[Up: Table of Contents]
[Previous: Section III]
[Next: Section V]
|This section and Section V contain a description of selected
findings from a recently completed study of bicycle/motor-vehicle accidents (Cross & Fisher, 1977). The findings of a traditional
analysis of descriptive data are summarized in this section. Section V contains a
description of frequently occurring types of bicycle/motor-vehicle accidents-referred to
as "problem types" -- and a discussion of educational countermeasures for the
various problem types.
The general objectives of the Cross and Fisher study were to compile data on the causes of bicycle/motor-vehicle accidents and to use the data to identify the full range of countermeasure approaches that have potential for reducing the number of accidents of this kind. The project was national in scope and encompassed both urban and rural accidents.
Although the information presented in Sections IV and V will meet the needs of most readers, it may not be complete and detailed enough to meet the needs of readers who are involved in the development of a new bicycle-safety-education program or in the assessment of existing programs. Persons with such responsibilities are advised to obtain and study a copy of the original report.
Data on bicycle/motor-vehicle accidents were collected in four sampling areas in the United States. The sampling areas were selected to provide maximum coverage of the characteristics of the bicycling population and the environmental conditions in which they ride. The sampling areas, each consisting of several contiguous counties, were located in California (Los Angeles area), Colorado (Denver/Boulder areas), Florida (Tampa/Orlando areas), and Michigan (Detroit/Flint areas). Within each sampling area, a proportionate sample of non-fatal cases was selected from those occurring during each month of calendar year 1975; an attempt was made to select equal numbers of urban and rural accidents at each sampling area. A non-fatal case was rejected from the sample if it was an unwitnessed hit-run accident or if both of the involved operators refused to be interviewed. Because of the small number of fatal accidents that occurred within each sampling area, none were rejected from the sample. Data were compiled on 166 fatal accidents and 753 non-fatal accidents--919 cases in all.
A conceptual model of the accident-generation process was used in defining the data requirements for this study. This model focused on the sequence of functions and events preceding the accident and the factors that influenced the function-event sequence. Data on each accident case in the sample were compiled by trained Field Investigators. Field Investigators compiled and recorded data from several sources, including: the official traffic accident report, observations and measurements taken at the accident site, and detailed interviews with the vehicle operators and persons who witnessed the accident. A structured questionnaire and a detailed scale-drawing of the accident site were used to conduct the operator interviews.
Some questionnaire items were designed to provide information about the characteristics of the operator, his vehicle, and his trip. However, most items were designed to provide detailed information about the accident-generation process. The interview procedures and instruments were designed to provide a clear notion of the pre-crash path of each vehicle, the function failure of each operator, and the combination of factors that were causally related to the function failures.
A classification system was developed and the accident cases were classified into mutually exclusive "problem types." Cases classified into the same problem type exhibited commonality in the following attributes: the traffic context in which the accident occurred, the operators' function failures, and the combination of factors causally related to the function failures.
All data items were analyzed by problem type. In addition, selected descriptive-data items were analyzed for the fatal and non-fatal samples -- pooled over problem types. The characteristics of individual problem types and the results of the descriptive-data analyses were examined systematically in an attempt to identify general countermeasure approaches having the potential for reducing the incidence of bicycle/motor-vehicle accidents.
The vehicle operators in the study sample -- both bicyclists and motorists -- were predominantly males. Furthermore, the proportion of males was greater for the fatal sample than for the non-fatal sample. Seventy-one percent of the non-fatal accidents and 85% of the fatal accidents involved a male bicyclist; a male motorist was involved in 65% of the non-fatal and 72% of the fatal accidents. It is probable that the overrepresentation of males is due partly to a greater amount of exposure for males -- particularly male bicyclists. However, it is also probable that there are some important behavioral differences between male and female bicyclists.
The overrepresentation of male bicyclists may suggest that a bicycle-safety-education program should concentrate principally or exclusively on males. Although it may be true that our educational dollars might be most cost-effectively spent in educating males to the exclusion of females, such an approach would be unfair and shortsighted. The absolute number of accidents involving females is far too large to warrant their exclusion from an educational program. Moreover, because an increasing number of females are becoming interested in bicycling, it can be expected that the differences in male and female involvement in accidents will diminish in the future.
The age distribution of the motorists in the study sample was found to be highly similar to the age distribution of motor-vehicle operators involved in all other types of traffic accidents. Since the age distribution of accident-involved motorists is well known (see National Safety Council's Accident Facts, 1977), the following discussion will be limited to the age distribution of accident-involved bicyclists.
Bicyclist age distributions for fatal and
non-fatal accident cases in the study sample.
|The age distributions of the fatally-injured and non-fatally injured bicyclists in the
study sample are shown in Figure 4. (It should be noted that accident frequency is plotted
for two-year age intervals.) Beginning at age four, accident frequency rises steadily to
the age of 12 and remains at this high level through the age of 15. Thereafter, accident
frequency declines dramatically and remains at a relatively low and constant level for
ages beyond 30 years. The general shape of the curves for fatal and nonfatal accidents is
similar, but fatal accidents are more frequent among the very young and the very old
bicyclists. About 4.5% of the fatal cases involved a bicyclist younger than six years of
age, whereas only two percent of the non-fatal cases involved a bicyclist younger than six
years. Similarly, it can be seen that 18.2% of the fatal cases involved a bicyclist older
than 35 years of age, and only 4.2% of the non-fatal cases involved a bicyclist older than
35 years. Although not shown in Figure 4, over 10% of the fatalities involved a bicyclist
older than 55 years and three percent involved a bicyclist older than 75 years of age. It
is of interest to note that the age distributions shown in Figure 4 are quite similar to
the age distributions found in a number of other studies of bicycle/motor-vehicle
accidents, including studies by: the American Automobile
Association (1973), the California Highway Patrol (1974),
the Virginia Department of Highways (1974), Walsh and Watt (1974), and the Washington
State Patrol (1973).
The age distribution of bicyclists in an accident sample is most meaningfully evaluated in terms of the relative exposure for each age group. Although exposure data are not available that take into account the combined frequency and amount of bicycle usage for each age group, Barton-Aschman Associates conducted statewide household surveys to assess the relative proportion of persons within each age group who rode a bicycle at least once during the year preceding the interview. Separate surveys were conducted for the State of Tennessee (Barton-Aschman, 1974b) and the State of Pennsylvania (Barton-Aschman, 1975). The age distributions revealed by these surveys are shown in Table 9 along with corresponding age distributions for the fatally injured and non-fatally injured bicyclists in the study sample
COMPARISON OF AGE DISTRIBUTIONS FOR
ACCIDENT SAMPLE AND THE GENERAL BICYCLING POPULATION
|ACCIDENT SAMPLE||BICYCLE USERS1|
|BICY- CLIST AGE||FATAL (N=166)||NON- FATAL (N=753)||TENNES- SEE (N=3141)||PENNSYL- VANIA (N=6372)||COM- BINED2 (N=9513)|
|1User data from household surveys completed by Barton-Aschman Associates,
Inc., for the .Tennessee Departments of Conservation and Transportation (Barton-Aschman, 1974b) and the Pennsylvania Department of
Transportation (Barton-Aschman, 1975).
*Proportion differs significantly from the proportion of users (combined Tennessee and Pennsylvania samples) in the corresponding age group (p <.05).
|An analysis was performed to determine whether the age distribution for either the
fatal or non-fatal sample differed significantly from the age distribution of the user
population -- as measured by the combined sample for Tennessee and Pennsylvania (see
column five of Table 9). In columns one (FATAL) and two (NON-FATAL), asterisks were placed
beside the percentage values that differed significantly from the corresponding percentage
value in column five (user population).
An examination of the data for the fatal sample shows that bicyclists younger than 30 years of age and those between 45 and.59 years of age are involved in fatal accidents in about the same proportion as their numbers in the user population. Bicyclists between 30 and 44 years of age are involved in fatal accidents significantly less often than would be expected from their numbers in the user population; bicyclists 60 years of age or older are involved in fatal accidents significantly more often than would be expected from the proportion of persons in this age group who ride bicycles. Stated differently, these data suggest that the likelihood of being killed in a bicycle/motor-vehicle accident is less than average for bicyclists in the 30-44 age group and greater than average for bicyclists who are 60 years old or older.
Examine next the age distribution for the non-fatal sample. It can be seen that bicyclists between six and 15 years of age are involved in non-fatal accidents more often than would be expected from their numbers; bicyclists younger than six years of age and those between 20 and 59 years of age are involved less often than would be expected from their numbers in the user population. It is of particular importance to note that:
The reader must exercise caution when using these data to define the educational target group. The finding that both accident frequency and accident rate is highest among bicyclists between the ages of 12 and 15 may suggest that safety education should be aimed at this age group. However, if safety education did not commence until the age of 12, about one-fourth of all fatal accidents and one-third of all non-fatal accidents will have occurred before the bicyclists receive the education.
It was found that most bicyclists and motorists were experienced vehicle operators who operated their vehicles regularly. In addition, most operators were driving/riding a vehicle they were thoroughly familiar with at the time the accident occurred. About 95% of the motorists and bicyclists had more than one year's driving experience and routinely operated their vehicles two or more hours each week. Seventy-five percent of the bicyclists and 93% of the motorists reported that they had driven the accident vehicle at least 50 times before the accident occurred; only seven percent of the bicyclists and three percent of the motorists had driven their vehicle fewer than five times before the accident.
No data have been located that indicate the amount of driving/riding experience that is required to acquire and maintain a reasonable level of vehicle-handling skill. However, it seems reasonable to assume that a relatively high level of vehicle-handling skill can be acquired by most persons in about one year and that this skill can be maintained by operating a vehicle for one or two hours each week. If these assumptions are valid, it can be concluded that few motorists and bicyclists in the non-fatal study sample lacked basic vehicle-handling skill at the time of the accident. In short, these data fail to support the assumption that a large proportion of bicycle/motor-vehicle accidents result from a lack of basic vehicle-handling skill.
With the exception of intoxication, few operators reported that they were suffering from any type of impairment at the time of the accident. It was found that less than one percent of the bicyclists were impaired by alcohol. However, evidence that the motorist had been drinking was found in 3.5% of the non-fatal accidents and 16.9% of the fatal accidents. Alcohol was judged contributory in nearly every case in which it was found present. Evidence of drug use was found only infrequently, but the type of data collected during this study cannot be expected to provide reliable information about the number of operators who were under the influence of drugs when the accident occurred.
BICYCLISTS' KNOWLEDGE OF THE LAW
For all accidents that resulted from the bicyclist's violation of a traffic law, the bicyclist was questioned in detail about his reasons for violating the law. It was found that the violation was due to ignorance of the law in only one case.
OTHER OPERATOR CHARACTERISTICS
Listed below are other items of information obtained from the interviews with operators in the non-fatal sample. The percentages reported are based on 525 bicyclist interviews and 385 motorist interviews.
The type of motor vehicle involved in the accident was usually recorded on the official traffic accident report form, but the specific type of bicycle was seldom reported. For this reason, information about motor-vehicle type was. obtained for nearly every case in both the fatal and non-fatal samples; information on bicycle type was obtained only for the non-fatal cases in which the bicyclist was interviewed.
The relative frequency with which different types of bicycles were ridden by male and female bicyclists in the non-fatal sample is shown in Table 10. Also shown is the distribution of bicycle types for the combined (male and female) sample. Considering the combined sample, it can be seen that most bicyclists were riding a lightweight bicycle at the time the accident occurred and that a smaller, but significant, number were riding a standard or middleweight bicycle. About five percent of the bicyclists were riding a highrise bicycle; less than two percent were riding another type of bicycle (child tricycle or big wheel, adult tricycle, folding or collapsible bicycle, tandem bicycle, or custom design).
TYPE OF BICYCLE RIDDEN BY MALE AND
FEMALE BICYCLISTS IN THE NON-FATAL SAMPLE
|A comparison of the distributions of bicycle type for males and females shows that nearly identical percentages of males and females (about 50%) were riding a lightweight bicycle. A standard or middleweight bicycle was ridden by a slightly larger percentage of females (46.5%) than males (40.50, whereas a slightly larger percentage of males than females were riding a highrise or "other" type bicycle. Statistical tests revealed that none of the differences between corresponding percentage values were statistically significant (p <.05). Therefore, these data suggest that there are no important differences in the types of bicycles ridden by male and female accident victims.|
DISTRIBUTION OF BICYCLE TYPES FOR THE
STUDY SAMPLE (NON-FATAL CASES) AND
A RECENT HOUSEHOLD SURVEY
|BICYCLE TYPE||STUDY SAMPLE (N=524)||HOUSE-
HOLD SURVEY1 (N=3187)
|There have been few survey studies that attempted to assess the relative number of
bicycles of each type that are in use by the general bicycling population. Most surveys
that have addressed the issue of bicycle type are limited to only one segment of the
population (school-age children, college students, etc.) or are outdated. One recent study
has been located that surveyed the general population in Santa Clara County, California (Diridon Research Corporation, 1973). The distribution of
bicycle types revealed by this survey is shown in Table 11 along with the distribution of
bicycle types for the study sample. It can be seen that lightweight bicycles are
overrepresented in the accident sample, and that all other bicycle types are
underrepresented. Although no data are available on the distribution of bicycle types in
use within the areas from which the accident sample was drawn, it is unlikely that the
number of lightweight bicycles in use within the sampling areas would be greater than the
lightweights in use within Santa Clara County, California, where the adult ridership is
very high. For this reason, the data shown in Table 11 suggest that a disproportionate
number of bicycle/motor-vehicle accidents involve lightweight bicycles. Although
it is possible that accident rate would be constant across bicycle types if exposure
(type, frequency, and amount of riding) was held constant, it is also possible that
accident rate is higher for lightweight bicycles because the average speed is far greater
than for other types of bicycles.
The distributions of motor-vehicle type for the fatal and non-fatal samples are shown in Table 12. The parenthetical values adjacent to the name of the vehicle type represent the percentage of total vehicle registrations for the- associated vehicle type (National Safety Council, 1976). For instance, 77.5% of all vehicles registered in the United States are passenger cars, 18.4% are trucks, and so on.
TYPE OF MOTOR VEHICLE DRIVEN BY MOTORISTS
IN THE FATAL AND NON-FATAL SAMPLES
|PASSENGER CAR (77.5%)||126||79.8||658||88.1|
|Pickup or Van||24||15.2||61||8.2|
|1Parenthetical values show percentage of total vehicle registrations for the associated vehicle type.|
|As would be expected, most of the motor vehicles involved in bicycle/motor-vehicle
accidents are passenger cars. It can be seen that about 80% of the fatal accidents and 88% of the non-fatal accidents involved a passenger car (a significantly
larger percentage of non-fatal than fatal accidents involved a passenger car
[p < .01]). Comparison of the distribution for the study sample with the
distribution of all registered motor vehicles shows that passenger cars are only slightly
overrepresented in the fatal sample but are overrepresented in the non-fatal sample by
more than ten percent. Although the reasons for this overrepresentation of passenger cars
in bicycle/motor-vehicle accidents is not known for certain, the most probable reason is
that passenger cars are more often driven in the areas where bicycle density is greatest.
Table 12 shows that trucks are involved in a proportionately greater number of fatal accidents (19%) than non-fatal accidents (9.4%). More than 80% of the trucks were pickups or vans; the remainder were larger types of trucks. These data suggest that the likelihood of fatal injuries increases as a function of the size of the vehicle. For instance, dividing the proportion of fatal cases by the proportion of non-fatal cases yields a ratio of .9 for passenger cars, 1.9 for pickups and vans, and 3.2 for larger types of trucks. However, because of the small number of cases involving a truck, these data can only be considered suggestive.
Only one fatality resulted from a collision between a bicycle and a motorcycle. Motorcycles were involved in a proportionately greater number of non-fatal accidents (2.4%). Although motorcycles were involved in bicycle/motor-vehicle accidents less often than would be predicted from their numbers, it is possible that the accident rate per mile driven may be greater than for other types of motor vehicles.
The small number of bicycle/motor-vehicle accidents involving a bus was somewhat surprising. Considering the width of a bus and the types of areas in which they travel, it seems reasonable to expect a greater number of bicycle-bus accidents than was revealed by the sample. This result is probably a function of the skill of the bus drivers and a recognition by bicyclists that buses constitute a serious threat.
The bicyclists who were interviewed were asked to identify both the safety equipment and the vehicle defects for the bicycle they were riding at the time of the accident. To minimize the effects of recall, checklists of safety equipment and defects were provided. The motorists who were interviewed were asked to identify equipment defects for the motor vehicle they were driving at the time the accident occurred. A checklist was also used to assess motor-vehicle defects.
Bicycle Safety Equipment
Bicyclists were asked to identify the safety equipment that was on the bicycle they were riding when the accident occurred and to indicate whether or not the items they checked were in good working order. The bars in Figure 5 indicate the proportion of bicycles in the non-fatal sample that were equipped with the associated safety item. The shaded portion of the bar indicates the proportion of cases in which the item was defective.
Safety equipment on the bicycles in the sample of non-fatal accidents.
|It can be seen that the vast majority of bicycles were not equipped with all the
safety items that most experts consider essential for safe riding and, in some cases, that
are required by law. Only four of the safety-equipment items were found on the majority of
bicycles: handlebar grips or tape (83%), rear reflector (76%), reflectorized pedals (68%),
and chainguard (62%). Although a front reflector and a forward and rear side reflector are
required by law, it can be seen that only about 47% of the bicycles were equipped
with a front reflector and about 38% were equipped with a forward and rear side reflector.
Twenty percent or fewer of the bicycles were equipped with the remaining safety items. It is interesting to note that although about 20% of the bicycles were
equipped with a taillight and headlight, about five percent of all taillights and
headlights were defective or otherwise inoperable at the time the accident occurred. It is
also of interest to note that only seven percent of the bicycles were equipped with a
safety flag and that less than five percent were equipped with a rear-vision mirror (this
percentage includes head-mounted rear-vision mirrors).
It might be argued that although many bicycles are not equipped with the necessary lighting equipment, such ill-equipped bicycles are not often ridden at night. For this reason, the availability of lighting equipment was tabulated separately for daytime and nighttime accidents. This tabulation is shown in Table 13. It can be seen that the proportion of bicycles equipped with the various lighting equipment was similar for the daytime and nighttime accidents. The proportions differed significantly only for reflectorized clothing where it was found that a significantly larger percentage of bicyclists involved in nighttime accidents were wearing reflectorized clothing (p < .05). However, the absolute number of bicyclists who were wearing reflectorized clothing at the time of the accident was so small that this difference has little practical significance.
LIGHTING EQUIPMENT ON BICYCLES INVOLVED IN DAYTIME
AND NIGHTTIME ACCIDENTS (NON-FATAL ACCIDENT SAMPLE)
EQUIPPED WITH ITEM
|REARWARD SIDE REFLECTOR||36.7%||38.5%|
|FORWARD SIDE REFLECTOR||35.4%||40.4%|
|These data would be most meaningful if it were possible to compare the safety
equipment on bicycles in the accident sample with the safety equipment on the general
population of bicycles in the sampling areas. Unfortunately, no data have been located
that enable one to estimate the percentage of bicycles in the general population that are
equipped with the safety items investigated in this study. However, based upon casual
observations, it is believed that bicycles in the accident sample would not differ
significantly from those in the general population.
As is discussed in more detail later, lighting equipment and devices to increase the daytime conspicuity of the bicycle (safety flags, for example) are clearly the most crucial items of safety equipment.
Other items are either present on most bicycles or, if absent, seldom contribute to bicycle/motor-vehicle accidents.
During the interviews, the bicyclists were first asked to indicate on the checklist the equipment that was defective at the time of the accident, and then were asked to indicate whether the defect contributed to the accident in any way. The bars in Figure 6 indicate the proportion of bicyclists who reported the presence of the associated defect. The shaded portion of the bar indicates the proportion of cases in which the defect was present and judged contributory by the bicyclist.
Bicycle defects reported and defects judged contributory
by bicyclists in the non-fatal accident sample.
|Although a significant proportion of the bicycles were defective, few of the defect
were judged contributory by the operator. The one exception to this observation is
defective brakes. Nearly 11% of the bicyclists reported that their brakes were defective
at the time of the accident, and over half of them indicated that their defective brakes
contributed to the accident. The researchers' assessment of the contribution of bicycle
defects did not always correspond with the judgment of the bicyclists. In a significant
number of cases, it was found that the accident was imminent by the time the bicyclist
first attempted to brake; so the defective brakes were judged non-contributory, even
though the bicyclists believed that the brake defect did, in fact, contribute to the
The main implication of these findings is that programs to eliminate bicycle defects, with the possible exception of defective brakes, cannot be expected to make a significant impact on the number of bicycle/motor-vehicle accidents that occur. This conclusion is supported by the findings of a study by the Virginia Department of Highways (1974) in which a bicycle defect was found to be a contributory factor in less than three percent of all bicycle/motor-vehicle accidents.
It was found that nearly all motor vehicles in the sample were properly equipped and free of defects when the accident occurred. This finding corresponds closely with the findings of other studies which indicate that less than one percent of all bicycle/motor vehicle accidents involve a defective motor vehicle (see Waller and Reinfurt, 1969; Washington State Patrol, 1973).
CHARACTERISTICS OF ACCIDENT TRIP
About 80% of the bicyclists and 96% of the motorists were on a utilitarian trip to a specific destination when the accident occurred. Approximately equal numbers of bicyclists were traveling for the following purposes: shopping or errands (22%), commuting to a place of recreation (21%), visiting friends (19%), and commuting to school or work (19%). Although only 18% of the accidents occurred while the bicyclist was on a recreational trip with no destination, household surveys have revealed that between 50% and 60% of all bicycle trips are of this type.
The most common trip purposes for motorists include: shopping or errands (41%), commuting to school or work (29%), visiting friends (14%), and commuting to a place of recreation (13%).
Most operators were on a relatively short trip when the accident occurred. The median one-way trip length was 1.1 miles for bicyclists and 5.8 miles for motorists. Less than five percent of the bicyclists were on a trip exceeding a one-way length of 3.4 miles; less than five percent of the motorists were on a trip that exceeded about 30 miles, one way.
DAY OF WEEK
The accidents in the study sample did not exhibit the weekend rise that is typical for other types of traffic accidents. In fact, the frequency of non-fatal accidents was less on Saturday and Sunday than on any day of the week. For all practical purposes, there is no day of the week that is clearly more or less important than any other day.
Figure 7 shows the distributions of fatal and non-fatal accidents in the study sample by time of day. Also shown (solid circles) is the distribution of all motor-vehicle accidents by time of day (National Safety Council, 1976). It can be seen that the distribution of bicycle/motor-vehicle accidents is similar but somewhat more pronounced than the distribution of all motor-vehicle accidents. That is, there is a minor peak during the morning rush hours between 7:00 and 9:00 AM and a major peak during the evening rush hours between 3:00 and 7:00 PM.
Distributions of fatal and non-fatal accidents by time of day.
|The distributions of fatal and non-fatal accidents differ in two important respects.
First, a relatively smaller proportion of fatal than non-fatal accidents occur during the
evening rush hours. While the absolute number of fatal accidents is greatest during these
hours, the likelihood of a fatal accident apparently does not increase as a simple
function of exposure. Secondly, the relative proportion of fatal accidents occurring after
8:00 PM is almost surely due to darkness. As will be shown later, the types of accidents
that occur during darkness are more likely to result in fatal injuries to the bicyclist.
Nearly identical distributions of accidents as a function of time of day are reported by Waller and Reinfurt (1969), Walsh and Watt (1974), and the Washington State Patrol (1973). All three of these studies show a secondary peak during the morning rush hours and a major peak during the evening rush hours. Furthermore, the reported percentage values are nearly identical to one another and to the percentage values for the non-fatal accidents presented in this study.
About 17% of all accident trips were made during darkness. However, it was found that a significantly greater proportion of fatal (30%) than non-fatal (10%) accidents occurred during darkness. These findings provide strong support for the contention that the likelihood of sustaining fatal injuries from a bicycle/motor-vehicle accident is significantly greater when the accident occurs at night.
In addition to a greater likelihood of fatal injuries at night, it is probable that accident rate is also far higher at night. Although no data have been located that provide an accurate estimate of the amount of all bicycle riding that is done during darkness, casual observation and discussions with a large number of bicyclists indicate that night riding accounts for no more than three or four percent of most bicyclists' total riding time.
Most of the accident trips were made during conditions of fair weather. A small, but significant, number of accidents occurred when rain was falling (three percent of the non-fatal cases and six percent of the fatal cases). Only a fraction of one percent of the cases occurred when it was snowing, during a period of heavy fog, or in an area with blowing sand or dust.
CHARACTERISTICS OF ACCIDENT LOCATION
URBAN VERSUS RURAL ACCIDENTS
Law enforcement agencies most commonly differentiate urban and rural areas in terms of either the incorporation status of the area or the number of inhabitants who reside within a built-up area. As a consequence, many accidents that are officially designated as rural occur in densely populated residential communities located in the unincorporated fringe of a large population center. Similarly, some accidents officially designated as urban occur in areas that are truly rural in character.
To avoid the ambiguity associated with the official designation, the accidents were classified as urban or rural based upon information obtained from the on-site inspections. Accidents usually were classified as rural if they occurred in an area where (a) the posted speed limit was 45 MPH or more, (b) there were no curbs or sidewalks adjacent to the roadway, (c) street lights were not present at the intersections, and (d) at least 50% of the area within one-half mile radius of the accident site was open. Cases that did not meet all four of these classification criteria were classified as urban.
A comparison of the official designations and the designations based upon the on-site inspections revealed the following:
According to the National Safety Council (1976), (a) 604 of the fatal accidents occur in incorporated areas, and 40% occur in unincorporated areas; (b) 80% of the non-fatal accidents occur in incorporated areas, and 20% occur in unincorporated areas. These data are based upon police reports, so are subject to the biases discussed above. Therefore, the above estimates of the magnitude of this bias were used to adjust the National Safety Council's estimates of the distribution of incorporated and unincorporated accidents. The adjusted estimates are shown below:
These data leave no doubt that the likelihood of sustaining fatal injuries is greater for accidents that occur in rural areas. It is also probable that accident rate is higher in rural areas, but it will be necessary to obtain data on the relative amount of riding that is done in urban and rural areas in order to assess the differences in accident rate.
PROXIMITY TO OPERATOR'S RESIDENCE
Most accidents occurred in close proximity to the operator's residence. The median distance between the accident site and the operator's residence was .6 miles for bicyclists and 2.6 miles for motorists. These findings, along with the finding that most operators had driven through the accident site many times before the accident occurred, enable one to confidently conclude that lack of familiarity with the accident site is seldom a factor in bicycle/motor-vehicle accidents.
POSTED SPEED LIMIT
The majority of accidents occurred on roadways with a posted speed limit of 30 MPH or less. However, the likelihood of fatal accidents was found to be positively correlated with the posted speed limit for the roadway on which the accident occurred. The distribution for non-fatal accidents showed that over 806 of the non-fatal accidents occurred on roadways with a posted speed limit of 35 MPH or less. In contrast, more than half of all fatal accidents occurred on roadways with a speed limit greater than 35 MPH; less than one-third of the fatal accidents occurred on roadways with a posted speed limit of 25 MPH or less.
It was found that one or both operators' pre-crash path was on a laterally curved roadway in only 3.6% of the cases. About seven percent of the motorists and ten percent of the bicyclists were traveling on a measurable hill at the time of the crash or shortly before. For motorists, equal numbers were traveling uphill and downhill. However, a significantly larger proportion of the bicyclists were traveling downhill than uphill. This finding undoubtedly is due to the higher speeds bicyclists travel when riding downhill, and indicates that, on the average, accident risk is greater when traveling downhill. Riding downhill at an excessive speed was judged contributory in about six percent of the cases.
About 12% of the accidents occurred on a roadway with one or more significant defects. However, roadway-surface defects were found to be contributory in less than three percent of the cases.