1. RISK HOMEOSTASIS: Theory, Evidence and Practical Implications*
07/16/96
RISK HOMEOSTASIS: Theory, Evidence and Practical Implications
Gerald J.S. Wilde, Ph.D.,
Professor Emeritus of Psychology
Queen’s University, Kingston, Ontario, Canada
No notes
Presented at Road Safety Fundamentals and Perspectives, University of Palermo, Sicily, Feb , 2009 *

2. *
07/16/96
There seem to be two common, but conflicting and contradictory philosophies of accident prevention.
The first is: “The accident loss can be reduced by protecting drivers from risky (inattentive, etc.) behaviour.” This is depicted in Slide #1. Note also that the concept of the seatbelt is more than a century old!
The second philosophy of accident loss reduction consists of “making the consequences of risky (inattentive, etc.) behaviour worse.”This is depicted in the cartoon in Slides #3 and #4. If you laugh or think this is ridiculous, realize that there are several such devices in the environment, like speed bumps or barbed wire. These seem to be designed to “scare people into safe behaviour.” Note also that both accident prevention strategies along the two lines seen in the cartoons so far are often implemented by one and the same safety authority.
Can one have it both ways? Or is this a case of superficial thinking? 2 *

3. *
07/16/96
No notes 3 *

4. *
07/16/96
No notes 4 *

5. *
07/16/96
Then there is a third view. This says that people will adjust to changes in the environment such that the risk they incur remains very much the same. In this cartoon one sees a driver who obeys the law and buckles his seatbelt. Notice that – as he does so – the right foot goes down a little heavier on the accelerator. He thus offsets the added security offered by the seatbelt by an increase in speed. When one feels better protected there is less need for caution.
Experiments carried out in the 1990s have shown that, when drivers who normally do not use the seatbelt are told to do so, they increase their driving speed and follow more closely the car in front (Janssen, W. (1994). Seatbelt wearing and driving behavior: an instrumented-vehicle study. Accident Analysis and Prevention, 26, ). 5 *

6. *
07/16/96
The cartoon on this slide and the next give a more exuberant demonstration of the same idea as Slide #5. Copied from a Dutch newspaper 6 *

7. *
07/16/96
No notes 7 *

8. Accident rates : Basic distinctions:*
07/16/96
Accident rates : Basic distinctions:
# per km driven
# per hour of road use
# per head of population per year
Important distinction : “safe cigarette”
# per cigarette smoked
# per smoker
# per head of population per year
If we wish to reduce the accident rate, we first of all must agree on the criterion of progress. What is it that we want: fewer accidents per unit distance driven (say per 100 million km) or fewer accidents per head of population? If we do not carefully distinguish between these two yardsticks of success, discussion about accident prevention becomes hopelessly muddled (as it of has been in the past).
Consider the example of the “safe cigarette”. Suppose current cigarettes are replaced by cigarettes that contain half the toxic and noxious substance as compared with the present cigarette. Is that progress?
Yes, because it allows double the smoking pleasure against the same frequency of smoking-related disease and death.
No, there will be no progress if smokers smoke twice as much. Then the smoking-related death rate per smoker will not go down. Moreover, the new cigarette may lead fewer smokers to kick the habit, because it is not as dangerous as it used to be. For the same reason more current non-smokers may start the habit. If this were to happen, the smoking-related death rate in the population will go up. In other words: there is no such thing as a safe cigarette; safety – if there is any – is in people, in their response to the product. 8 *

9. What is more important? per km driven? per head of population?*
07/16/96
What is more important?
per km driven?
per head of population?
to reduce accidents
Both objectives are perfectly legitimate. Reduction in the accident rate per km means greater benefit against the same loss.
Reduction in the accident rate per capita means a reduced loss in human life. 9 *

10. Sometimes obvious: # per km of available rope?*
07/16/96
Sometimes obvious:
# per km of available rope?
# per cubic meter of gas?
# per head of population
suicide rate
Suppose we were in the business of trying to reduce the suicide rate in the nation. It is obvious that our criterion of success would be a reduction in the per capita suicide rate, not in the reduction of the frequency of suicide by a particular means of killing oneself.
A similar objective apples to efforts to reduce the homicide rate.
# per pistol?
# per capita
murder rate 10 *

11. What is the goal of occupational accident/injury prevention?*
07/16/96
What is the goal of traffic accident prevention?
1. More mobility per traffic death ?
(economic gain)
2. Fewer injuries/deaths per year?
(public health gain)
What is the goal of occupational accident/injury prevention?
1. More productivity per case of injury/death?
2. Fewer injuries/deaths in the workforce?
(example: abdominal belts)
In road accident prevention, the first objective is the one that is often used in traffic safety engineering.
The second is the dominant objective in efforts to promote public health. *

12. Estimates of occupational traffic deaths:*
07/16/96
Estimates of occupational traffic deaths:
When commuting accidents are excluded (Bureau of Labor Statistics (USA, 2005):
43% of all occupational deaths occurred in transportation,
24% on highways alone.
When commuting accidents are included (world-wide estimates of the Int’l Labour Org. (Geneva; Takala, 1999; a total of 493,000 deaths were considered):
19% of all deaths occurred during driving for work,
32% while employees were commuting,
Thus, no less than one-half of all employee deaths occurred on the road: Road deaths are a major concern in occupational safety. *

13. Deaths per 100 million miles*
07/16/96
Traffic death rate per distance travelled, traffic death rate per capita, and the road distance travelled per capita in the U.S.,
Deaths per 100,000 residents.
Miles (x1000)
per resident
Deaths per 100 million miles
The relevance of the distinction between the two types of accident rates is clearly demonstrated by this graph.
The death rate per 100 million miles driven dropped roughly by a factor 10 in three quarters of the 20th century.
The death rate per 100,000 residents shows plenty of fluctuation over the years, but no dominant upward or downward trend.
So, there has been progress against one criterion, but only stagnation against the other.
The Ministry of Transport has reason for satisfaction, but The Ministry of Health laments the fact that we still have the same number of people in the morgues or in hospital beds.
In 1987 the fatal traffic accident rate per capita was about the same as it had been in 1927;
In 1996 it was about the same as in 1923 (National Safety Council, various years).
In the absence of a dominant upward or downward trend in the course of the larger part of this century, there have, however, been major fluctuations in the annual traffic death rate per head of population; from a low of 16.1 per 100,000 residents to a high of 30.8 in the time period considered. 12 *

14. *
07/16/96
In this 17-year period in Ontario, we see a marked drop in the accident per per mile, but at the same time a positive trend in the death rate per capita.
This was a period of economic growth.
Traffic deaths per distance driven and per capita, and distance driven per capita in a period of economic growth; Ontario 13 *

15. *
07/16/96
Annual variations in the unemployment rate and the traffic death rate per capita in the USA,
top profile: traffic deaths per 100,000 inhabitants
Bottom profile: unemployment rate in %; r = -.68
USA
1974: oil crisis
The importance of the economy (business cycle) can also be seen in this graph. In virtually every year in which the unemployment rate dropped, the per capita death rate on the road showed an increase, and in virtually every year in which the unemployment rate increased, there was a drop in the per capita traffic death rate.
Note the effect of the “energy crisis” in 1974. 14 *

16. *
07/16/96
MODELLED
DEADTRAF
DEADTRAF
ACTUAL
That same relationship has been observed in many countries, including Canada. Here we have the case of Switzerland. Using quarterly changes in the index of industrial production to model (predict) the death rate on the road, we found that the two curves are virtually indistinguishable.
It is remarkable that this should be true for a country as small as Switzerland, with some 600 road deaths per year.
Actual (DEADTRAF ACTUAL) and modelled (MODELLED DEADTRAF, i.e., predicted by the ARIMA procedure) traffic death rate per 100 residents in Switzerland on the basis of the index of industrial production; quarterly data (after Wilde and Simonet, 1996) 15 *

17. Chile: Blue: jobless rate – Red: traffic deaths per capita*
07/16/96
Here we have data from a similar analysis of Chilean data on the fatal traffic accident rate per head of population in that country related to annual variations in the rate of joblessness. The pattern is not as clear as in the case of the USA and Switzerland, yet it is clearly discernable.
Chile: Blue: jobless rate – Red: traffic deaths per capita
Source: Francisco Fresard (personal communication, 2007) *

18. The Target Level of Risk*
07/16/96
The Target Level of Risk
Four determining factors:
The expected advantages of risky behaviour alternatives:
Examples: gaining time by speeding, making a risky manoeuvre to fight boredom, rush production to meet a deadline, trying to catch up after having been delayed. +
The expected costs of risky behaviour alternatives:
Examples: automobile repair expenses after an accident, insurance surcharges for being at fault in an accident, equipment wear and tear. -
The expected benefits of safe behaviour alternatives:
Examples: an insurance discount for accident-free driving, building a reputation of responsibility. -
The influence of the economy can be understood when we realize that in good economic times more money can be made by driving faster and by driving more kilometres. It makes thus sense that the degree of exposure to accident risk is enhanced.
Moreover, the cost of risky behaviour is reduced as car repairs, insurance surcharges for accidents, and wear and tear on the vehicle and gasoline are all comparatively low relative to disposable income.
The slide indicates that the economy, although a dominant factor, is not the only factor determining the amount of risk people are willing to take. Some of the other factors are psychological or cultural in nature.
The expected costs of safe behaviour alternatives:
Examples: using an uncomfortable seatbelt, being called a wimp by one’s peers, time loss. + 16
+ INCREASE IN TARGET RISK
- DECREASE IN TARGET RISK *

19. *
07/16/96
As the level of exposure to risk increases, so does the expected gain, and so does the expected loss. The challenge is, therefore, to choose a level of exposure to risk for which it is true that the expected net benefit is the greatest. That is the most sensible, rational, level of risk.
Of course, anybody can reduce the chances of a road accident to zero by refraining of making use of the roads, but that person would also lose out on all the benefits that mobility can offer.
As there is not a single behaviour with total certainty of outcome, all behaviour is risk-taking behaviour. “ZERO RISK” is a meaningless objective. The challenge, instead, is to be neither too cautious, nor too prudent.
Theoretical representation of road users as net benefit maximizers and thus as risk optimizers. They choose an amount and manner of mobility such that the associated level of subjective risk corresponds with the point at which the expected net benefit is maximal. This is the smartest level of risk! (Note that the curve y3 has been drawn so that each y3 value equals the corresponding value y1 minus the corresponding value y2 absolute.) 17 *

20. *
07/16/96 1 2
We do not normally arrive at a decision as to what is the right amount of risk by carefully calculating the potential costs and benefits and their relative frequencies, if it were only because these costs and benefits are unknown. We must act on our intuition of what is the right amount of risk.
Te owner of the bicycle at the top has parked his vehicle with no lock or chain. So, the chances of him losing the bike to theft are quite high. The lady of the bike at the bottom is very averse to the risk of theft; she uses many chains and locks to secure her bike. But the poor lady has very little time left for anything else in life! We would say that she is overly cautious.
The lady of the bike in the middle picture seems to be the risk-smartest of the three bike owners. She provides a reasonable amount (but less than complete) of security to her bike, without being obsessive or compulsive about safety from theft. 3 18 *

21. Opportunities Missed There was a very cautious man*
07/16/96
Opportunities Missed
There was a very cautious man
Who never laughed or played;
He never risked, he never tried,
He never sang or prayed.
And when he one day passed away
His insurance was denied;
For since he never really lived,
They claimed he never died!
Here is another demonstration of the “danger of zero risk”
As Wildavsky has said: Not taking any risk may be the greatest risk of all. 19 *

22. Mrs. Cautious Driver: a case of driving phobia*
07/16/96
Mrs. Cautious Driver: a case of driving phobia
Used seatbelt before the law, always drove very slowly
4 collisions in 4 years
Not at fault in any of them
Likelihood much lower than I in 10,000 drivers
Very conservative in gap acceptance
3 times hit from behind, the last time by a police car
Interesting paradox: “dangerously cautious”
Consider the case of “Mrs. Cautious Driver.” This is the name we will give to a lady whose problems were presented at the Grand Rounds in Psychiatry in the Kingston General Hospital in Ontario in the mid-’70s. She suffered a serious nervous breakdown and an acute case of driving phobia. This is her story.
In the course of four years she experienced four traffic accidents. This is a rare occurrence, with a frequency of less than one in 10,000 drivers. In none of these accidents was she at fault. That makes her predicament a much rarer occurrence still. In all cases she was driven into at an intersection by another car. She was an extremely cautious driver in the sense that she drove her station wagon well below the speed limit on four-lane highways and she always buckled her seatbelt in a period before this was compulsory by law. At stop and yield signs, she was in the habit of waiting very long before she would accept a gap wide enough to her liking and go ahead. In such circumstances she would sometimes brake and stop again. Three times she was rear-ended in this situation, the third time by a police patrol car.
Mrs. Cautious Driver presents an interesting paradox. She was very careful indeed, in fact, so careful that her behaviour was rather unpredictable to other drivers. This made her liable to having accidents. On the other hand, if everybody were to behave as cautiously as she, there would be fewer crashes. (from Target Risk 2, p. 189) *

23. *
07/16/96
It is obvious that the accident rate in a nation depends on the level of risk that the people in that nation is willing to accept (the “Target Risk”). What this diagram proposes is that the accident rate in turn affects the perceived level of risk, and thus road-user behaviour.
So, behaviour influences accident occurrence and accident occurrence influences behaviour in a circular process of causation.
Homeostatic model relating the accident rate per head of population in a jurisdiction to the level of caution in road-user behaviour and vice versa, with the average target level of risk as the controlling variable. 20 *

24. *
07/16/96
You are familiar with this circular mechanism through the functioning of the thermostat in you house or room. The actions of the thermostat (comparing the actual temperature with the desired temperature and kicking the furnace in our at when there is a discrepancy) control the temperature in the room and the temperature in the room controls the thermostat. When averaged over time, the temperature is ultimately determined by one factor only: the desired temperature. This is turn is chosen as a function of the costs and benefits of setting the target temperature either higher or lower.
The operating principles may be explained as follows:
Box 1: You, the user of this control system, consider various factors in determining the preferred temperature. The temperature preferred usually is a compromise between the degree of physical comfort you ideally wish, on the one hand, and the cost of the energy needed for heating or cooling, on the other.
Box a: The preferred temperature is set on the thermostat control; this is called the set-point variable. It is a variable, because you have the choice between an entire range of set points. If energy costs go up, you are likely to choose a different compromise between considerations of comfort and cost, and you set the desired temperature to a different level.
Box b: The thermostat control continuously compares the actual temperature reading of the thermometer with the set-point temperature; this comparison is made at a point in the regulating process that is called the comparator or summing point.
Box c: Whenever there is a discrepancy (symbolized as [a-b]) between the thermometer reading and the set point, and this discrepancy is greater than a given tolerance of, say, 2% to 5%, the generator of warm air (furnace) or cool air is activated. The purpose of this is to keep the difference between a and b close to zero and this is achieved through a temperature-sensitive switch that tells the unit to produce either warm air or cool air, or to do nothing at all.
Box d: In order to adjust the house temperature to the set point, the air being forced into the house is somewhat warmer than the set point in the case of thermostatic heating, and somewhat cooler in the case of air conditioning.
Box e: As a result of this adjustment action, the house temperature is changed in the direction of the set-point temperature.
Symbol f: Because the thermostat control is usually (and for an obvious reason) not located in the vicinity of the air vents, and be­cause it takes some time for the altered air temperature to diffuse throughout the house and to finally reach the location of the thermometer, there is some time delay between the production of the adjusted house temperature and the reading on the thermometer. This brings the process back to Box b and starts another adjustment cycle. Hence the term “closed loop.”
Homeostatic model relating house temperature to heating system activity and vice versa: relating heating system activity to house temperature, with the set-point (target) temperature as the controlling variable 21 *

25. *
07/16/96
No further notes
Various amplitudes and wavelengths of fluctuations of homeostatically controlled variable (solid curves) around a value that is stable when averaged over time (dotted line). 22 *

26. Accident Causation as a Closed-Loop Control Process*
07/16/96
Accident Causation as a Closed-Loop Control Process
This is the simplest possible way of presenting the essence of Risk Homeostasis Theory.
The first implication of this model is that, at any point in time where the past accident rate is lower than the level of risk that people are willing to accept, road users will subsequently adopt a riskier manner and/or amount of mobility. The second implication is that they will do the opposite when the past record, and the personal experience associated with it, exceeds the preferred or target level of accident risk.
The first of these implications of risk homeostasis theory provides an explanation for what happened in Sweden and Iceland when these countries changed from left-hand to right-hand traffic at an early morning hour in the late 1960s. To the great initial surprise of many—including experts, laymen and politicians in Sweden and Iceland—the traffic accident rate per head of population dropped immediately and considerably after the change-over, but it subsequently returned to pre-existing trends, within two years in Sweden and, in Iceland, after about ten weeks.
According to risk homeostasis theory, these findings may be explained as follows. Because of the change-over’s major impact and fear-arousing interference with existing skills and habits, road users in these countries at first overestimated the level of accident risk that it would create. The thought of having to get up the following morning and drive on the opposite side of the road made drivers very apprehensive. Some road safety experts expected disastrous consequences.
Thus, the perceived level of risk surged to an unusual level that far exceeded the target level of risk. As a result, Swedish road users took unusually cautious adjustment actions, which in turn caused an unusual dip in the accident rate. During the 12-month period after the change-over date there was a 17% reduction in the number of traffic fatalities as compared to the preceding 12 months. After some time, however, the Swedes discovered, through their individual experiences and reports in the news media, that the new situation was not as dangerous as they had thought. Thus. the perceived level of risk went down, coming closer and closer again to the target level of risk. Consequently, the perceived need for prudent adjustment declined, cautious actions became less prevalent, and the accident rate re­turned to normal.
[1]Wilde, G.J.S. (1982). The theory of risk homeostasis: implications for safety and health. Risk Analysis, 2,
Sometimes adaptation to new conditions is much faster and the accident rate hardly changes at all:
The Munich taxicab experiment (Aschenbrenner, M. and Biehl, B. (1994). Improved safety through improved technical measures? Empirical studies regarding risk compensation processes in relation to anti-lock braking systems. In R. M. Trimpop and G.J.S. Wilde, Challenges to accident prevention: The issue of risk compensation behaviour. Groningen, the Netherlands: Styx Publications).
Part of a taxi fleet in Munich was equipped with an anti-lock brake system—also known as ABS. This type of brake system prevents the wheels from locking up under extreme braking conditions. It offers the advantage of improved steering control over the vehicle during rapid deceleration, especially on slippery road surfaces. The system makes it possible to change the direction of the car and abruptly reduce speed at the same time, at a considerably reduced risk of losing control. ABS brakes offer a perfect example of what was called a change in intrinsic risk towards the end of Section 3.4 in “Target Risk 2” a change in the objective accident loss expected if drivers don’t change their behaviour when a “safer vehicle” is made available. However, according to risk homeostasis theory, drivers are expected to change their behaviour and to maintain their accident likelihood per hour of driving as long as the target level of risk is not altered.
The cars with and without ABS in that Munich taxi fleet were of the same make and identical in all other respects. The majority of cab drivers were randomly assigned to one or the other of the two types of cars and the remaining drivers rotated between driving one type or the other. The exposure to traffic of each of the ABS taxicabs was carefully matched with cabs with traditional brakes over a period that lasted three years. Due to the matching procedure there was no difference in the time of day, the day of the week, the seasons, and the weather conditions in which both types of cabs were in operation.
Among a total of 747 accidents incurred by the company’s taxis during that period, the involvement rate of the ABS vehicles was not lower, but slightly higher, although not significantly so in a statistical sense. These vehicles were somewhat under-represented in the sub-category of accidents in which the cab driver was judged to be culpable, but clearly over-represented in accidents in which the driver was not at fault. Accident severity was independent of the presence or absence of ABS.
In another part of their investigation, the researchers installed accelerometers in ten ABS and ten non-ABS cars, without the drivers’ knowledge. These sensors measured the g-force of acceleration and deceleration once every ten milliseconds for a total of 3276 hours of driving. It was found that extreme deceleration, that is, extremely hard braking, occurred more often in the vehicles with ABS.
A third part of the study consisted of observations of driving style. Observers were trained in the systematic observation of a person’s driving style and in recording their evaluations on rating scales. They were then instructed to call a taxi and to observe the traffic manners of the driver while they were passengers. A total of 113 such trips were made, 57 in cabs with ABS and 56 in cabs without. All trips covered the same 18 km route. Speed measure­ments were taken at four pre­determined points of this route.
The drivers were not aware that their driver behaviour was be­ing observed and the observers did not know whether they were in a taxi with ABS or without. The drivers did, of course, know whether or not they were operating an ABS cab, because of their familiarity with the car they were driving.
Subsequent analysis of the rating scales showed that drivers of cabs with ABS made sharper turns in curves, were less accurate in their lane-holding behaviour, proceeded at a shorter forward sight distance, made more poorly adjusted merging manoeuvres and created more “traffic conflicts”. This is a technical term for a situation in which one or more traffic participants have to take swift action to avoid a collision with another road user.[1] Finally, as compared with the non-ABS cabs, the ABS cabs were driven faster at one of the four measuring points along the route. All these differences were significant.
In a further extension of their study, the researchers analysed the accidents recorded by the same taxi company during an additional year. No difference in accident or severity rate between ABS and non-ABS vehicles was observed, but ABS taxis had more accidents under slippery driving conditions than the comparison vehi­cles. A major drop, however, in the overall accident rate occurred in the fourth year as compared with the earlier three-year period. The researchers attributed this to the fact that the taxi company, in an effort to reduce the accident rate, had made the drivers responsible for paying part of the costs of vehicle repairs, and threatened them with dismissal if they accumulated a particularly bad accident record.
To sum up, in response to the installation of ABS brakes, drivers changed their driver behaviour. First, they utilized ABS to their advantage, but no improvement in the accident loss per time unit of exposure to traffic could be seen. Second, regardless of whether they were driving with or without ABS, a reduction in the accident rate did occur when the drivers’ target level of risk was reduced by measures increasing their expected cost of risky behaviour.
The Munich taxicab experiment attracted a great deal of attention, not only in the professional circles, but also in the popular press. Newspapers carried articles about it and Bavarian Television wanted to show the viewers what had happened. As the experiment had already been completed, they decided to re-enact the experimental manipulation and the way the drivers had responded. Airing of this documentary added further to the popular debate. The results of this experiment were also discussed by a group of international experts from the Organisation for Economic Co-operation and Development, commonly abbreviated as OECD. In their final report, these experts from sixteen different countries stated that: “Behavioural adaptations of road users which may occur following the introduction of safety measures in the transport system are of particular concern to road authorities, regulatory bodies and motor vehicle manufacturers, particularly in cases where such adaptations may decrease the expected safety benefit.”[2]
[1]Hauer, E. and Garder, P. (1986). Research into the validity of the traffic conflicts technique. Accident Analysis and Prevention, 18,
[2]OECD (1990). Behavioural adaptations to changes in the road transport system. Paris: Organization for Economic Co-operation and Development; Road Transport Research, p. 5.
Note 1) Sweden 1967 and 2) Iceland 1968 23 *

27. *
07/16/96
This is a cross-sectional study. The researcher obtained data on the accident rate per million vehicle-miles (called spatial accident rate) in 40 different road sections in Detroit. He then undertook to measure moving speeds in these road sections. It can be seen from the graph that in locations where the accident rate per mile is low, moving speeds are high and that in road section where the danger of accident per mile driven is greater, people drive more slowly.
As can be seen in this figure, the author related the accident rate (A) per million vehicle miles of each of the road sections to total travel time per road section (T), in an exponential function instead of a simple linear one. Additional calculations show, however, that the non-linear component, reflected in the curved solid line, is not statistically significant from a linear one. It would seem that the data are in reasonable agreement with what the risk homeostasis theory would expect: where the spatial accident rate is half as high, people drive twice as fast. This expectation is presented by the dotted line in this slide.
In other words, the accident rate per time unit of exposure remains essentially constant from road section to road section. People adjust to the variations in spatial accident risk they perceive in such a manner that the actual accident risk constant over time and independent of where the driving is done. (See p. 59 in Target Risk 2).
Accident rates per million vehicle miles (m.v.m.) related to average total travel time per mile and moving speeds in various road sections of different road design (graph adapted after May, 1959). 24 *

28. Virginia Commonwealth University*
07/16/96
Journal of Law and Economics, 38, 1995, ARE DRIVERS OF AIR-BAG-EQUIPPED CARS MORE AGGRESSIVE? A TEST OF THE OFFSETTING BEHAVIOR HYPOTHESIS*
STEVEN PETERSON,
Virginia
Commonwealth
University
GEORGE HOFFER,
and
EDWARD MILLNER,
Virginia Commonwealth University
ABSTRACT
An increasing number of researchers have hypothesized that regulatory attempts to improve automotive safety through product design would be at least partially offset by driver behavioral changes. This article analyzes two independent data sets to test whether differences in driver behavior exist between cars equipped with air bags and those not so equipped. An analysis of an insurance industry generated data set reveals that relative injury claims increase following adoption of an air bag system. Since there is no indication that the increase diminishes over time, the results appear to be attributable to offsetting behavior as opposed to a sorting of auto buyers. Analyses of 1993 Virginia State Police accident reports indicate that air-bag-equipped cars tend to be driven more aggressively and that aggressiveness appears to offset the effect of the air bag for the driver and increases the risk of death to others.
Another case of compensatory behaviour in response to “improved vehicle safety”.
air-bag-equipped cars tend to be driven more aggressively and that aggressiveness appears to offset the effect of the air bag for the driver and increases the risk of death to others. 25 *

29. *
07/16/96
This slide in combination with the next present an interesting combination of findings. The slide above clearly indicates that the chances of surviving or not getting injured at all in an accident are much better when the seatbelt is in use than when it is not. These data were collected in Sweden before there was any obligation to wear the seatbelt.
Can one logically infer from this that mandating seatbelt wearing will reduce injuries and death?
No, not logically, because there remains the possibility that drivers will react to the obligation by driving faster or following more closely, or by driving more under inclement weather conditions, or whatever.
See the next slide for fatality rates before and after introduction of the seatbelt-wearing laws in a variety of countries.
Graph showing the effectiveness of voluntarily wearing seat belts in reducing driver fatalities and injuries at different impact speeds. Data base 28,870 accidents. 26 *

30. *
07/16/96
This figure clearly shows that the fatal traffic accident rate in the countries that introduced mandatory seatbelt wearing legislation dropped to levels well below of what had been experienced before. We should, of course, be warned that the economic juncture might have something to do with this (see earlier slides, and consider the effect of the energy crisis in 1973 and 1974).
But what this figure shows, too, is that traffic fatalities also decreased in countries without such legislation. In fact, the drop was even somewhat greater in the latter. Could this possibly have been due to a lulling effect ? Thus, could it be due to the fact that, in countries in which seat­belt wearing became mandatory, the public was told over and over again in mass media campaigns that “seatbelts save lives”? In other words, could it be due to the public coming to believe that wearing the seatbelt would give a greater safety advantage than it actually does?
(from Adams, J.G.U. (1985). Risk and freedom; The record or road safety regulation. London: Transport Publishing Projects).
Bars at top indicate dates on which law came into effect in different countries. (Adams, 1985) 27 *

31. R.C. James (photographer). Courtesy of the photographer.*
07/16/96
What you are likely to see here is a pretty random set of dots.
R.C. James (photographer). Courtesy of the photographer. 28 *

32. *
07/16/96
The first impression one typically gets from this fragmented figure is a meaningless pattern of black dots against a white background. The figure becomes meaningful, or reorganized, when the viewer is told that the pattern depicts a dog sniffing the ground.
The first impression one typically gets from this fragmented figure is a meaningless pattern of black polygons against a white background. The figure becomes meaningful, or reorganized, when the viewer is told that the pattern depicts a dog sniffing the ground. 29 *

33. *
07/16/96
Here you have the slide again without the highlighting of the dog. Most of you will now be able to see the dog.
The significance of the last three slides lies in their demonstration that – once you have learned that something is there – you will actually see it. “Seeing is believing”, but also: believing is seeing. If you know something is in your fridge, you will eventually find it, though it may be hidden behind something else.
The earlier publications on Risk Homeostasis has a similar effect. More and more researchers saw its operation in their findings.
The next few slides give some examples.
The first impression one typically gets from this fragmented figure is a meaningless pattern of black polygons against a white background. The figure becomes meaningful, or reorganized, when the viewer is told that the pattern depicts a dog sniffing the ground. 30 *

34. Drivers keep a shorter distance*
07/16/96
INSTITUTE OF TRANSPORT ECONOMICS(TØ).NORWAY
Behavioural adaptation to antilock brakes:
Drivers keep a shorter distance
Drivers of taxis with antilock braking systems (ABS) have shorter time headways than drivers of taxis without such brakes. This is one of the results from a new research programme concerning taxi drivers’ behavioural adaptation to airbag and antilock brakes, performed at the Institute of Transport Economics.
Road lighting increases safety [that is per km driven]
- But motorists drive slightly faster and pay less attention
The introduction of road lighting leads to drivers slightly increasing their speed and paying less attention. The traffic safety effect of road lighting will there-fore not be as significant as it could have been, according to a research report from the Institute of Transport Economics. Road lighting is nevertheless an effective means of reducing accidents in darkness.
No additional notes
Mandatory course of driving on slippery roads does not reduce the accident risk
The mandatory course of driving on slippery roads for drivers of heavy vehicles, which has been introduced in parts of southern Norway, has not resulted in a reduced accident risk. This is the conclusion in a report from the Institute of Transport Economics. Some of the analyses indicate in fact that the accident risk has increased as a result of the course.
NORDIC ROAD & TRANSPORT RESEARCH NO 31 *

35. Orande forskningsresultat: Trafiktränade barn löper*
07/16/96
Barn och trafik
Orande forskningsresultat:
Trafiktränade barn löper
störe olycksrisk
“Children with traffic safety training run a higher accident risk”
From VTI aktuellt (Linköping, Sweden), 4, 1997.
“Der helikopter wird mich schon herausholen”. Chancen und Risken neuer Methoden im Flugrettungswesen.
By Peter Donatsch in Berg ‘97, Alpenvereinsjahrbuch (Band 121), edited by Dr. P. Grauss, pp Austrian Alpine Association (ISBN ).
“The helicopter will get me out for sure.” Opportunities and risks of new methods in rescue operations by air.
No notes 32 *

36. Munich taxicab experiment with ABS Dutch seatbelt wearing experiment *
07/16/96
Further evidence
Munich taxicab experiment with ABS
Dutch seatbelt wearing experiment
US motorcycle helmet laws
US seatbelt wearing rate and accidents
Accident migration – German Autobahn
Accident metamorphosis – alcohol, BC
US flood protection and flood victims
Skydiver parachute ripcord
Michelangelo computer virus
Railway crossing visibility improvement
Re: 1, see Aschenbrenner, M. and Biehl, B. (1994). Improved safety through improved technical measures? Empirical studies regarding risk compensation processes in relation to anti-lock braking systems. In R. M. Trimpop and G.J.S. Wilde, Challenges to accident prevention: The issue of risk compensation behaviour. Groningen, the Netherlands: Styx Publications.
Re: 2, see Janssen, W.H. (1994). Seatbelt wearing and driving behavior: an instrumented-vehicle study. Accident Analysis and Prevention, 26,
Re: 3, see Branas, C.C. and Knudson, M.M. (2001). Helmet laws and motorcycle rider death rates. Accident analysis and Prevention, 33,
Re: 4, see Derrig, R.A., Segui-Gomez, M., Abtahi, A. and Liu, L.L. (2002). The effect of population safety belt usage rates on motor-vehicle-related fatalities. Accident Analysis and Prevention, 34,
Re:5, see Pfafferott, I. and Huguenin, R.D. (1991). Adaptation nach Einführung von Sicherheitsmassnahmen. Zeitschrift für Verkehrssicherheit, 37,
Re: 6, see Rockerbie, R.A. (1980). Counterattack: perspectives and assessment. Vancouver: Policy Planning Division, Ministry of the Attorney General.
Re: 7, see Lave, T.R. and Lave, L.B. (1991). Public perception of the risks of floods: Implications for communication. Risk Analysis, 11,
Re:8, see Napier, V. (2000). Open canopy fatalities and risk homeostasis: a correlation study. Department of Psychology, Western Oregon University, March 5.
Re: 9, see Sawyer, J.E., Kernan, M.C., Conlon, D.E. and Garland, H. (1999). Responses to the Michelangelo computer virus threat: The role of information sources and risk homeostasis theory. Journal of Applied Social Psychology, 29,
Re:10, see Ward, N.J. and Wilde, G.J.S. (1996). Driver approach behaviour at railway crossings before and after enhancement of lateral sight distances: An experimental investigation of a risk perception and behavioural compensation hypothesis. Safety Science, 22, *

37. *
07/16/96
From the editorial “Low Tar, High Toll” by Warner, K.E. and Slade, J. In the American Journal of Public Health, 82, 1992,
“Smokers regulate their nicotine ingestion, compensating for lower yields by smoking more cigarettes, puffing more frequently, and inhaling more deeply.”
“Survey evidence demonstrates that the public, and particularly smokers, perceive low tar and nicotine cigarettes as carrying less risk [….].”
“Combined, nicotine compensation and switching instead of quitting suggest the very real prospect that the existence of low tar and nicotine cigarettes has actually caused more smoking than would have occurred in their absence and thereby raised the morbidity and mortality associated with smoking.”
Back to the so-called “safe cigarette”! 33 *

38. *
07/16/96
If the average number of sexual relationships were to increase by a small amount, from six to seven, the number of different pathways by which any sexually transmitted disease might spread from one person to another would increase more than 17-fold - to 326,592
It has been estimated by Patrick Dixon in The Truth About Aids that condoms fail in their contraceptive function about one time in 12, and that their Aids-prophylactic failure rate is likely to be even higher. They make sex safer, but not safe.
According to Canadian Psychologist Gerald Wilde, it’s all to do with our personal risk thermostats.
Whether or not condom advertisements impede the spread of Aids depends on whether the decreases that condoms effect in the risks associated with a particular sexual encounter are offset by the increases that the advertisements stimulate in the numbers of encounters. If the advertising results in more people, more often, attempting manoeuvres with defective safety equipment that they would not try without it, it is likely to promote the spread of Aids.
No additional notes
The message seems to be that these relationship are all enjoyable, and safe so long as people use the advertised product. 34 *

39. *
07/16/96 *

40. Deaths and Property Damage*
Deaths and Property Damage
Due to Floods, USA,
07/16/96
Note: Katrina, New Orleans, 2005: at least 1,836 people lost their lives in the deadliest flood since the 1928 Okeechobee hurricane, and in terms of damage the costliest tropical cyclone in USA history. *

41. The four utility factors that determine the target level of risk.*
07/16/96
Some have called the theory of risk homeostasis “Wilde’s law of the conservation of misery, of the conservation of accidents.” Funny, maybe, but they were mistaken. There is nothing fixed about the target level of risk. It can be changed, just as you can change the target temperature setting on your thermostat.
The four utility factors that determine the target level of risk. 35 *

42. *
07/16/96
The accident rate here is stated in the form of accidents per 1 million km driven. Note the marked drop from 1956 to 1957, which could not have been due to a change in vehicle fleet or highway construction.
The fact that the green line dropped even deeper than the red-dotted one indicates that is was the more serious accidents in particular that were prevented by the incentive program.
Drop in accident rate and annual accident costs in German trucking fleet after institution of safe-driving incentive program in 1957 36 *

43. The “Triple E” approach versus the “Single M” approach*
07/16/96
The “Triple E” approach versus the “Single M” approach
E for Engineering
E for Enforcement
E for Education (i.e., skill training)
Two crucial issues here:
The ability to be safe does not imply the willingness to be safe. Safe behaviour depends upon the desire to be safe.
The ability to be safe vs the willingness to be safe
The desire to avoid a fine vs the desire to be safe 37 *

44. *
07/16/96
M for Motivation
“To protect people from the negative consequences of risky behaviour is to encourage risky behaviour”
“To offer people positive consequences for cautious behaviour is to encourage cautious behaviour”
To reward or to punish? Incentives vs disincentives
There are in principle two ways of motivating people:
1. Give them a reward for desirable behaviour, or
2. Punish them for undesirable behaviour.
The second is not a very successful approach 38 *

45. Problems with penalties/disincentives*
07/16/96
Problems with penalties/disincentives
1. The “self-fulfilling effect of attribution,” labelling effect;
2. Process controls can never be exhaustive; low enforcement rates; “accident migration” and “accident metamorphosis;”
3. Negative side effects of punishment: resentment, antagonism, sabotage, reactance (alias “forbidden fruit effect”).
Re: the self-fulfilling effect of attribution: labelling people with undesirable characteristics, and expecting that they will show them unless kept in check by the threat of punishment, may cause individuals to be­have as if they had these characteristics. To illustrate this by its classic example: suppose you pretend that a perfectly solvent bank is about to go bankrupt (your attribution), and that you spread that rumour. The rumour you spread may cause depositors to withdraw their funds, with the end effect that the bank develops solvency problems and ends up by going bankrupt. Similarly, the very imposition of a speed limit may provoke some people to drive faster than they otherwise would. It is well known that some drivers find pleasure in activating the electronic devices installed along high­ways that tell the speeding driver: “You are going too fast.”
“Accident migration” refers to accidents moving to a different location. Likely to happen when there is heavy police enforcement on a particular road. Drivers will choose and alternative trajectory. So, the traffic migrates, and so do the accidents with it.
“Accident metamorphosis” refers to a situation in which a the particular immediate cause of accidents is replace by another immediate cause. For instance, drunk accidents being replaced by sober accidents, without their overall frequency diminishing. 39 *

46. Sample features of safety incentive programmes*
07/16/96
Sample features of safety incentive programmes
Industrial employees, truck and van drivers, passenger car drivers, public transit bus drivers
Target Group :
So, we turn to the reward approach.
What follows on the next several slides is a digest of the literature of some 150 publications on the topic. These come from North-America, South-America, Europe and Russia.
workers/drivers only
workers, foremen, supervisors
and middle management
Scope: 40 *

47. Sample features of safety incentive programmes*
07/16/96
Sample features of safety incentive programmes
Nature of bonus:
cash, savings bonds, public praise, certificates of merit, merchandise, extra holidays, lottery tickets, insurance discounts/rebates, free driver’s licence renewal, savings stamps for merchandise
No additional nots
Type of bonus:
for individual performance only,
for team performance only,
for both. 41 *

48. Sample features of safety incentive programmes*
07/16/96
Sample features of safety incentive programmes
Condition for eligibility:
being accident-free, displaying specified safety behaviours,
a combination of both
Incubation period:
one month,
three months,
six months,
one year
“Incubation period” refers to the minimum period of time during which the employee has to remain accident-free.
Penalty for failure to report an accident:
yes, no 42 *

49. Sample features of safety incentive programmes*
07/16/96
Sample features of safety incentive programmes
Implementations:
incentive programme only,
or combined with other accident countermeasure (usually safety training/education)
No further notes
Programme evaluation:
sometimes high standard, sometimes weak methodology, sometimes absent 43 *

50. 1. Truck drivers (Germany): approx. 80% reduction in accidents, *
07/16/96
Some findings:
1. Truck drivers (Germany): approx. 80%
reduction in accidents,
2. City transit drivers (USA): 25-35% reduction,
3. Passenger car drivers (California): 22-33% reduction; effect even greater in drivers under 25,
4. Novice passenger-car drivers (Norway): 35%
reduction. *

51. *
07/16/96
Date of intervention
Derived from: Fox, D.K., Hopkins, B.L. and Anger, W.K. (1987). The long-term effects of a token economy on safety performance in open pit mining. Journal of Applied Behavior Analysis, 20,
Note: “token economy” is synonymous with “incentive program in force.” 44
The yearly number of work-related injuries, per million person hours worked, requiring 1 or more days lost from work. *

52. Conditions favouring incentive effectiveness*
07/16/96
Conditions favouring incentive effectiveness 1.
Managerial vigour and commitment 2.
Program designed in cooperation with target group 3.
Involve multiple levels in the organization
The importance of Point #2 should not be underestimated. An incentive can only be an incentive to the extent that is being experienced as an incentive. You cannot “impose” an incentive program as you can impose a set of penalties. 4.
Keep rules simple
Equitable judgement of responsibility for accidents (with appeal procedure in place) 5. 45 *

53. Conditions favouring incentive effectiveness*
07/16/96
Conditions favouring incentive effectiveness 6.
Reward accident-free performance, not some safe behaviour 7.
Choose attractive rewards 8.
Progressive rewards for longer periods of being accident-free
No further notes 9.
Insure that reward is being perceived as equitable
Insure that reward is being perceived as attainable
10. 46 *

54. Conditions favouring incentive effectiveness*
07/16/96
Conditions favouring incentive effectiveness
11.
Consider supplementing incentives with safety training
12.
Discourage under-reporting of (minor) accidents
13.
Strength peer pressure toward safe conduct
14.
Keep incubation periods short
Re Pint #12. This is the only known undesirable side effect of incentive programs.
Re: Point #13. This can be achieved by (additional) team bonuses for accident-free performance of the entire team.
Maximizing net savings versus benefit/cost ratio
15.
Provide a research and evaluation component
16. 47 *

55. General conclusions from incentive studies*
07/16/96
General conclusions from incentive studies
1. Lost-day case rate or doctor’s cases per 100,000 hours worked reduced by 50 to 80%
2. Benefit-cost ratios usually at least 2 to 1; ratios as high as about 25 to 1 have been reported
3. Effectiveness usually does not dwindle over time. Some programmes have been in effect over some 30 years without losing effectiveness
4. A company can make money on its safety incentive programme. Who is paying for the added safety?
No further notes. 48 *

56. Side effects of incentive programmes*
07/16/96
Side effects of incentive programmes
1. Negative: under-reporting of minor accidents
2. Positive: improved (company) morale, leading to more productivity and less personnel turn-over
The unreported accidents (unreported under the pressure of the incentive system) are minor ones. It is rather easy to hide a broken pinky, much more difficult to conceal a corpse. 49 *

57. Why are incentive programmes so effective?*
07/16/96
Why are incentive programmes so effective?
They enhance the expected value the future and hence the desire for safety and health. “Expectationism”
People can be expected to be more careful with their health and safety:
1. as they rate the value of their future higher than the value of present time,
2. as they more actively plan for the future.
Quoted from Target Risk 2, Section 12.1:
In connection with incentives for safe performance, it is of interest to note that they may be viewed as just one example of a wider class of interventions that hold the promise of reducing much more than the fatal accident rate per capita. They also hold promise for reducing lifestyle-dependent disease and death rates, as well as diminishing the level of violence in society.
Offering a person a reward for not having an accident in the future implies offering that person a reason for looking forward to the future with increased expectations. Therefore, it also amounts to motivating that person to be more careful with life and limb and to take the measures necessary to be alive and well when that future comes.
As an alternative to the traditional “Triple E” (Engineering, Education and Enforcement) ideology for increasing safety—an ideology that we have seen to be rather ineffective in reducing the accident rate per head of population— an ‘expectationist’ approach is being suggested for the purpose of reducing the accident rate per head of population.
Expectationism is the name of the preventative strategy for reducing the accident rate and lifestyle-dependent disease and death rate per head of population by enhancing people’s perceived value of the future.[1] It comes in two varieties, and these may be called “specific” and “general.” A specific expectationist strategy demands that a person fulfill a particular requirement at some future point in time, such as not having been at fault in a road accident or not suffering from alcohol-related cirrhosis of the liver, not suffering from smoking-related respiratory disease, or some other specific criterion of health. The “general” variety sets no detailed criteria; all a person has to do to receive the reward is to be alive at that future date at which the incentive has been promised to materialize.
Consider, as a very simple example of general expectationism, a society in which every citizen is promised that upon reaching the age of retirement, a sum of money that equals five or ten times the average annual wage, in addition to current pensions and old-age benefits. It would seem reasonable to expect that this prize would stimulate people to use and develop their survival skills in such a way that more people than presently is the case will reach that age and be fit enough to enjoy the bonus.
The amount of money to be paid out in this example may appear large, but the benefit may well be found to be larger still, if one considers the amount of money that could be saved. The savings would take the form of reduced costs of medical care, physical damage and disability compensation, as well as a reduction in the current economic loss due to forgone wages and a person’s contribution to the gross national product. Consider, too, the savings that would accrue to a nation if governments no longer relied on safety legislation, engineering technology, enforcement practices, and educational measures of various kinds. These represent large expenditures, but fail to reduce the accident rate per capita to a significant degree, as we have seen from the available evidence.
Additional benefit would come from a reduction in the social cost of violence. To be violent is to run a major risk to one’s life or health, because of the damage that the perpetrator may incur in attacking another person. So, expectationism also holds the promise of reducing violence, not just accidents and lifestyle-dependent diseases.
[1]Wilde, G.J.S. (1986). Beyond the concept of risk homeostasis: Suggestions for research and application towards the prevention of accidents and lifestyle-related disease. Accident Analysis and Prevention, 18, 50 *

58. *
07/16/96
Future-orientedness
Study participants: 628 undergraduate students at Queens’ University (Bjorgvinsson and Wilde, 1998).
Health and safety habits under study:
1. Safe driving
2. Regular seatbelt use
3. Not smoking
4. Healthy diet
5. Regular exercise
6. Moderate drinking
These habits were more common in people who:
place less value in present time,
place a higher value on future time, and
who have a stronger tendency towards future planning.
See for references to this study:
Björgvinsson, T. and Wilde G.J.S. (1996). Risky health and safety habits related to perceived value of the future. Safety Science, 22,
Björgvinsson, T. (1998). Health and safety habits as a function of the perceived value of the future. Doctoral dissertation, Department of Psychology, Queen’s University, Kingston, Ontario, Canada.
Other studies of incentives for wellness include:
Stevens, M.M., Paine-Andrews, A. and Francisco, V.T. (1996). Improving employee health and wellness: A pilot study of the employee-driven Perfect Health Program. American Journal of Health Promotion, 11,
Hennrikus, D.J. and Jefferey, R.W. (1996). Worksite intervention for weight control: A review literature. American Journal of Health Promotion, 10, )
Mavis, B.E. and Stöffelmayr, B.E. (1994). Multidimensional evaluation of monetary incentive strategies for wieight control. Psychological Record, 44, 51 *

59. “Simple Safety Song” Give me a ladder that is twice as stable,*
07/16/96
“Simple Safety Song”
Give me a ladder that is twice as stable,
And I’ll climb it twice as high.
But double the cause for caution,
And I will be twice as shy.
This is the shortest summary of risk homeostasis theory that I have been able to come up with. 52 *

60. Target Risk 2 (2001) cover; PDE Publications, Toronto; pde@drivers.com*
07/16/96
Target Risk 2 (2001) cover; PDE Publications, Toronto;
The second, enlarged, edition Contains about 15% more material than the first edition (NOT ACCESSIBLE ON THE WEB).
There are translations of Target Risk in Spanish:
Wilde, G.J.S. Riesgo Desado? El comportamiento humano ante el peligro. Versión al español: Ing. L. Daniel Ramírez I. Mexico City: Asociacion Mexicana de Higiene y Seguridad, 2001 (323 pages), ACCESSIBLE ON THE WEB: <
and partially in Russian (UNAUTHORIZED): ACCESSIBLE ON THE WEB: <
In Portuguese:Portuguese version (updated from Target Risk 2, translated by Reinier J.A. Rozestraten) : “O Limite Aceitável do Risco, Uma nova Psicologia de Segurança e de Saúde, O que funciona? O que não funciona? E por que...”(São Paulo, Brasil, 2005); for publisher check . (NOT ACCESSIBLE ON THE WEB)
In Japanese: translation, (by Shigeru Haga, Tokyo, updated from Target Risk 2), published by Shinyo-sha, Publishers, Tokyo, March < NOT ACCESSIBLE ON THE WEB. 54 *

61. *
07/16/96 1 2
Asociación Mexicana de Higiene y Seguridad, México, 2001; translation by Daniel Ramirez, P. Eng.:
2) Translation by Prof. Reinier Rozestraten, published by (São Paulo, 2005) *

62. Translation by Prof. Shigeru Haga*
07/16/96
“Target Risk 2” in
Japanese
translation,
Tokyo, Feb. 2007
"Why do traffic accidents keep on happening?"
With subtitle "The psychology of risky behaviour"
Translation by Prof. Shigeru Haga *

63. *
07/16/96
Time for me to stop. Thank you for your patience. I *

64. *
07/16/96
Surely, traffic accident are often tragic, but at times they can be rather funny. *

65. Good thing he is wearing a helmet*
07/16/96
Good thing he is wearing a helmet *