close-following drivers on two-lane highways

Pergamon

Accid. Anal. and Prev., Vol. 29, No. 6, pp. 723-729, 1997 0 1997 Elsevier Science Ltd

All rights reserved. Printed in Great Britain 0001.4575,‘97 $17.00 + 0.00

PII: s0001-4575(97)00041-9

CLOSE-FOLLOWING DRIVERS ON TWO-LANE HIGHWAYS

SIRPA RAJALIN,’ SVEN-OLOF HASSEL~ and HEIKKI SUMMALA~*

‘Central Organization for Traffic Safety in Finland, Sitratie 7, 00420 Helsinki, Finland, 2Mobile Police, Western Finland Unit, P.O. Box 51,201Ol Turku, Finland and 3Department of Psychology,

Traffic Research Unit, P.O. Box 13, 00014 University of Helsinki, Helsinki, Finland

Abstract-This study was intended first to replicate, on two-lane highways, of the Evans and Wasielewski (Accident Analysis & Prevention 14, 57-64, 1982; 15, 121-136, 1983) results on the connection between close- following driving and traffic offenses and, second, to reveal reasons for close-following. A sample of close- following drivers (N= 157) and control drivers (N= 178) was picked from the flow on two-lane main highways. The driver records of the past 3 years showed retrospectively that the close-followers had accumulated 2.3 times more traffic offenses than had the control drivers and 2.0 times more when mileage was taken into account. The result is in agreement with the Evans and Wasielewski results for multi-lane highways, with the additional check for mileage in these data. However, the effect only occurred in males and was more marked in young males. Close-following females even indicated a tendency of having fewer offenses than their controls when their higher mileage was taken into account. Another sample of close-followers interviewed on the road revealed that hurry or desire to overtake the car ahead was the justification for the close-following in the majority of cases. It was suggested that on two-lane highways close-following substantially stems from overtaking needs and maneuvering connected to higher target speeds. This study partly confirms the connection between close- following and an increased number of offenses in comparisons between drivers. However, the suggested connection between close-following and accident involvement, as based on interindividual comparisons, still remains somewhat open. 0 1997 Elsevier Science Ltd.

Keywords-Accidents, Age, Car following, Driver behavior, Gender, Speed, Traffic offenses

INTRODUCTION

Close-following is generally considered risky even to the degree that several countries now have laws against driving at short following distancies (e.g.

Schweitzer et al., 1995). On theoretical grounds, close-following, especially at high speeds in platoons is almost bound to lead to a disaster even if a minor disturbance propagates down the flow through cumu- lating response latencies and increasing response (brake) strength (Evans, 1991; Herman et al., 1959).

Other things being equal, short following distancies provide less time to react to a lead car’s braking or major disturbancies ahead, and therefore close- following is indeed usually connected with an increased accident risk in comparison to longer following distancies.

Efforts have been made to show the dangers involved in close-following driving by means of comparisons between drivers, and police increasingly apply an enforcement strategy of booking tailgaters

*Author for correspondence. E-mail heikki.summala@ helsinki.fi

in many countries. Evans and Wasielewski (1982)

showed, using photographs taken from a bridge over

a freeway to identify cars and drivers, that drivers who keep shorter distance on a freeway indeed have more often accumulated accidents and offenses. Of those who drove with a time interval of < 1 second, 49.7% had been involved in an accident and 65.4%

had at least one offense on their record. The corresponding figures for those who drove with a time

interval of at least 1 second were 44.4% and 57.9%. As to the violations, this result was confirmed with

more extensive data (Evans and Wasielewski, 1983),

both in univariate and in multivariate analyses, taking into account a set of relevant variables such as driver

age and sex, vehicle mass and model year. The

relationship between accident involvement and close- following was also replicated in univariate analyses but not in multivariate ones.

However, citations and accidents were not checked against the exposure (mileage or some related

measure), and therefore, we cannot necessarily con- clude from the Evans and Wasielewski studies that close-following must be risky and that tailgaters are

123

124 S. RAJALIN et al

more prone to accidents [see also Makinen and Kulmala ( 1987)]. Furthermore, a sample of repeated observations of the same drivers in the Evans and Wasielewski data showed a certain degree of consis-

tency in close-following which appears to be a relatively permanent driver disposition just as speed selection is (e.g. Hauer et al., 1982; Rajalin, 1994; Summala et al., 1984). As a matter of fact, these two dispositions may be closely related, at least on two- lane roads.

Higher accident and violation rates per driver

among close-followers may be due to higher mileage. A possible causal link simply goes through higher target (desired) speed level (Summala, 1996, 1997).

The variation in desired speeds necessarily results in the need to overtake slower vehicles. Drivers with a

higher target speed encounter slower vehicles and have to overtake more often. On two-lane roads particularly, close-following may largely result from this desire to overtake, and waiting for an opportunity to do so. Summala (1980) found that a long overtaking prohibition on a two-lane highway reduces short headways, and Summala and Vierimaa ( 1980) showed that the shorter the following distance, the more left the average lateral position is with respect to the car ahead. On multi-lane freeways, too, fast drivers often ‘force’ slower ones to leave the faster

lane by driving very close behind them. Our point is that people who drive more are also assumed to drive faster and collect more speeding (and other moving) tickets while they also follow other cars at shorter distance at least when preparing to overtake. Indirect support for this hypothesis comes from findings noted above that drivers show consistency both in following behavior and speed choice. These drivers may also

incur more accidents due to higher exposure, both in terms of mileage and of number of overtakings.

However, because accidents do not increase directly with mileage, this connection may not be as strong.

Therefore, it is important to also replicate the Evans and Wasielewski (1982, 1983) studies on two- lane roads, and check whether close-followers accu-

mulate more offenses when mileage is counted for. In the second part of this study, a sample of close- followers was interviewed about their motives, in order to test the hypothesis that close-following on two-lane roads is largely connected to overtaking.

STUDY I: DRIVER RECORDS OF CLOSE-FOLLOWERS

Methods A sample of close-following drivers was taken

from a normal traffic flow as well as a control sample from the same site at the same time. The four sites

used were located on two-lane main highways with speed limits of 80 and 100 km hour - ‘. The data were

collected by the police in conjunction with their routine monitoring of vehicle headways, by means of

the inductive loop system (Marksman 600) during April-May 1993 in the Finnish province of Turku- Pori. Car and van drivers who drove with a net time headway of 0.8 seconds or less were defined as being close-following drivers. The corresponding time headway for the control drivers was 225 seconds. This spans the lower limit of a recommended safe following distance and the upper limit for platooning in traffic

engineering (Highway Research Board, 1965). The police stopped the drivers, took the driver license number and asked about the number of kilometres driven over each of the past 3 years. These data were

obtained from 157 close-following drivers and 178 control drivers. The controls were taken from the flow as soon as the interview of the corresponding close-following driver was completed. For both groups, traffic offenses were counted from the driver records for the past 3 years. Records of accidents paid by insurance companies during the same period were taken from the data base of the Traffic Insurance Center.

Results Men accounted for 73% and women 27% of the

drivers in the group of close-following drivers. The

corresponding percentages for the control group were 77% and 23%. The gender difference between the two groups was not statistically significant. Table 1 shows, however, that there were relatively more younger people ( 1 S-34 years) and relatively less elderly people (> 54 years) among the close-following drivers.

Both the close-following and control drivers had driven quite a large number of kilometers each year- about half the drivers had driven > 25,000 km a year.

The self-reported yearly average was a little higher for close-followers (33,876 km) than for the control drivers (29,065 km), the difference being significant [F( 1,332)=4.05; p=O.O47].

Table 1. The number and percentage distribution of the drivers by age group

Close-following

Age group N %

18-24 17 10.8 25-34 48 30.6 35-54 83 52.9 55-81 9 5.1

Total 157 100.0

x2=10.44, df=3,p=0.015.

Control group

N %

14 7.9 40 22.5 96 53.9 28 15.7

178 100.0

Close-following drivers on two-lane highways 125

The close-following drivers had committed 162 traffic offenses altogether during the preceding 3-year period, compared with 79 for the control drivers (excluding offenses for tailgating). The distribution of drivers with regard to traffic offenses is given in Table 2. The close-following drivers had accumulated significantly more traffic offenses than had the control drivers [x2( 3) = 18.78, df = 3, p = 0.0003]. The percentage of drivers with more than two offenses on their driving license records was 26% for the close-followers but 9% for the control drivers. About half the drivers in the close-following group had not committed any offenses during the 3-year period. The corresponding proportion in the case of the control drivers was almost three-quarters. For the close-following group the number of offenses per driver was 1.0 per 3 years and 10.2 per million kilometres while the corresponding figures for the control drivers were 0.4 and 5.1 (see Table 3).

Gender It was found that the higher offense rate of close-

followers is only true for males drivers in these data [F( 1,330) = 6.98 for the Gender x Following condition interaction, p<O.Ol; Table 41. Close-following females even had somewhat fewer offenses per mileage than their controls.

For statistical tests of mileage-related effects, in these relative small data with zero offense counts included, a ‘conformity-to-traffic-law’ measure was computed for each subject indicating mileage driven

Table 2. Distribution of the drivers by number of accumulated traffic offenses

No.of offenses per driver Close-following (%) Control (%)

0 52.2 71.3 1 22.3 19.7 2 13.4 5.1 3f 12.1 3.9

Total 100 100.0

x2= 18.78, df=3, p=O.O003.

per offense during past 3 years:

Conformity Index = Mileage/

(No. of registered offenses + 1)

Bearing in mind that speeding was the most common offense in these data, this is a rough estimate of how well each individual driver conforms to speed limits, applicable to drivers with offenses as well as to those without any. The basic assumption is merely that, say, 50,000 km without offenses equals to 100,000 km with one offense and to 150,000 km with two offenses. Logarithmic transformation of this index results in approximately normal distribution applicable to a standard ANOVA.

A two-way ANOVA, Gender x Following condition, showed a significant interaction in mileage- related offenses [conformity index; F( 1,330) = 8.3 1, p = 0.0041 while separate one-way ANOVAs showed a significant following condition effect in males [higher offense rate in close-followers, E;( 1,249) = 5.16, p =0.024] and a marginal effect [lower offense rate in close-followers, F( 1,8 1) = 3.86, p = 0.0531 in females.

Age Age difference between close-followers and con-

trols also only occurred in males [means for the two groups of males 38.9 and 43.1 years, F( 1,249) = 7.86, p=O.O05, and of females 35.2 and 35.6 years, not significant]. Since the number of female drivers was too small to test possible differential age effects, only male drivers were included in a two-way ANOVA (Age x Following Condition), classified by age into three groups of approximately equal size. The effects of age [P(2,245)= 14.51, p<O.OOl] and following condition [E;( 1,249) = 5.16, p = 0.0241 were significant and their interaction marginally significant [F( 2,242) = 2.38, p = 0.0951. Figure 1 shows that close-followers have more offenses in all age groups while the youngest close-followers indicate a tendency to collect more convictions for a given mileage.

Table 3. The number of offenses during the past 3-year period as well as per driver and kilometers driven, by age group, separately for close-following drivers and control drivers

Close-following Control

Age No. of offenses Offenses per driver Offensesper lo6 km No. of offenses Offenses per driver Offenses per lo6 km

18-24 38 2.24 26.6 5 0.36 5.5 25-34 53 1.10 10.9 32 0.80 9.5 35-54 69 0.83 7.7 37 0.39 4.0 55-81 2 0.22 3.3 5 0.18 2.6

Total 162 1.03 10.2 19 0.44 5.1

126 S. RAJALIN et al.

Table 4. Mean number of offenses, mileage and mileage-related offense rate by gender and following condition

Males Females

Close-followers Far-followers Close-followers Far-followers

Mean number of offenses for 3 years 1.325 0.496 0.256 0.275 Mean yearly mileage (1000 km) 37.51 32.78 24.09 11.01 Mean number of offenses lo6 km per 11.76 5.04 3.54 5.37

- Close-followers * Far-followers

-30 31-45 46- yrs

Age group Fig. 1. Number of offenses as related to mileage by age group for

close-followers and for far-followers. Males only included.

Accidents There were only 25 insurance-covered accidents

in the whole sample, 10 among close-followers (0.063 per driver and 0.63 per million kilometres, as based on the estimates given for the before period) and 15 among control drivers (0.084 per driver and 0.97 per million kilometres). In comparison, the corresponding rates for all insurance-covered crashes were ca 0.1 per driving licence and 2.5 per million kilometres (Traffic Insurance Center, 1994). The accident data, although very limited, therefore indicated no evidence of any higher involvement in accidents related to

close-following.

STUDY II: DRIVER INTERVIEWS

Methods To reveal the reasons for close-following, another

sample of close-followers (N= 140) was collected in the same area. The headways were recorded by means of the inductive loop system. The police stopped drivers with a net interval of less than half a second,

and these were interviewed by the research assistants. To keep the interview data as unbiased as possible, the police neither checked driving licenses nor issued citations in this study.

The drivers were first told that they drove too close behind another vehicle and asked to explain why they did so. They were then given a list of factors

which might result in following too close (see Table 6) and were asked to say which made them do so.

Results The open answers were divided into five

categories:

(1)

(2)

(3)

(4)

(5)

Vehicle spacing ‘adequate’: the drivers did not

regard the distance observed (co.5 second) as too short. Overtaking intentions: the short vehicle interval was justified on the grounds that the driver was waiting for an opportunity to overtake.

Behavior of the other vehicle: the behavior of the other vehicle was put forward as the cause of the short interval, for example, the vehicle in front was driving too slowly or had decelerated sud-

denly to turn into a side road. Close-following driving had become a habit: the drivers said that they had got into the habit of close-following driving, or were unable to give any reason for their behavior. Other reasons: for example, traffic congestion, being in a hurry, sudden change of speed. Overtaking intentions were most often cited

(34% of the drivers) as the reason for close-following. Explanations such as waiting for an opportunity to overtake, preparing to overtake or looking for a place

to overtake, were spontaneously offered by these

drivers. The next two most commonly cited reasons for close-following driving were those associated with the behavior of the other vehicle, actually resulting in a need to overtake, and close-following having become a habit (Table 5).

Table 6 shows the structured answers given by drivers interviewed as reasons for their close- following. Apart from spontaneous reasons, not all of these are exactly reasons but rather excuses and explanations at several levels. However, Tables 5 and


Table 5. The distribution of reasons for close-following into the five categories, open answers

Vehicle spacing adequate Overtaking intentions Other vehicle Habit Other reasons

N %

10 7 47 34 33 24 21 20 20 15

Table 6. The reasons for close-following, structured answers

%

I am waiting for an opportunity to overtake 40 I follow closely when in a hurry 12 The risk of a rear-end collision is negligible 9 I have acquired a habit of close-following 9 I trust to my quick reactions 9 I want to prevent other drivers from merging in front of me 7 Close-following is not prohibited 6 Short distances reduce congestion 6 I show the driver ahead that he/she drives too slowly 2 I cannot estimate a safe distance 1 A short distance is safer than keeping too great a distance 1 I try to get the driver ahead to speed up 1 I consciously drive at the risk threshold Other reason 12 I do not usually follow closely 24

6 show that the majority of respondents give an explanation which is directly connected to speed level and overtaking aims. Thus, 40% of drivers say directly that the reason is overtaking, 12% say they are in a hurry, and 2% say they want to show the driver in front that he/she is driving too slowly; a total of 54%.

DISCUSSION

These results first confirmed the early Evans and Wasielewski (1982, 1983) finding that close-followers incur more offenses than other drivers. This study showed that this also applies to two-lane roads, and furthermore, that the effect cannot be explained by means of mileage. However, the effect only occurred in males. Female close-followers had accumulated approximately the same number of offenses as their controls, and even less when their higher mileage is taken into account. Although the data suggest that the effect is more marked in young males, it occurs in all age groups and, therefore, age does not explain it. As female close-followers did not differ in age from far-followers, age cannot explain the major gender effect either.

The other sample showed that when interviewed as to the reasons for close-following, the majority of drivers ascribe it to desire to overtake or hurry. This confirms earlier results which indicate a close relation-

ship between close-following and overtaking maneuvers.

The problem of close-following is conceived to be largely related to learning and adaptive processes that make drivers feel short safety margins as not hazardous (Evans, 1991; Naatanen and Summala, 1976; Summala, 1988). However, a very basic, higher- level reason for close-following, explaining a large part of very small headways especially on two-lane highways, appears to be the variation in desired speeds in the traffic flow which results in a need to overtake. Overtaking maneuvers often precede a prior-maneuver phase when drivers wait for an opportunity to overtake, and apparently stay too close to the vehicle ahead. [Another reason, as noted by Colbourn et al. (1978) may be that drivers aim to prevent other drivers from merging in front of them, but this explanation was not frequent among the reasons mentioned in the present interview: 7% of all the close-followers interviewed.] Higher target speed and the resulting overtaking need is connected to high mileage (Summala, 1987) which, as supported by these results, results in higher offense rate. It is to be noted that most offenses recorded in the Finnish driver records are for speeding, and thus reflect fast driving habits.

Close-following is now widely prohibited and enforced in highly motorized countries, as it is under- stood to be associated with increased risk, and great efforts are being put into developing intelligent distance warning and control systems. The latter should be effective in reducing the rear-end crash risk in steady-state following which stems from inattention and does not necessarily involve very small following distances. However, the role of overtaking should be taken into account in generating rather small headways. In addition to prohibition and enforcement, we should therefore consider road and traffic engineering and educational countermeasures which could remove the higher-level cause, at least on two-lane roads (Summala, 1996). Overtaking lanes and wide lanes are being used nowadays as inexpensive means to counter capacity problems (Harwood and Hoban, 1987). All measures which provide safe opportunities for overtaking without time pressures should also reduce the problem of close-following. Educational measures such as chevrons painted on the pavement have also been proposed and used against close- following (e.g. Hamalainen, 1993; Helliard-Symons et al., 1995).

However, there are two kinds of close-following, one temporary and situation-specific and another more consistent steady-state following, and we might hypothesize that whenever a faster driver reaches a slower vehicle, he/she chooses the headway differently

728 S. RAJALIN et al

depending on whether he/she plans to overtake the car or not. Frequent overtaking may lead to habitual close-following in other circumstances as well.

Our data showed a major gender effect, not in the rate of close-following but in offense rates associated with close-following. We hypothesize that this difference comes from a different attitude toward speeding and different behavior of males and females (e.g. Blockey and Hartley, 1995; Dahlstedt, 1994, pp. 60-64; Makinen, 1990; Parker et al., 1995). Although close-following is connected to higher mileage and, presumably, to higher target speeds also with females, they keep to the speed limits and accumulate fewer speeding tickets.

Interindividual variance among drivers certainly include other differences than age, driving experience and gender. For example, Heino et al. (1996) showed that sensation seekers (Zuckerman et al., 1979) follow a lead car more closely in a test drive than people who report avoiding extreme sensations. However, the former do not feel that their chosen distance is more risky than the latter, and a question remains how these two groups selected from the extreme ends of sensation-seeking continuum actually have acquired different following strategies. It may even be due to higher speeds which sensation seekers presumably use (Zuckerman and Neeb, 1980), if not impeded by the experimenter or other traffic, and due to resulting overtaking need and more frequent prepa- ration for overtaking, which of course may become habitual.

This study confirms the connection between close-following and increased number of offenses in terms of between driver comparisons. However, the respective connection between close-following and accident involvement still remains somewhat open. These accident data however, although being very limited, indicated no evidence that the close-followers would have more accidents whether they be related to close-following or not.

There are several reasons which may make it difficult to show the connection between close- following and increased accident involvement. Active overtakers-people who drive more and faster than the average and follow other vehicles at shorter distance when waiting for an opportunity to overtake-may be more alert, and may also have more practice in managing with slower vehicles (see e.g. Bjornskau, 1996) to the degree that their motor control to a decelerating lead vehicle is faster (van Winsum and Heino, 1996; van Winsum and Brouwer, 1997). ‘Safe’ far-followers may instead be less alert and may also share time and attention between driving and in-car tasks which results in problems in detecting lead car braking; as a matter of fact, the

detection of a lead car’s braking even gets more difficult with increasing distance (Lamble et al., 1996). The use of mileage-related accident figures also favor people who drive a lot because the connection between total mileage and accident involvement is not linear (Janke, 1991; Maycock et al., 1991; Ernvall and Pirtala, 1996). Therefore, research focused on driver differences should relate close- following to the relevant accident involvement to show its dangers while the possible connections with other accident types may reflect more general driving style differences (Summala, 1996).

Finally, a critical question arises: should we apply minimum following distances and should police enforce them if conclusive evidence for the dangers involved is lacking? The answer is definitely yes. Other things being equal, including attention, close- following very likely increases accident risk, while it also get drivers in front anxious and irritated (Hamalainen, 1988). In steady-state, habitual short following, increased attentiveness probably does not compensate for the risks of close-following. Even until more conclusive evidence is provided, prohibition of short headways and enforcement is needed in order to improve habitual following behavior.

Acknotvfedgements~The authors thank Juha Valtonen for help in the data analysis, the Traffic Insurance Center and Esa Nysten for providing the accident data, Liisa Hakamies-Blomqvist for comments on an early version of this paper, and three anonymous referees for their comments during the review process.

REFERENCES

Bjornskau, T. (1996) Why are the “safest” norms, attitudes and types of behaviour not typical for the safest drivers?. Transport Reviews 16, 169- 18 1.

Blockey, P. N. and Hartley, L. R. (1995) Aberrant driving behaviour: errors and violations. Ergonomics 38, 1759-1771.

Colbourn, C. J., Brown, I. D. and Copeman. A. K. (1978) Drivers’ judgments of safe distances in vehicle following. Human Factors 20, 1-l 1.

Dahlstedt, S. (1994) SARTRE-tabellerna. En redovisning av nigra europeiska bilfijrares isikter om trafik och trafiksakerhet (The SARTRE-tables. Opinions about traffic and traffic safety of some European drivers). Report No. VT1 403/403A. The Swedish Road and Traffic Institute, Linkiiping.

Ernvall, T. and Pirtala, P. (1996) Annual mileage as an exposure to accident risk. Second Road Safety Confer- ence, Bled, Slovenia, 12-15 June.

Evans, L. (1991) TrafJic Safety and the Driver. Van Nostrand Reinhold, New York.

Evans, L. and Wasielewski, P. (1982) Do accident-involved drivers exhibit riskier everyday driving behavior? Acci- dent Analysis & Prevention 14, 57-64.

Evans, L. and Wasielewski, P. (1983) Risky driving related to driver and vehicle characteristics. Accident Analysis & Prevention 15, 121-136.


Htimalainen, V. (1988) Vuorovaikutus, nopeus ja koetut riskit (Drivers’ Interaction, Speed, and Fears) The Central Organization for Traffic Safety in Finland, Helsinki, Finland.

Hamalainen, V. (1993) Opastuksen vaikutus ajoneuvoviilei- hin (The Effects of Chevron Road Markings on Vehicle Intervals). The Central Organization for Traffic Safety in Finland, Helsinki, Finland.

Harwood, D.W. and Hoban, C.J. ( 1987) Low-cost Methods for Improving Trafic Operations on Two-lane Roads. Fed- eral Highway Administration, Washington, DC.

Hauer, E., Ahlin, F. J. and Bowser, J. S. (1982) Speed enforcement and speed choice. Accident Analysis & Pre- uention 14, 261-218.

Heino, A., van der Molen, H. and Wilde, G. J. S. (1996) Differences in risk experience between sensation avoiders and sensation seekers. Personality and Individual Dtffer- ences 20, 71-79.

Helliard-Symons, R., Webster, P. and Skinner, A. (1995) The Ml chevron trial. TrafJic Engineering & Control 36, 563-567.

Herman, R., Montroll, E. W., Potts, R. B. and Rothery, R. W. (1959) Traffic dynamics: analysis of stability in car following. Operations Research 7, 86-106.

Highway Research Board (1965) Highway Capacity Manual, Special Report 87. Highway Research Board, Washington, D.C.

Janke, M. K. (1991) Accidents, mileage, and the exaggera- tion of risk. Accident Analysis & Prevention 23, 183-188.

Lamble, D.E., Laakso, M. and Summala, H. (1996) The effects of driving experience on drivers’ ability to detect braking and decelerating of a lead vehicle using peri- pheral vision. International Conference on TrafJic & Transport Psychology, Valencia, 22-25 May.

Makinen, T. (1990) Liikennerikkomusten subjektiivinen kiinnijtiamisriski ja sen lisagmisen vaikutukset kuljetta- jien toimintaan (Subjective risk of apprehension from traffic violations and effects of increasing subjective risk on driver behaviour). Research Reports No. 707. Techni- cal Research Centre of Finland, Espoo, Finland.

Makinen, T. and Kulmala, R. (1987) Aikavalien yhteys turvallisuuteen taajamaliikenteessa (Safety and headways in urban traffic). Technical Research Centre of Fin- land, Espoo, Finland.

Maycock, G., Lockwood, C.R. and Lester, J. (1991) The accident liability of car drivers. Report No. 315. Trans- port and Road Research Laboratory, Crowthorne, UK

Naatanen, R. and Summala, H. (1976) Road-user Behavior and TrafJic Accidents. North-Holland/Elsevier, Amsterdam.

Parker, D., Reason, J. T., Manstead, A. S. R. and Stradling,

S. G. (1995) Driving errors, driving violations and accident involvement. Ergonomics 38, 1036-1048.

Rajalin, S. ( 1994) The connection between risky driving and involvement in fatal accidents. Accident Analysis & Prevention 26, 5555562.

Schweitzer, N., Apter, Y., Ben-David, G., Liebermann, D. G. and Parush, A. (1995) A field study on braking responses during driving. II. Minimum driver braking times. Ergonomics 38, 1903-1910.

Summala, H. (1980) How does it change safety margins if overtaking is prohibited: a pilot study. Accident Analysis & Prevention 12, 95-103.

Summala, H. (1987) Ajaminen suomalaisella ja ruotsa- laisella moottoriliikennetiella (Driving on a Finnish and Swedish semi-motorway). University of Helsinki Traffic Research Unit Reports 13, Helsinki.

Summala, H. (1988) Risk control is not risk adjustment: the zero-risk theory of driver behaviour and its implica- tions. Ergonomics 31, 491-506.

Summala, H. (1996) Accident risk and driver behaviour. Safety Science 22, 1033117.

Summala, H. (1997) Hierarchical model of behavioral adaptation and traffic accidents. In TrufJ and Transport Psychology. Theory and Application, eds T. Rothengatter and E. Carbonell. Elsevier, Amsterdam.

Summala, H., Naatanen, R. and Vaisanen, M. (1984) Devi- ant speeds are dangerous: who uses them?. Travail Humain 47, 177-182.

Summala, H. and Vierimaa, J. (1980) Ajoneuvojen nopeu- det, nopeuserot ja aikavalit : analyyseja autonkuljettajien kayttaytymisesta (Driving speeds, speed differences and headways: analyses on driver behavior). Road Admin- istration, Helsinki.

Traffic Insurance Center (1994) TrafJic Accidents Statistics of Insurance Companies 1993. Traffic Insurance Center, Helsinki.

van Winsum, W. and Heino, A. ( 1996) Choice of time headway in car-following and the role of time-to-collision information in braking. Ergonomics 39, 5799.592.

van Winsum, W. and Brouwer, W. (1997) Time-headway in car-following and operational performance during unexpected braking. Perceptual and Motor Skills 84, 1247-1257.

Zuckerman, M., Eysenck, S. and Eysenck, H. J. (1979) Sensation-seeking in England and America: cross-cul- tural, age and sex comparisons. Consulting and Clinical Psychology 46, 139- 149.

Zuckerman, M. and Neeb, M. (1980) Demographic influ- ences in sensation seeking and expression of sensation seeking in religion, smoking and driving habits. Personal- ity and Individual Differences 1, 197-206.

close-following drivers on two-lane highways

Documents