdefining
andassessing riskas a game, in which the rules must be socially negotiated within the
context of a specific problem. This contextualist view of risk provides insight into why
technical approaches to risk management often fail with problems such as those involving
radiation and chemicals, where scientific experts and the publicdisagree on the natureof
therisks. It also highlights theneed for allowing the interested and affected parties to
define and play the game, thus emphasizing the importance of institutional, procedural,
and societal processes in risk management decisions.
The Risk Game
page 4
1.0 Introduction
The practice of risk assessment has steadily increased in prominence during thepast
several decades, as risk managers in government and industry have soughtto develop
more effective ways to meet public demands for a safer andhealthier environment.
Dozens of scientific disciplines have been mobilized to provide technical information
aboutrisk, andbillions of dollars have beenexpended to create this information and
distill it in the context of risk assessments.
Ironically, as oursociety and other industrialized nations have expended this great
effort to make life saferand healthier, many in the publichavebecome more, rather than
less, concerned about risk. These individuals see themselves asexposed to more serious
risks than were faced bypeople in thepast, and they believe that this situation is getting
worse rather than better. Nuclear and chemical technologies (except for medicines) have
been stigmatized by being perceived as entailing unnaturally great risks.1 As a result, it
has been difficult, if notimpossible, to find host sites for disposing ofhigh-level or low-
level radioactive wastes, or for incinerators, landfills, and other chemical facilities.
Public perceptions ofrisk have been found to play an important role indetermining
the priorities and legislative agendas ofregulatory bodies such as the Environmental
Protection Agency, much to the distress ofagency technical experts who argue that other
hazards deserve higher priority. The bulk ofEPA's budget inrecent years has gone to
hazardous waste primarily because the public believes that the cleanup ofSuperfund sites
is themost serious environmental threat that the country faces. Hazards such as indoor air
The Risk Game
pageS
pollution are considered more serious health risks byexperts butare notperceived that
wayby the public.2
Great disparities in monetary expenditures designed to prolong lifemay alsobe
traced to public perceptions ofrisk. As noteworthy asthe large sums ofmoney devoted to
preventing a statistical fatality from exposure toradiation and chemical toxins are the
relatively small sums expended to prevent a fatality from mundane hazards such as
automobile accidents. Other studies have shown that serious risks from national disasters
suchas floods, hurricanes, andearthquakes generate relatively littlepublic concern and
demand for protection.3,4
Such discrepancies are seen as irrational bymany harsh critics of public perceptions.
These critics drawa sharp dichotomy between the experts and the public. Experts are
seen as purveying risk assessments, characterized as objective, analytic, wise, and
rational — based upon the real risks. In contrast, thepublic is seen to rely upon
perceptions ofrisk that are subjective, often hypothetical, emotional, foolish, and
irrational (see, e.g., refs. 5 or6). Weiner7 defends the dichotomy, arguing that "This
separation ofreality and perception is pervasive in a technically sophisticated society, and
serves to achieve a necessary emotional distance ..." (p. 495).
Insum, polarized views, controversy, and overt conflict have become pervasive
within riskassessment andriskmanagement. Frustrated scientists and industrialists
castigate the public for behaviors they judge to be based on irrationality or ignorance.
Members ofthe public feel similarly antagonistic toward industry and government. A
desperate search for salvation through risk-communication efforts began in the mid-1980s
The Risk Game
page 6
— yet, despite some localized successes, this effort has not stemmed the major conflicts
or reduced much of the dissatisfaction with risk management. This dissatisfaction can be
traced, in part, to a failure to appreciate thecomplex and socially determined nature of the
concept "risk." In the remainder of this paper, I shall illustrate thiscomplexity andpoint
toward the need for new definitions of risk and new approaches to risk management.
2.0 The Need for a New Perspective
Newperspectives and new approaches are needed to manage risks effectively in our
society. Social science research has provided some valuable insights into thenature of the
problem that, without indicating a clear solution, do point to some promising prescriptive
actions.
Forexample, early studies of risk perception demonstrated thatthepublic's concerns
could notsimply beblamed on ignorance or irrationality. Instead, research has shown that
many ofthe public's reactions to risk can be attributed to a sensitivity to technical, social,
and psychological qualities ofhazards that are not well-modeled in technical risk
assessments (e.g., qualities such as uncertainty in riskassessments, perceived inequity in
the distribution of risks and benefits, and aversion to being exposed to risks that are
involuntary, not under one's control, ordreaded). The important role ofsocial values in
risk perception and risk acceptance has thus become apparent.8
More recently, another important aspect of therisk-perception problem has come to
be recognized. This is the role oftrust. Inrecent years there have been numerous articles
and surveys pointing out the importance oftrust in risk management and documenting the
The Risk Game
page 7
extreme distrust we now have in many of the individuals, industries, and institutions
responsible for risk management.9 Thispervasive distrust has alsobeenshown to be
strongly linked to the perception that risks are unacceptably high and to political activism
to reduce those risks.
A third insight pertains to the very nature of the concept "risk." Current approaches
to risk assessment and risk management are based upon the traditional view of risk as
some objective function ofprobability (uncertainty) and adverse consequences. I shall
argue for a conception of risk that is starkly different from this traditional view. This new
approach highlights the subjectiveand value-ladennature of risk and conceptualizes the
act of defining and assessing risk as a game in which the rules must be socially negotiated
within the context of a specific problem.
3.0 The Subjective and Value-Laden Nature of Risk Assessment
Attempts to managerisk must confrontthe question: "What is risk?" The dominant
conception views risk as "thechance of injury, damage, or loss."10 The probabilities and
consequences of adverse events are assumed to be produced by physical and natural
processes in ways that canbe objectively quantified by riskassessment. Muchsocial
science analysis rejects this notion, arguing instead that risk is inherently subjective.11"16
In this view, risk does not exist "out there," independent of our minds and cultures,
waiting to be measured. Instead, human beings have invented theconcept riskto help
them understand and cope with the dangers and uncertainties of life. Althoughthese
dangers are real, there is no such thing as "real risk" or "objective risk." The nuclear
The Risk Game
page 8
engineer's probabilistic risk estimate for a nuclear accident or the toxicologist's
quantitative estimate of a chemical's carcinogenic risk are both based on theoretical
models, whose structure is subjective and assumption-laden, and whose inputs are
dependent on judgment. As we shall see, nonscientistshave their own models,
assumptions, and subjective assessment techniques (intuitive risk assessments), which are
sometimes very different from the scientists' models.
One way in which subjectivity permeates risk assessments is in the dependence of
such assessments on judgments at every stage of the process, from the initial structuring
of a risk problemto deciding whichendpoints or consequences to include in the analysis,
identifying and estimating exposures, choosing dose-response relationships, and so on.
For example, even the apparently simpletask of choosing a risk measure for a well-
defined endpoint such as human fatalities is surprisingly complex andjudgmental. Table
1 shows a few of the many different ways that fatality risks can be measured. How should
we decide which measure to use when planning a risk assessment, recognizing that the
choice is likelyto makea big difference in how the risk is perceived and evaluated?
Insert Table 1 about here
Anexample taken from Wilson and Crouch17 demonstrates how thechoice of one
measure or another can make a technology look either more or less risky. For example,
between 1950 and 1970, coal mines became much less risky in terms of deaths from
accidents per ton ofcoal, but they became marginally riskier in terms ofdeaths from
accidents peremployee. Which measure one thinks more appropriate for decision making
The Risk Game
page 9
depends on one's point of view. From a national point ofview, given that a certain
amount of coal has to be obtained, deaths per million tons of coal is the more appropriate
measure of risk, whereas from a labor leader's point of view, deaths per thousand persons
employed may be more relevant.
Each way of summarizing deaths embodies itsown setof values.18 Forexample,
"reduction in life expectancy" treats deaths ofyoung people as more important than
deaths of older people, who have less life expectancy to lose. Simply counting fatalities
treats deaths of the old and young as equivalent; it also treats as equivalent deaths that
come immediately after mishaps and deaths that follow painful and debilitating disease or
long periods during whichmanywho will not sufferdisease live in daily fear of that
outcome. Using "number of deaths" as the summary indicator of risk implies that it is as
important to prevent deaths of people who engage in an activity by choice and deaths of
those who have been benefiting from a risky activity or technology as to protect those
who get no benefitfrom it. One can easily imagine a range of arguments to justify
different kinds of unequal weightings for different kinds of deaths, but to arrive at any
selection requires a valuejudgmentconcerning whichdeaths one considers most
undesirable. To treat the deaths as equal also involves a value judgment.
3.1 Framing the Risk Information
After a risk analysis has "negotiated," all the subjective stepsof defining the
problem and its options, selecting and measuring risks in terms of particular outcomes,
determining thepeople at risk and their exposure parameters, and so on, one comes to the
The Risk Game
page 10
presentation of this information to the decision maker, often referred to as "framing."
This process of presentation is also rife with subjectivity.
Numerous research studies have demonstrated that different (but logically
equivalent) ways of presenting the same risk information can lead to differentevaluations
and decisions. One dramatic example of this comes from a study by McNeil, Pauker, Sox,
and Tversky,19 who asked people to imagine that they had lung cancer and had to choose
betweentwo therapies, surgery or radiation. The two therapies were describedin some
detail. Then one group of subjects was presented with the cumulative probabilities of
surviving for varying lengths of time after the treatment. A second groupof subjects
received the same cumulative probabilities framed in terms of dying rather than surviving
(e.g., instead of being told that 68% of those having surgery will have survived afterone
year, theywere told that 32%will have died). Framing the statistics in terms of dying
changed the percentage of subjects choosing radiation therapy oversurgery from 18% to
44%. The effect was as strong for physicians as for laypersons.
Equally striking changes in preference result from framing the information about
consequences in terms ofeither lives saved or lives lost20 or from describing an
improvement in a river's water quality as a restoration of lostquality or an improvement
from the current level.21
We now knowthat every form of presenting risk information is a frame that has a
strong influence onthe decision maker. Moreover, when we contemplate the equivalency
of lives saved vs. lives lost, mortality rates vs. survival rates, restoring lost water quality
The Risk Game
page 11
vs. improving water quality, and so forth, we see that there are often no "right frames" or
"wrong frames" —just "different frames."
3.2 The Multidimensionalitv of Risk
As noted above, research has also shown that the public has a broad conception of
risk, qualitative and complex, that incorporates considerationssuch as uncertainty, dread,
catastrophic potential, controllability, equity, risk to future generations, and so forth, into
the risk equation. In contrast, experts' perceptionsof risk are not closely related to these
dimensions or the characteristics that underlie them. Instead, studies show that experts
tend to see riskiness as synonymous with expected mortality, consistent with the
dictionary definitiongiven above and consistent with the ways that risks tend to be
characterized in risk assessments (see, for example, ref. 22). As a result of these different
perspectives, many conflicts over "risk" may result from experts and laypeople having
different definitions of the concept. In this light, it is not surprising that expert recitations
of "risk statistics" often do little to change people's attitudes and perceptions.
Thereare legitimate, value-laden issues underlying the multiple dimensions of
public risk perceptions, and these values needto be considered in risk-policy decisions.
For example, is risk from cancer (a dread disease) worse than risk from auto accidents
(notdreaded)? Is a risk imposed on a child more serious than a known riskaccepted
voluntarily by an adult? Are the deaths of 50passengers in separate automobile accidents
equivalent to the deaths of 50 passengers inone airplane crash? Is the risk from a polluted
Superfund site worse if the site is located ina neighborhood that has a number of other
The Risk Game
page 12
hazardous facilities nearby? The difficult questions multiply when outcomes other than
human health and safety are considered.
3.3 The Risk Game
There are clearly multiple conceptions of risk.23 Thompson and Dean24 note thatthe
traditional view of risk characterized by the event probabilities and consequences treats
subjective and contextual factors suchas those described above as secondary or
accidental dimensions of risk, just as coloration might be thought of as a secondary or
accidental dimension of an eye. Accidental dimensions mightbe extremely influential in
the formation of attitudes toward risk, just as having blue or brown coloration may be
influential in forming attitudes toward eyes. Furthermore, it may be that all risks possess
some accidental dimensions,just as all organs of sight are in some way colored.
Nevertheless, accidental dimensions do not serve as criteria for determining whether
someone is or is not at risk,just as coloration is irrelevant to whether something is or is
not an eye.
I believe that the multidimensional, subjective, value-laden, frame-sensitive nature
of risky decisions, as described above, supports a very different view, which Thompson
and Dean call "the contextualist conception." This conception places probabilities and
consequences onthe list of relevant risk attributes along with voluntariness, equity, and
other important contextual parameters. On the contextualist view, the concept ofrisk is
more like theconcept of a game than theconcept of theeye. Games have time limits,
rules ofplay, opponents, criteria for winning or losing, and soon, but none of these
attributes is essential to the concept of a game, nor is anyof themcharacteristic of all
The Risk Game
page 13
games. Similarly, a contextualist view ofrisk assumes that risks are characterized by
some combinationof attributes such as voluntariness, probability, intentionality, equity,
and soon, butthat no one of these attributes is essential. The bottom line is that, justas
there is no universal set of rules for games, there is no universal set of characteristics for
describing risk. The characterization must depend on which risk game is being played.
4.0 Resolving Risk Conflicts
4.1 Technical Solutions to Risk Conflicts
There hasbeen no shortage of high-level attention given to the riskconflicts
described in the introduction to thispaper. One prominent proposal by Justice Stephen
Breyer25 attempts to break what he sees as avicious circle ofpublic perception,
congressional overreaction, and conservative regulation that leads to obsessive and costly
preoccupation with reducing negligible risks as well as to inconsistent standards among
health and safety programs. Breyer sees public misperceptions ofrisk and low levels of
mathematical understanding atthe core ofexcessive regulatory response. His proposed
solution is to create a small centralizedadministrative group charged with creating
uniformity and rationality inhighly technical areas ofrisk management. This group
would be staffed by civil servants with experience in health and environmental agencies,
Congress, and OMB. Aparallel is drawn between this group and the prestigious Conseil
d'Etat in France.
Similar frustration withthe costs of meeting public demands led the 104th Congress
to introduce numerous bills designed to require all major new regulations to be justified
The Risk Game
page 14
by extensive risk assessments. Proponents of this legislation arguedthat suchmeasures
were necessary to ensure that regulations are based upon"sound science" and effectively
reduce significant risks at reasonable costs.
The language of this proposed legislation reflects thetraditional narrow view of risk
and risk assessment based "... only on the best reasonably available scientific data and
scientific understanding ..." Agencies are further directed to develop a systematic
program for external peer review using "expert bodies" or "other devices comprised of
participants selected onthe basis of their expertise relevant to the sciences involved ...
Public participation inthis process isadvocated, but no mechanisms for this are specified.
The proposals byBreyer and the 104th Congress are typical in their call for more
andbettertechnical analysis andexpert oversight to rationalize risk management. There
is no doubt that technical analysis is vital for making risk decisions better informed, more
consistent, and more accountable. However, value conflicts and pervasive distrust in risk
management cannot easily bereduced by technical analysis. Trying to address risk
controversies primarily with more science is, in fact, likely to exacerbate conflict.
4.2 Process-Oriented Solutions
A major objective ofthis paper has been todemonstrate the complexity ofrisk and
its assessment. To summarize the earlier discussions, danger is real, but risk is socially
constructed. Riskassessment is inherently subjective andrepresents a blending of science
and judgment with important psychological, social, cultural, and political factors. This
complexity leads to a "contextualist view" in which risk is conceptualized as a game
whose rules must be socially negotiated within the context of specific decision problems.
26
The Risk Game
page 15
Whoever controls the definition of risk (i.e., determines the rules of the risk game)
controls the rational solution to the problem at hand. If you define risk one way, then one
optionwill rise to the top as the most cost-effective or the safestor the best. If you define
it another way, perhaps incorporating qualitative characteristics and other contextual
factors, you will likely get a different ordering ofyour action solutions.27 Defining risk is
thus an exercise in power.
The limitations of risk science, the importance and difficulty of maintaining trust,
and the subjective and contextual natureof the risk game point to the need for a new
approach—one that focuses upon introducing more public participation intobothrisk
assessment and risk decision making in order to make the decision process more
democratic, improve the relevance and qualityof technical analysis, and increase the
legitimacy andpublic acceptance of the resulting decisions. Work by scholars and
practitioners in Europe andNorth America hasbegun to lay the foundations for
improved methods of public participation within deliberative decision processes that
include negotiation, mediation, oversight committees, andother forms of public
involvement.18,28"31
Recognizing interested and affected citizens as legitimate partners in the exercise of
risk assessment is no short-term panacea for the problems of risk management. But
serious attention to participation andprocess issues may, in the longrun, leadto more
satisfying and successful waysto manage risk.
The Risk Game
page 16
Acknowledgment
Preparation of this paper was supported by the Alfred P. Sloan Foundation, the
Electric Power Research Institute, and the National Science Foundation under Grant Nos.
91-10592 and SBR 94-122754.
This paper contains material adapted from two sources:
1) Kunreuther, H. & P. Slovic (1996). Science,values, and risk. The Annals ofthe
American Academy ofPolitical and Social Science, 545, pp. 116-125.
2) Slovic, P. (1997). Trust, emotion, sex, politics, and science: Surveying the risk-
assessment battlefield. In M.Bazerman et al. (Eds.), Environment, Ethics, and
Behavior (pp. 277-313). San Francisco: New Lexington.
The Risk Game
page 17
5.0 References
1. Gregory, R., Flynn, J., & Slovic, P. Technological stigma. American Scientist, 83,
1995, pp. 220-223.
2. United States Environmental Protection Agency (U.S. EPA). Unfinished business: A
comparative assessment ofenvironmentalproblems. Washington, DC: Author.
1987.
3. Palm, R. I. Natural hazards:An integrativeframeworkfor research andplanning.
Baltimore: Johns Hopkins. 1995.
4. Kunreuther, H. Mitigating disaster losses through insurance. Journal ofRiskand
Uncertainty. In press.
5. DuPont, R. L. Nuclearphobia — Phobic thinking about nuclearpower.
Washington, DC: The Media Institute. 1980.
6. Covello, V. T., Flamm, W. G., Rodricks, J. V., & Tardiff, R. G. The analysis of
actual versusperceived risks. New York: Plenum. 1983.
7. Weiner, R. F. Comment on Sheila Jasanoffs guest editorial. Risk Analysis, 13, 1993,
pp. 495- 496.
8. Slovic, P. Perception of risk. Science, 236, 1987, pp. 280-285.
9. Slovic, P. Perceived risk, trust, and democracy: A systems perspective. Risk
Analysis, 13, 1993, pp. 675-682.
10. Webster, N. Webster's newtwentieth century dictionary (2nd). New York: Simon &
Schuster. 1983.
The Risk Game
page 18
11. Funtowicz, S. O., & Ravetz, J. R. Three types of risk assessment and the emergence
of post-normal science. In S. Krimsky & D. Golding (Eds.), Social theories ofrisk
(pp. 251-274). Westport, CT: Praeger. 1992.
12. Krimsky, S., & Golding, D. Social theoriesofrisk. Westport, CT: Praeger-
Greenwood. 1992.
13. Otway, H. Public wisdom, expert fallibility: Toward a contextual theory of risk. In
S. Krimsky & D. Golding (Eds.), Social theories ofrisk (pp. 215-228). Westport,
CN: Praeger. 1992.
14. Pidgeon, N., Hood, C, Jones, D., Turner, B., & Gibson, R. Risk perception. In
Royal Society Study Group (Ed.), Risk: Analysis, perception andmanagement
(pp. 89-134). London: The Royal Society. 1992.
15. Slovic, P. Perception of risk: Reflections on the psychometric paradigm. In
S. Krimsky & D. Golding (Eds.), Social theories ofrisk(pp. 117-152). New York:
Praeger. 1992.
16. Wynne, B. Risk and social learning: Reification to engagement. In S. Krimsky&
D. Golding (Eds.), Social theories ofrisk(pp. 275-300). Westport, CT: Praeger.
1992.
17. Wilson, R., & Crouch, E. Risk/Benefit analysis. Cambridge, MA: Ballinger. 1982.
18. National Research Council. Understanding risk: Informing decisions in a
democratic society. Washington, DC: National Academy Press. 1996.
The Risk Game
page 19
19. McNeil, B. J., Pauker, S. G., Sox, H. C, Jr., & Tversky, A. On the elicitation of
preferences for alternative therapies. New EnglandJournal ofMedicine, 306, 1982,
pp. 1259-1262.
20. Tversky, A., & Kahneman, D. The framing of decisions and the psychology of
choice. Science, 211, 1981, pp. 453-458.
21. Gregory, R., Lichtenstein, S., & MacGregor, D. G. The role of past states in
determining reference points for policy decisions. Organizational Behavior and
Human Decision Processes, 55, 1993, pp. 195-206.
22. Cohen, B. L. Criteria for technology acceptability. RiskAnalysis, 5, 1985, pp. 1-2.
23. Shrader-Frechette, K. S. Riskand rationality. Berkeley: University of California.
1991.
24. Thompson, P. B. & Dean, W. R. Competingconceptions of risk. Risk: Health,
Safety & Environment, 7, Fall 1996, pp. 361-384.
25. Breyer, S. Breaking the vicious circle: Toward effective riskregulation. Cambridge,
MA: Harvard University. 1993.
26. U.S. Senate. The comprehensive regulatory reform act of1995 [Dole/Johnson
discussion draft to S. 343], 1995, pp. 57-58.
27. Fischhoff, B., Watson, S., & Hope, C. Defining risk. Policy Sciences, 17, 1984,
pp. 123-139.
28. English, M. R. Siting low-level radioactive waste disposalfacilities. NewYork:
Quorum. 1992.
The Risk Game
page 20
29. Kunreuther, H., Fitzgerald, K., & Aarts, T. D. Siting noxious facilities: A test of the
facility siting credo. RiskAnalysis, 13, 1993, pp. 301-318.
30. Renn, O., Webler, T., & Johnson, B. Citizen participation for hazard management.
Risk — Issues in Health and Safety, 3, 1991, pp. 12-22.
31. Renn, O., Webler, T., & Wiedemann, P. Fairness and competence in citizen
participation. Dordrecht, The Netherlands: Kluwer. 1995.
Table 1. Some Ways of Expressing Mortality Risks
Deaths per million people in the population
Deaths per million people within x miles of the source of exposure
Deaths per unit of concentration
Deaths per facility
Deaths per ton of air toxic released
Deaths per ton of air toxic absorbed by people
Deaths per ton of chemical produced
Deaths per million dollars of product produced
Loss of life expectancy associated with exposure to the hazard
The Risk Game
page 21