1
Q
what do we mean by risk?
A
- not one agreed definition because there is not one definition that is suitable for all problems
- concerns the potential of gaining or losing something of value resulting from a given action or inaction which can be foreseen or unforeseen
- values = physical/emotional health, social status, financial wealth
2
Q
what is risk perception?
A
- the subjective judgement that people make about the severity and probability of a risk
- varies from person to person
- the study of risk perception arose out of the observation that experts and lay people often disagree about how risky various technologies and hazards are
3
Q
actual risk
A
- risk perception differs from actual risk
- actual risk is the actual dangers of certain things
- people can perceive things to be more or less risky than what they actually are
4
Q
Slovic, Fischhoff and Liechtenstein (1981)
A
- 4 groups of people asked to rate events/activities based on how risky they were
–> women voters, college, students, active club members, experts - technical expert judgements are in line with objective data
- lay judgements do not agree with objective data
- lay judgements of risk do not even agree with lay estimates of current yearly deaths
–> clear disconnect between risk perception and chance of death
–> e.g women voters and students assigned nuclear power the highest risk value but the lowest in terms of annual fatalities
5
Q
lack of objection / news coverage in risk perception
A
- the way in which events / risks are covered varies, e.g:
1. Boston marathon
–> 15 April 2013
–> 3 dead
– Acres of news coverage
2. West fertilizer company explosion
–> 17 April 2013
–> 15 dead
–> Virtually no coverage
3. Beirut Port explosion (fertilizer company)
–> 4th August 2020
–> 204 dead
–> 7500+ injuries
–> international coverage - coverage depends on risk perception
6
Q
risk perceptions affect political attitudes
A
- Huddy, Feldman, Taber & Lahav (2005):
–> level of perceived risk predicts support for aggressive anti-terrorism strategies - Barrios & Hochberg (2020):
–> the lower perceived risk of COVID-19 in a state, the higher percentage of Trump voters
–> Trump didn’t do a lot for COVID
7
Q
risk perceptions affect behaviour
A
- Gigerenzer (2006):
–> “An estimated 1,500 Americans died on the road in the attempt to avoid the fate of the passengers who were killed in the four fatal flights.”
–> after 9/11 people flew less and drove way more - Dryhurst et al. (2020):
–> risk perception correlated significantly with preventative health behaviors in ten countries across Europe, America, and Asia
8
Q
risk perceptions affect economy and society
A
- decrease in footfall and tourism following Salisbury novichok poisonings (led to £3.7m fund from Government to help city recover)
- UK enters recession after gross domestic product plunged 20.4% in second quarter of 2020
9
Q
risk perceptions allow us to communicate better
A
- Acton, Rogers & Zimmerman (2007):
–> people are much more concerned about radiation from dirty bomb, than about the explosion
–> but guess which will kill more people?
–> the explosion
–> talking about risk perception increases communication - Neumann-Bohme et al. (2020):
–> concerns about vaccine side-effects/safety reducing willingness to be vaccinated against COVID-19
10
Q
risk perceptions allows us to direct policy
A
- Taylor, Dessai & Bruine de Bruin (2014):
–> public perception of climate risk predicts support for climate change policies - Foad et al. (2021):
–> the more people perceive the threat of COVID-19 as severe, the more they supported lockdown
11
Q
risk perception and behaviour
A
- the effect of risk perception on behaviour is particularly important as this can have ripple effects on other areas
–> e.g. economy, society, government policy - risk perceptions feature in a number of theoretical models describing the psychological determinants of behaviour
12
Q
Protection Motivation Theory (Rogers & Prentice-Dunn, 1997)
A
- developed to explain the behavioural impact of health risk information
- protective behaviours occur when individuals have high protection motivation
- protection motivation = a motive that arouses, sustains and directs activity
- usually measured using intentions for the relevant protective behaviour
- protection motivation arises as the result of combining two cognitive appraisals:
1. Threat appraisal
2. Coping appraisal
13
Q
threat appraisals
A
- considers the risky behaviour
–> e.g. smoking - the perceived severity of the threat:
–> “If I continue to smoke, I could get lung cancer” - probability of being vulnerable to the threat:
–> “I won’t get cancer from smoking because my grandfather smoked and lived healthily” - rewards of the risky behaviour:
–> “Smoking helps me concentrate, calms my nerves”
14
Q
threat appraisal equation
A
perceived severity of threat + probability of bein vulnerable to the threat - the rewards of the threat
15
Q
coping appraisals
A
- considers the adaptive / protective behaviours
- response efficacy
–> will behaviour deal with the threat
–> “If I quit smoking, I can reduce my risk of getting lung cancer compared to people who continue to smoke” - self-efficacy or confidence for being able to engage in the behaviour:
–> “I tried to quit five times before and it never works” - response costs of the adaptive behaviour:
–> “If I stop smoking, the calming effects of smoking will be gone”
16
Q
coping appraisal equation
A
reponse efficacy + self-efficacy - response costs
17
Q
protection motivation equation
A
threat appraisal + coping appraisal = protection motivation
18
Q
interaction between appraisals
A
- threat and coping appraisals interact to influence motivation for the protective behaviour
–> e.g If people believe they can cope with the threat, then the greater the perceived threat, the greater their intentions for the protective behaviour - however, if people don’t believe they can cope, increases in perceived threat, decreases intentions for the protective behaviour
19
Q
evidence for the protection motivation theory (Floyd, Prentice-Dunn & Rogers, 2000)
A
- review of 65 studies that represented over 20 health behaviours
- increases in threat severity, threat vulnerability of the risky behaviour
AND - increases in response efficacy and self-efficacy of the protective behaviour
–> increased adaptive intentions or behaviors - increases in risky behaviour rewards and protective behaviour response costs
–> decreased adaptive intentions or behaviors - in general, coping variables showed slightly stronger relations with the adaptive behaviors than did the threat variables
20
Q
evidence for protection motivation theory (Malmir et al., 2018)
A
- RCT to test effects of PMT intervention on preventing cervical cancer
- completed a questionnaire including PMT constructs and demographic variables
- experimental group received educational intervention based on PMT
–> 6 sessions of lectures, group discussions, and questions–answers with pamphlets/booklets
–> covering topics such as epidemiology, signs and symptoms, risk factors, importance of early detection via cervical screening
21
Q
results of Malmir et al (2018)
A
- intervention had significant effects on the experimental groups’ PMT construct scores
- cervical screening attendance and health center enquiries about screening were significantly increased after 3 months in the experimental group, but not the control group
22
Q
critical evaluation of studies testing PMT
A
- many studies measure intentions rather than behaviour (intention-behaviour gap)
- measurement of intentions is often vague with no time frame (e.g. in the future…)
–> people with more perceived vulnerability may intend to change their behaviour earlier than others - when behaviour is assessed, self-report measures are often used (social desirability and recall bias)
- measurements of PMT constructs are often not validated/reliable
23
Q
critical evaluation of studies testing PMT (continued)
A
- smaller relationships between threat appraisals and protective behaviour:
–> could be because threat appraisals are better predictors of behaviours intended to reduce specific health threats (e.g., getting vaccinated)
–> but less for behaviours that have a variety of health, social and economic consequences (e.g. exercise, diet) - PMT predictions have been tested and supported using experimental manipulations
–> experimental design = strength
–> but criticised for tending not to use mediation analyses to test if the impact of the intervention is due to changes in the PMT constructs
–> follow-up tends to be short
24
Q
Brewer et al (2007)
A
correlation between perceived vulnerability and vaccination intentions is stronger than behaviours with wider consequences
25
theories / approaches to understanding risk perception
1. Cultural theory
2. Heuristics and biases
3. The psychometric paradigm of risk
4. Social amplification of risk
26
cultural theory
- risk perceptions are socially constructed by institutions, cultural values and ways of life (less about personal control)
- individuals are expected to form perceptions of risk that reflect and reinforce their commitment to one or another “cultural way of life”
--> individuals gravitate toward perceptions of risk that advance the way of life to which they are committed
- 4 ‘ways of life’ across the dimensions: Group and Grid
--> Fatalism
--> Hierarchy
--> Individualism
--> Egalitarianism
27
cultural theory - ways of life
- Fatalists feel regulated by social groups they do not belong to, indifferent about risk
--> weak group, strong grid
- Hierarchy fear things that disrupt the natural order of society
--> strong group, strong grid
- Individualism fear things that obstruct their individual freedom
--> weak group, weak grid
- Egalitarianism fear things that increase inequalities amongst people
--> strong group, weak grid
28
cultural theory - group & grid
- grid:
--> strong = life should be organised through role differentiation based on sex, ethnicity, wealth etc...
--> weak = life should be equal where anyone can participate in any social role
- group:
--> strong = view life as a sense of community, people depend on each other
--> weak = life as competitive, people fend for themselves
29
empirical evidence for cultural theory (Dake, 1991)
- cross-sectional telephone survey
- measured associations between 3 ways of life (Hierarchy, Individualism, Egalitarianism) and societal risk concerns
--> hierarchy and individualism are related to concerns about social deviance and war
--> but individualism more concerned about market issues
--> egalitarianism is related to concerns about technology, the environment and breakdown of democracy
- suggest that each cultural way of life is associated with different risk perceptions
30
limitations of cultural theory
- doesn’t allow for individual rational choice
- empirical support based on studies using unreliable and non-validated measures of ‘the ways of life’
- respondents could - and do - rate high on more than one world view
- only explain a small amount of variance in individual perceptions of risk
--> e.g. studies typically find about 5% of variance in risk judgments are explained by cultural theory measures
31
heuristics and biases
- found that people use a number of heuristics to evaluate information.
- heuristics = mental models on which to base judgements of risk
--> often useful shortcuts for thinking
--> but may lead to inaccurate judgments in some situations (cognitive biases)
--> based on experience
- especially likely to be used where time is short or self-efficacy for studying the facts is low
- provide a way of arriving at a decision about risks where the information is incomplete, unfathomable or bewildering in terms of its possibilities
32
types of heuristic
1. Affect heuristic
2. Availability heuristic
3. Optimism Bias
33
Affect heuristic
- draws attention to the way that feelings about a risk object are used by individuals to arrive at conclusions about levels of risk and their acceptability
- if feelings towards a risk object are positive, there is a tendency to underestimate the potential harm and overestimate the benefits
- altering people’s emotions, often affects how they perceive a risk
--> positive perception = lower risk perception
--> negative perception = higher risk perception
34
empirical support for the affect heuristic
- Hsee & Kunreuther (2000)
--> what woud you pay to insure this clock...
--> a) gift from grandparents
--> b) no real sentimental value
--> people are willing yo pay twice as much for object with sentimental value
- Yamagishi (1997)
--> what's worse a disease that kills...
--> a) 1,286 out of every 10,000 people
--> b) 24.12% of people
--> people most often choose option a, even thought option b has a higher risk of death (12% vs 24%)
- manipulating emotion and affect change your risk perception
34
availability heurisitic
- risk perceptions depend on how easily someone can see the risk as happening
- risk perceptions based on how easily one can think of an example of a risky situation happening
--> direct and indirect suffering explains most of the variance in risk judgements
- usually works, but the advent of the mass media causes problems
--> rare causes of death tend to be overestimated because they hit the media
--> common causes of death tend to be underestimated because they don’t hit the media
35
empirical support for availability heuristic - Keller et al (2006)
- 1598 participants were asked to imagine:
--> that they were planning to buy a house, and that they received some information about the probability of a flood
- availability heuristic based on direct experience affected risk perceptions:
--> participants with own experience of flooding, rated the riskiness of living in the house much higher than those with no previous experience
36
empirical support for availability heuristic - Facione (2002)
- survey of 770 women regarding cancer screening behaviour and perceived risk of developing breast cancer
- availability heuristics based on direct or indirect experience influenced risk perceptions:
--> women with history of benign breast disease or with a female relative with breast cancer or both overestimated their risk
37
optimism bias
- bad things happen to other people
- risk of my spouse being affected is higher than the risk of me being affected
--> defensive denial = protects from worry
--> downward comparison = “compared to people really at risk...”
--> egocentricism = “Ah, but I take precautions!” [and others don’t?]
- optimism bias = more likely to engage in risky behaviour
38
empirical support for optimism bias - Dejoy (1989)
- college-age drivers compared their risk of being involved in a variety of described traffic accidents relative to their peers
- optimism bias:
--> believed in almost all situations they would be at less risk than other drivers
39
empirical support for optimism bias - Weinstein et al (2005)
- a US national telephone survey of current smokers asked about their own risks of cancer vs the risk of the average smoker
- optimism bias:
--> believed they had a lower risk of developing lung cancer than the average smoker
--> perceived risk of cancer did not increase proportionally with the number of cigarettes smoked per day
--> agreement with several myths (e.g. that exercise undoes most smoking effects)
40
the psychometric paradigm
- suggests risks perceptions are based on a range of dimensions
- the aim is to unveil the dimensions that determine risk perceptions in order to create a taxonomy for hazards that can be used to understand and predict estimates of risk
- look at correlations between rated riskiness and dimension scores to answer the question: --> which dimensions seems to matter in to predicting our estimates of risk?
- massively influential
41
9 dimensions used in early psychometric paradigm
1. Voluntary vs involuntary
2. Chronic vs catastrophic
3. Common vs dreaded
4. Certain not fatal vs definitely fatal
5. Known to those exposed vs unknown
6. Immediate effects vs delayed
7. Known to science vs not known to science
8. Not controllable vs controllable
9. New vs old
42
psychometric paradigm (continued)
- factor analysis of dimensions typically identifies two factors
1. dread
--> e.g. lack of control, fatal, risk increasing, involuntary
2. Unknown
--> e.g. new, unknown to science/those exposed, effect delayed
- dread shows strongest correlations with global estimates of risk
43
empirical support for psychometric paradigm - Jenkins, Harris & Osman (2020)
- nationally representative UK sample
- rated 11 old and new food hazards on a total of 12 risk dimensions
--> e.g. newness, severity, immediacy
- principal components analysis identified two main components:
--> ‘dread’ and ‘knowledge’
--> explained 80.8% of the variance in perceptions
44
critical evaluation of the psychometric paradigm
- basic gist of psychometric paradigm replicated many times
- different dimensions matter in different situations
--> but agreement that Dread and Unknown are important
- factor structure is pretty stable
- factors account for risk perceptions pretty well
--> studies in the tradition of the psychometric paradigm typically explain 70% or even more of the variance in risk perceptions
45
criticisms of psychometric paradigm
- works best on aggregate level data (i.e., comparing different risks among groups)
--> poorer associations at individual level or for individual risks
--> akin to “ecological fallacy”
- replication of correlations and factors could be due to artefacts in the question wording (they are pretty similar)
- “Risk” is often targetless
--> risk to me? to my family? to non-specific others?
- different properties, and different dimensions, seem to matter in different studies
46
social amplification of risk
- brings together previous fragmented theories to better understand how risk information is disseminated and changed through different actions and interactions
- outlines how communications of risk events pass from the sender through intermediate stations to a receiver explaining how risks are amplified or attenuated
- the main thesis of SARF states that risk events interact with individual psychological, social and other cultural factors in ways that either increase or decrease public perceptions of risk
47
routes in social amplification of risk
- sources of information
- information channels
- social stations
- individual stations
- institutional and social behaviour
- ripple effects
- impacts
48
empirical support for the social amplification of risk - Frewer et al (2002)
- investigated the effect of increased media reporting about the risks associated with genetically modified food on public attitudes to the technology
--> first survey in February 1998 before media attention to genetic modification peaked
--> second survey in March 1999 when media coverage was at its peak
--> third survey in July 2000, when the level of media attention was subsiding
- perceptions of risk associated with genetically modified food increased during the highest levels of media coverage
- BUT risk perceptions subsequently reduced as coverage diminished
49
critical evaluation of social amplification of risk
- strengths:
--> Can explain why certain risks experts characterize as small “produce massive public reactions”
--> Helps to clarify phenomena, such as the key role of the mass media in risk communication and the influence of culture on risk processing
- limitations:
--> Amplification may be directed as to what are regarded as ‘exaggerated’ risks
--> May be too general to test empirically and particularly to seek outright falsification
--> Formulated 30 years ago (before online media environment we know today, studied within more traditional media)
