The First Cross Study of Car-Bike Collisions

home              safety

In the early 1970s, California was implementing a policy of forcing cyclists off the roadways onto bikeways, using the excuse that this would make cycling much safer, but actually, as events proved, with the intent of making motoring more convenient, regardless of the danger to cyclists. I was fighting this in the California Statewide Bicycle Committee, which was supposed to recommend the appropriate restrictive legislation to the Legislature. Crucial to this policy is the superstition that the greatest danger to cyclists is same-direction motor traffic. Experienced cyclists such as myself knew that the major collision hazards came from conditions ahead of the cyclist, just as they do for motorists, but we had no scientific data on this point. California contracted with Ken Cross to make a statistical study of car-bike collisions, in the expectation that this study would demonstrate the truth of the superstition that the greatest hazard to cyclists came from same-direction motor traffic. Ken's study was presented to the California Statewide Bicycle Committee at a meeting room in the Sacramento Airport. After the presentation, I rather naively pointed out that the Cross study supported all that I had been saying and utterly disproved the supposed basis for California's policy. As a result of what I said, the study was suppressed, so that the only copies available were those that had been already handed out to those present. The following is a copy of the first Cross Study converted into .htm format. JF]



Kenneth D. Cross Anacapa Sciences, Inc. Santa Barbara, California

[Presented to the California Statewide Bicycle Committee (established by Joint Resolution of the California Legislature), Sacramento, California, 19 June, 1974. JF]

Abstract: This paper describes the first part of a study designed to identify the causes of critical behavioral acts that lead to collisions between bicycles and motor vehicles. Traffic accident reports for 384 accident cases were studied in detail by two analysts. For each accident case, the analysts made independent judgments about the behavioral act that Zed immediately and directly to the accident; that is, the "critical behavior." Accident cases were then categorized into accident types based upon commonality of the critical behavior and the traffic context in which the accident occurred. An analysis was performed to assess the relative frequency of each accident type (and sub-type), and to determine the bicyclist's age and sex distribution for both the total accident sample and for each accident type.

It was found that 91.4 percent of the sample could be categorized into one of ten generic accident types and that each accident type could be subdivided into meaningful sub-types. The accident types and sub-types that were identified are illustrated schematically and discussed in the main body of the paper. The analysis revealed that accident type interacts strongly with the age but not the sex of the bicyclists. That is, the age distribution of bicyclists was found to differ widely from one accident type to another. The implications of the findings for bicycle safety education are discussed.


The principal objective of this study is to develop a cost-effective bicycle-safety education program for school-age children that will serve to reduce the incidence of accidents between bicycles and motor vehicles. The approach developed to meet this objective is based upon three important considerations. First, the education program must be based upon sound empirical data about the causes of bicycle-motor vehicle accidents (2)

Many experts have voiced opinions about accident causes that have been neither supported nor refuted by empirical data. Although these opinions may be perfectly valid, we believe that educational resources are far too limited to warrant the inclusion of training material dealing with causal factors whose relationship with accidents is questionable. Second, if a bicycle safety education program is to be actually implemented, it must be developed in full recognition of the constraints on time and resources available for bicycle-safety education: We feel it would be a simple matter to develop a program that would result in a large reduction in accident frequency if training time and resources were unlimited. However, since resources are always limited, it is necessary to develop a program that will optimize the payoff (in terms of accident reduction) within the relatively fixed limits of available training time and resources. Finally, we feel that a feasible program can be developed and implemented only by taking full advantage of the knowledge and experience of other persons who have been involved in bicycle-safety training. It is extremely costly to develop and validate training methods and the associated materials that are required. Therefore, although the program is to be tailored to the specific problems of Santa Barbara, every attempt should be made to minimize the initial developmental costs by utilizing existing methods, materials, and aids.

A secondary objective of the study is to develop data that will be useful in creating a law enforcement program that is cost effective and compatible with the safety education program. Like educators, law enforcement agencies have limited time and resources that must be allocated to many competing demands. The enforcement of bicycle laws has heretofore been a low priority item among law enforcement officers, partly because bicycle accidents have not been considered an important problem and partly because ticketing bicyclists--especially young bicyclists--is not considered a pleasant job. Law enforcement representatives believed that two types of data are needed to develop an effective program. First, data are needed to sensitize the patrolman to the magnitude of the accident problem. Second, data are needed to identify the high-risk bicyclist and to identify the violations that most often lead to accidents. The latter information enables officers to allocate a relatively greater proportion of their time to the "problem cyclist" and the "problem violation"; that is, to enforce selectively.


The development of an accident typology resulted from several abortive attempts to define specific training and enforcement objectives in other ways. A study of the literature revealed a large number of operator factors, environmental factors, and vehicular factors that have been shown (or assumed) to bear a causal or contributory relationship with the incidence of accidents. However, the description of the causal factors are usually so general that it is quite impossible to use them to define specific countermeasures. For example, it has often been stated that many accidents occur because bicyclists violate traffic laws. An appropriate countermeasure for this type of accident cannot be recommended unless one knows why bicyclists violate the law. If violations are due to ignorance of the law, an education program may prove to be the best countermeasure. . If bicyclists know the law but choose to disregard it, enforcement rather than education might be necessary. A similar problem arises with nearly every explanation of accident cause found in the literature.

The above problem is complicated by the lack of information about the relative importance of the various causal factors. Without such information it is quite impossible to allocate resources in a cost effective way.

A third problem concerns the definition of target groups. The literature identifies male bicyclists between 10 and 14 years of age as being the high-risk bicyclist. That is, this general group of bicyclists has a disproportionately large number of accidents in comparison to the population at large. Does this finding indicate that the involvement of young male bicyclists is the same for all types of accidents? We doubt it! Based upon preliminary findings from previous research, we suspect that the "target group" differs from one accident type to another.

In summary, the development of an accident typology resulted from: a need for more detailed definitions of accident cause, a need for data with which to evaluate the relative importance of various accident causes, and a need for data with which to identify the target group for each type of accident.


The cause of an accident can legitimately be defined in many different ways. The particular definition that is "best" depends entirely upon the purpose for which cause is being defined. Our purpose for defining accident cause is to create a logical framework for identifying potential accident countermeasures and for evaluating their relative cost effectiveness. Our approach to defining accident causation is based upon the fundamental premise that many factors contribute to an accident and that these factors are related in a hierarchical fashion. (By a hierarchical relationship, we mean that proximal causes can be explained in terms of secondary causes, secondary causes can be explained in terms of tertiary causes, and so on.) It follows that no single data item can be expected to provide a satisfactory indication of accident cause, and that accident cause can be defined at several different levels of specificity: The implication of this rationale is that accident cause must be defined in terms of a set of data items, and that the data items used to define cause will depend upon the level of specificity at which the accident analysis is focused.

The concept of causation adopted for this study is illustrated schematically in Figure 1. This conceptualization is based upon the premise that, in the final analysis, all accidents occur because operators behave in an inappropriate manner while in a particular traffic context. Thus, the most proximal cause of accidents is a "specific behavioral act" performed within a specific traffic context. An understanding of the environmental context is necessary to define the behavioral act, but is not considered a direct causal factor.

Figure 1. Concept of accident causation.

The "proximal cause" has descriptive value but has no explanatory value. In order to develop countermeasures, one must explain why the operator chose to behave in a way that led immediately and directly to the accident. The proximal cause (behavior) must be explained in terms of one or a combination of three secondary causes: operator knowledge, operator skill, and operator attitudes and values. Although other factors may influence behavior, the influence is indirect, being mediated by the factors referred to here as "secondary or mediating causes." Hence, the factors identified as "accident correlates" may or may not influence behavior, depending upon the particular knowledge, skills, and attitudes of the operator.

Accidents can be curtailed through behavior modification or by forbidding the operator access to the specific traffic context where the accidents occur. Behavior can be modified by: modifying the knowledge, skills, and attitudes of the operator; by modifying certain environmental factors or vehicle factors; or by modifying any combination of the factors. However, since the influence of environmental and vehicle factors are mediated by the operator's skill, knowledge, and attitudes, it is possible to modify environmental and/or vehicular factors and yet not achieve the desired behavioral modification.


This section describes only that portion of the method that was concerned with defining basic accident types and sub-types. Our objective was to determine whether meaningful accident types could be defined from a study of traffic accident reports alone. The next step in the procedure-which is not discussed here--is to gather data with which to define the factors which caused or contributed to the various accident types.


'The initial sample included traffic accident reports for all accidents involving a bicycle and motor vehicle that occurred in Southern Santa Barbara County between January 1971 and December 1973. Late reports and reports on hit-run accidents were eliminated from the sample because the information was incomplete or of questionable validity. The remaining sample consisted of 384 accidents; 247 accidents involved a male bicyclist and '(37 involved a female bicyclist. The sex of the motorists was not tabulated for this analysis. 


Each accident report was studied in detail by two evaluators. A data card was prepared for each accident that contained: a brief narrative of the accident, a reproduction of the accident diagram, and the age and sex of the bicyclist. The narrative description of the accident described the traffic context (type of street, location of crash relative to an intersection or driveway, traffic control devices if relevant, illumination, known contributory vehicle factors, and visual obstructions if any) and the overt movements or behavior of both vehicle operators.

The first evaluator proceeded through several sorting iterations in an attempt to categorize the accidents into groups according to the pattern of factors and events that led to the accident. The final sorting resulted in 91.4 percent of the accidents being sorted into one of ten major accident types. Brief narrative descriptions of each accident type were developed along with corresponding accident diagrams. The cards, narrative descriptions, and diagrams were given to a second evaluator who was asked to sort the cards into one of the ten categories or into an "other" category: The second evaluator sorted only 22 of the accidents into . categories different from the first evaluator. All but four of these differences were due to an error or a misinterpretation. The classification of the first evaluator was used in the analysis described below.



Figure 2. Cumulative distribution of accidents as a function of cyclist's age and sex.

Figure 2 shows the cumulative distribution of accidents for the total sample as a function of bicyclists' age and sex. The frequencies for males and females were converted to percentages in order to make it easier to compare the distribution of the two curves. Hence, the ordinate represents cumulative percentage rather than cumulative frequency.

The most important finding illustrated by these curves is that the age. range for female participants is smaller than the age range for males. Female involvement in accidents begins at about six years of age, and about 90 percent of the female participants are 25 years of age or younger. In contrast, male involvement in accidents begins at an age of about four years, two years younger than for females. Furthermore, the 90th percentile of the male distribution is not reached until an age of about 34 years of age, nearly nine years older than the 90th percentile for females.

Figure 3. Relative proportion of accidents as a function of cyclist's age and sex.

Figure 3 shows the relative proportion of accidents as a function of the bicyclist's age and sex. This figure provides a somewhat more dramatic illustration of the difference between the age distribution for males and females. The modal value for males occurs at an age of about 13 years, whereas, the modal value for females occurs at an age of about 20 years. We do not know whether the secondary peaks have any meaning or merely reflect sampling error. Nevertheless, this figure leaves little doubt that the "target group" for males is quite different from that for  females. For males, the problem age appears to be between 10 and 15 years of age; the problem age for females appears to be between about 12 and 22 years of age.

TABLE A, From Pie Chart Figure 4. Relative Contribution of Each Accident Type

Type of Accident Percent Critical Maneuver
A: Cyclist Exited Driveway Into Motorist's Path 8.59 Cyclist
B: Motorist Exited Driveway Into Cyclist's Path 5.73 Motorist
C: Cyclist Failed to Stop/Yield at Controlled Intersection 8.33 Cyclist
D: Cyclist Made Improper Left Turn 11.20 Cyclist
E: Cyclist Rode on Wrong Side of Street 14.32 Cyclist
F: Motorist Collided With Rear of Cyclist 4.17 Motorist
G: Motorist Failed to Stop/Yield at Controlled Intersection 7.81 Motorist
H: Motorist Made Improper Left Turn 12.76 Motorist
I: Motorist Made Improper Right Turn 11.20 Motorist
J: Motorist Opened Car Door into Cyclist's Path 7.29 Motorist
Other 8.60

Figure 4. Breakdown of total accident sample by accident type. See Table A

Generic titles for the 10 accident types identified are listed in the left-hand portion of Figure 4. The pie-shaped figure shows the  relative proportion of the total accident sample that was categorized into each type. It can be seen that 91.4 percent of the total sample was categorized into one of the ten accident types. The remaining 8.6 percent did not correspond with any of the major types and was therefore placed in the "other" category. The proportion of the total accident sample accounted for by a given accident type varied from 4.17 percent (motorist collided with rear of cyclist) to 14.32 percent (cyclist rode on wrong side of street). The key at the bottom of the page indicates, for each accident type, whether the critical maneuver was executed by the cyclist or the motorist.


Figures 5 through 13 are stylized, plan-view diagrams that were prepared to more clearly illustrate the circumstances that characterized each of the accident types. (A diagram was not prepared for the accident in which the motorist opened the car door into the bicyclist's path since the dynamics of this situation are easily visualized.) The figures are stylized in the sense that they are not drawn to exact scale and they do not accurately depict which vehicle struck the other. Otherwise, the major factors associated with the accidents are accurately depicted. All of the figures show variations or sub-types of the general accident type. {These figures are similar to those shown in the second Cross study, with diagrams of each subtype designated with its proportion of the type. JF]

Cyclist Exited Driveway into Motorist's Path

All of the accidents shown in Figure 5 occurred when a bicyclist exited a driveway into the path of an oncoming motorist. It was found that accidents of this type could be classified into three distinct subtypes. In about one-third of the cases, the traffic report indicated that a parked vehicle (as shown on the left in Figure 5) obstructed the vision of one or both of the operators. The distinguishing characteristics of another sub-type--which also represented about one-third of the cases--was that the bicyclist traveled across at least one full lane of traffic before colliding with the motor vehicle. The final sub-type occurred in the near lane and involved no visual obstruction whatsoever.

Information contained on the accident report indicated that a failure by the bicyclist to search for oncoming traffic was a factor in nearly all of this type accident, even in the cases where a parked vehicle obstructed vision. In the preponderance of cases, the accident occurred at the intersection of a driveway to a private residence; but in some cases the driveway was an exit from a business parking lot.


Figure 5. Cyclist exited driveway Into motorist's path.

Motorist exited driveway into cyclist's path

The accidents depicted in Figure 6 occurred as a motorist was exiting a private driveway. The motorist was backing out of the driveway in most, but not all cases. It was found that this type of accident occurred under two distinctly different sets of circumstances. In one case, the cyclist was riding illegally on the sidewalk. This sub-type accounted for 73 percent of this accident type. (A city ordinance prohibits sidewalk riding.) A visual obstruction was reported in a few cases but was not a factor in enough of the accidents to warrant a further subdivision. Bicyclist speed and motorist search were the contributory factors most often mentioned on the accident report.
The second type of accident involved a bicyclist riding legally in the street. This sub-type accounted for 27 percent of this accident type. Speed of the bicyclist was reportedly a factor in a few cases, but the. cause of this accident remains quite obscure. One can speculate that motorist search or bicyclist visibility are contributory-factors, but this information was not contained on the report form. Since these two sub-types would appear to involve quite different causal factors, they should probably be separated into major types. That is, the countermeasure for one type would involve removing the bicycle from the sidewalk, whereas, the countermeasure for the other sub-type would involve increasing the motorist's alertness, increasing the ability of the bicyclist to detect a hazard, reducing bicyclist speed, etc.

Figure 6. Motorist exited driveway Into cyclist's path.

Cyclist Failed to Stop/Yield at Controlled Intersection

The accidents in Figure 7 resulted from the bicyclist's failure to stop or yield at a controlled intersection. The diagram portrays a stop sign but a traffic signal was involved in nearly half the cases.

The dashed bicycle symbol in Figure 7 is used to depict a stop-and-go maneuver by the bicyclist. That is, the bicyclist made a legal stop at the intersection and then proceeded into the path of the motor vehicle approaching from the left. This sub-type was defined only to illustrate the infrequency with which this type of accident occurs when the bicyclist makes a legal stop. In contrast, 97 percent of this type of accident occurred as a result of the bicyclist's failure to stop.

The accident report revealed little information about why the bicyclist-failed to stop. A high rate of speed, attributable to riding down a hill, was an interesting factor mentioned in a few cases. Ignorance of the law is an unlikely factor since a previous study revealed few bicyclists who did not know they were supposed to obey all traffic signals.

Figure 7. Cyclist failed to stop/yield at controlled intersection.

Cyclist Made Improper Left Turn

Figure 8 shows accidents that occurred when the bicyclist was making a left-hand turn maneuver. It was found that this type of accident could be broken into four sub-types. Three sub-types--which occurred with about equal frequency--occurred in close proximity to the intersection of two streets or at the intersection of a street and a driveway. In two of these cases, the bicycle collided with a motor vehicle approaching from the front. Inattention, failure of the bicyclist to scan, and misjudgment of the motorist's speed are possible causal factors for these two subtypes, but the accident report contained no useful information to verify this assertion.

Figure 8. Cyclist made Improper left turn.

In two sub-types, the bicyclist collided with a motor vehicle approaching from the rear. In all of these cases, the bicyclist was making an overt left-hand turning movement rather than swerving. Hence,, lack of motor control cannot be held accountable for this type of accident. The most mysterious and frequently occurring accident sub-type (58 percent) happened mid-block and involved neither an intersection nor a parked vehicle. We do not know the bicyclist's motive for turning or the reason why they failed to detect the approaching motor vehicle. A possible cause for this type of accident was revealed during an interview with a female bicyclist who had been involved in an accident of this type. She stated that she was preparing to make a left-hand turn in order to ride on the other side of the street (facing traffic). Before turning, she looked behind her and detected an oncoming motor vehicle. She stated that she waited until the vehicle had passed and then initiated a turning movement only to be struck by a second car that was obscured from view by the first. This is an example of a type of accident that warrants detailed investigation. That is, the accident occurs often enough to constitute a problem, and potential countermeasures are not apparent from the data available on the accident report.

Cyclist Rode on Wrong Side of Street

Figure 9 shows that accidents occur in every conceivable way when the bicyclist is riding on the wrong side of the street. The cause of this type of accident is well known. That is, the motorist simply does not expect a hazard to be approaching from that portion of the environment. In examining the intersection accidents, it is interesting to note that a much smaller proportion of accidents occur when a full lane of traffic separates the bicycle and the motor vehicle. Apparently, the lane separation provides sufficient additional response time for the motorist and/or bicyclist to execute a successful evasive maneuver.

Although this type of accident occurs frequently, a detailed analysis to identify the cause does not appear necessary. Riding against traffic is clearly unlawful, and most bicyclists are aware of this fact. Hence, law enforcement and attitude modification appear to be obvious countermeasures for this type of accident.

Figure 9. Cyclist rode on wrong side of street.

Motorist Collided with Rear of Cyclist

Accidents in which the motorist collided with the rear of the bicyclist occur relatively infrequently. The five sub-types shown in Figure 10 constitute only 4.1 percent of the total accident sample. This is surprising since these types of accidents appear most hazardous to the bicyclist. Most bicyclists would predict that many accidents occur when a bicyclist is riding along a heavily trafficked street-with a line of parked cars along the right-hand curb. In fact, this type of accident occurred extremely infrequently in our sample. The accident that occurs when the bicyclist is swerving around a single parked car is also assumed by "experts" to be a frequently occurring accident type. In fact, this type of accident has been given a name in the literature, the "mousetrap" accident. Again, this type of accident occurred very infrequently in our accident sample.
It is extremely interesting to find that the type of accidents that are most often referred to in the bicycle-safety education literature as "most hazardous" are, in fact, among the most infrequently occurring accidents revealed by this analysis. We hypothesize that the reason these types of accidents occur infrequently is that the hazard is so apparent to both the motorist and bicyclist that they exercise much more caution than under ordinary circumstances.


Figure 10. Motorist collided with rear of cyclist.

Motorist Failed to Stop/Yield at Controlled Intersection

In Figure 11, it was the motorist who failed to stop or yield at a controlled intersection. Unlike the corresponding accident where the bicyclist failed to stop/yield, it was found that 80 percent of the motorists involved in this type of accident first made a legal stop at the signal and then proceeded into the path of the approaching bicyclist. In only 20 percent of the cases did the motorist fail to stop at a controlled intersection.

Although we do not yet know why this type of accident occurs, it is unlikely that accident frequency would be reduced by a more strict law enforcement program directed at the motorist. Evidence obtained during interviews with motorists involved in this type of accident suggests that the motorist simply did not see the bicyclist. Additional information will be required before we can assess why the motorists fail to see the bicyclist in this type of situation. The search for causal factors should center on such factors as: motorists' search and scan behavior, the visibility of bicyclists, bicyclists' ability to recognize hazards, and so on.


Figure 11. Motorist filled to stop/yield at controlled intersection.

Motorist Made Improper Left Turn

Figure 12 shows accidents in which the motorist made an improper left turn into the path of a legally ridden bicycle. Although three subtypes were identified, the majority of this type of accident falls into one sub-type (86 percent). It can be seen that this accident occurs as the motor vehicle is about to complete his turn and the bicyclist has just entered the intersection.

Little information about the cause of this type of accident could be obtained from the traffic accident reports, but information obtained from participants of this type of accident points to potential causal factors: poor bicycle visibility, an assumption by the bicyclist that he had been seen by the motorist, motorist underestimating bicycle speed, and uncertainty about which vehicle had-the right-of-way. This type of accident will require considerable study before countermeasures can be meaningfully defined.

Figure 12. Motorist made Improper left turn. 

Motorist Made Improper Right Turn

Figure 13 depicts the classical right-turn accident that is so often described in the literature. The only unique finding shown in this figure is the sub-type that occurs at the junction of a curved roadway and a freeway on-ramp. We discovered that every one of this accident sub-type occurred at exactly the same location. In order to remain on the roadway, the bicyclist must bear left and travel across the entrance to the freeway on-ramp. A motorist intending to enter the freeway must simply continue straight ahead. This type of accident apparently is due to confusion about which vehicle has the right-of-way.

We feel that the layout of the roadway is mainly responsible for the operator's confusion about right-of-way, and therefore, that modification of the environment would be the most cost-effective countermeasure.

Figure 13. Motorist made Improper right turn.

Summary Comment

In the next phase of our study, we intend to conduct interviews with motorists and bicyclists in order to clarify the causes of the various accident types and sub-types discussed in the previous paragraphs. In some cases, little additional information will be needed in order to identify effective countermeasures. In other cases, the causal factors are much more obscure and considerable effort will be required to identify them. In still other cases, the accident sub-type occurs so infrequently that it will not be cost effective to attempt to define the factors that cause or contribute to that type of accident.


In developing an education program, it is desirable to delay the education as long as possible because the desired training can usually be more easily achieved with older children. That is, within limits, the same amount of training can be achieved in a fewer number of hours with older than with younger children. On the other hand, training cannot be delayed .too long, for many accidents will have already occurred before the required training is administered.

Another complicating factor is the difficulty in fitting a significant amount of training into. the already busy schedule of the schools. There is certain to be more training required than could possibly be accomplished at any one grade level. Hence, it is desirable to distribute training over several grade levels if this can be done without incurring a large number of accidents before the required training is administered. For this reason, it is extremely important to determine if age interacts with accident type and, if so, at what age should education about each accident type be administered.

Figure 14 shows cyclist age distribution as a function of accident type. The key in the lower right hand corner defines the symbolization. This figure shows that there is indeed an interaction between age and accident type and that the interaction is a very significant one. From this figure it can be seen that: the median age for the accident types varies from about 10 years of age to about 21 years of age; the 25th percentile varies from about 7.5 years to about 17 years; and the 10th percentile varies from about 4.5 years to about 14 years of age.
It is interesting to note that use of the total sample distribution to identify "target groups" or to make decisions about the age at which education should be administered would be a grave error. For example, the median for only one accident type (motorist collided with rear of cyclist) is even close to the median for the total accident sample. The median for all other accident types differs from the total sample median by at least two years, and in some cases the difference is as much as six years. We feel this finding has great significance for the evaluation of nearly all types of potential countermeasures.

Figure 14. Cyclist age distribution as a function of accident type.

Figure 15 shows the cumulative distribution of accidents as a function of age and accident type. This figure shows that the age distribution varies widely from accident type to accident type, and furthermore, shows that the distributions tend to fall into two distinct groups. It will be noted that all of the accidents in which the critical maneuver was performed by the bicyclists are clustered together, whereas, all of the accidents in which the critical maneuver was performed by the motorists are also closely clustered.

Table 1 shows the estimated "pre-training accidents" as a function of the grade at which training is administered. It can be seen that safety training, or any other type of countermeasure, can be delayed until junior high for some accident types without incurring an unreasonable number of accidents. For other types of accidents, a rather significant proportion of the accidents will have occurred before the child reaches kindergarten. The boxes encircled by the heavy-weight line show the grade levels at which training could be administered without incurring _ more than ten percent of the total accidents, assuming that the age distribution remains the same.

Figure 15. Cumulative distribution of accidents as a function of age and accident type.


Grade Level
K 1 2 3 4 5 6 7 8 9 10 11 12
Cyclist Exits Driveway 16 22 29 36 43 52 60 67 74 82 85 88 93
Motorist Exits Driveway 11 16 18 27 33 38 43 47 52 55 58 65 72
Cyclist Runs Stop Sign 1 6 10 15 20 28 34 38 50 65 72 76 80
Cyclist Improper Left Turn 1 3 8 10 15 19 27 38 46 57 63 67 72
Cyclist Wrong Side of Street 1 3 6 9 13 16 18 27 39 52 58 65 71
Motorist Hit Rear of Cyclist 0 0 0 0 0 0 3 10 16 22 26 30 45
Motorist Ran Stop Sign 0 0 0 1 4 5 7 8 9 15 22 26 30
Motorist Improper Left Turn 0 0 0 0 0 1 3 4 6 10 17 23 25
Motorist Improper Right Turn 0 0 1 3 4 4 7 10 13 17 19 23 28
Motorist Opened Door 0 0 0 0 0 0 0 1 6 8 13 17 21


Much work needs to be done before causal factors are identified and countermeasures (in this case educational countermeasures) have been developed. However, we feel that the typology and the associated analysis have focused our attention on the critical behaviors that must be modified and the age groups at which the countermeasures should be addressed.


(1) This project is part of the California Traffic Safety Program and was made possible through the support of the Office of Traffic Safety, State of California, and the National Traffic Safety Administration. The work was performed under the direction of Mr. Clint L. Lefler, Santa Barbara Department of Public Works. The opinions, findings, and conclusions expressed in this publication are those of the authors and not necessarily those of the State of California or the National Highway Traffic Safety Administration.
(2) Hereafter, the term "accident" will refer to accidents between bicycles and motor vehicles unless stated otherwise.


Return to: John Forester's Home Page                         Up: Safety