III. RESULTS OF STUDY - ACCIDENT CLASSIFICATIONS

On Tables 3 and 4, the distribution of accidents by classes and types is shown. Table 3 shows the distribution using the original Cross classification scheme, allowing comparison of MAPC data with both the Cross and the Missoula, Montana studies. Table 3 shows the distribution using the modified Cross scheme, based on the Manual Accident Typing (MAT) system created by the National Highway Traffic Safety Administration, which added seven new types to the Cross scheme (see Introduction, sections on Methodology and the MAT).

In using this system, MAPC also removed one prominent accident type -- involving opening doors of parked cars - from the original Cross Type 17 and included it with two other types in a new Class G, "Slowed or Parked Cars." "Other" types were grouped together in Class H. MAPC believes that these revisions improve the classification system. This revised classification is used in the cross-tabulations with other variables in the study.

The classes are reviewed in the order of their frequency of occurrence in the MAPC study. Following the name of the class are four percentages: the MAPC revised MAT classification frequency; the MAPC original Cross system frequency; the Cross non-fatal sample frequency; and the Missoula, Montana frequency.

1) Class F: Motorist Unexpected Turn (MAPC Rev: 17.20, MAPC: 17.6%; Cross NF: 14.5%; Missoula: 20.0%)

This class involved accidents where a motorist who is turning right or left at an intersection (excluding right turns-on-red) collides with a bicyclist approaching from the motorist's front or rear. Only 14 percent of these accidents involved a wrong-way cyclist, compared to the 24 percent of all accidents involving wrong-way cyclists (however, 5 of the 6 accidents included in Type 22, "Motorist Left Turn; Parallel Paths; Same Direction" involved wrong-way cyclists).

Cyclists 15 years of age and over accounted for over 87 percent of these accidents. Those over 18 years of age accounted for over 55 percent of the cases. As with the other classes of accidents, approximately three-quarters of Class F accidents occurred on weekdays. Most often, these occurred during the afternoon peak, between 3-7 p.m. (40 percent). On weekends, these accidents were more likely to occur between 10 a.m. - 3 p.m. (50 percent).

Class F accidents showed a similar distribution in the incidence and type of injury as the sample as a whole.

The most frequent type of accident within this class is the motorist turning left in front of a cyclist coming from the opposite direction (Type 23). This was the most frequent accident type in the study. The next most frequent type within Class F is the motorist turning right in front of a cyclist coming from the same or opposite direction (Type 24 - 6 percent). Least frequent in this class was the accident type involving a motorist turning left in front of a cyclist coming from the same direction (Type 22 1 percent). As pointed out above, however, wrong-way riders accounted for 83 percent of Type 22 accidents.

2) Class C: Motorist Driveout (MAPC Rev: 16.8%; MAPC: 15.7%; Cross NF: 18.7%; Missoula: 23.3%)

This class involves a motorist emerging from an intersection, driveway or alley onto a roadway, and colliding with a bicyclist on that roadway. Right turns on red are included as Type 10. While Cross limited intersection accidents in this class to those where the motorist's approach was controlled by a sign or signal, the MAT added Type 43 to this class. Type 48 accidents involve a collision at an uncontrolled intersection where it is established that the motorist failed to yield to the cyclist.

Wrong-way cyclists were over-represented in this class, relative to the sample as a whole involved in 49 percent of Class C accidents, compared to 24 percent of all accidents.

Class C accidents occurred among a slightly older population than the other classes. Over 76 percent occurred among cyclists over 15 years of age, and 31 percent involved cyclists over 25 years of age.

Class C accidents occurred on weekdays in the same proportion as all accidents. Mid-day weekday accidents are over-represented in this class -36 percent occurred during the hours of 10 a.m. to 3 p.m. compared to 24 percent of all accidents. The afternoon peak period was the next most likely time period to experience these accidents (38 percent, compared to 41 percent of all accidents).

Weekend Class C accidents were most likely to occur during the 10 a.m-3 p.m. period (58 percent for Class C versus 47 percent of all weekend accidents).

Class C accidents were somewhat less likely than other classes to result in fatalities or the most serious injuries, and somewhat more likely to result in no injury at all.

The most common type of accident within Class C was Type 9 -- "Motorist failure to yield at stop or yield sign" -- 9 percent of all accidents. This was the second most common type of accident in the study. Wrong-way cyclists were involved in 53 percent or Type 9 accidents.

3) Class A: Bicycle Rideout From Driveway, Alley and Other Mid-block Locations (MAPC Rev: 16.6%; MAPC: 16.4%; Cross NF: 13.9% Missoula: 8.9%)

This class includes a cyclist emerging from a residential or commercial driveway, alley or sidewalk and colliding with a motor vehicle approaching on the roadway. Only 10 percent of these accidents involved a wrong-way cyclist (compared to the 24 percent of wrong-way cyclists in the sample).

Over 90 percent of Class A accidents involved cyclists 14 years of age and under. This class was by far the most likely to occur to younger cyclists.

Class A accidents most frequently occurred on weekdays (75 percent). Fifty-six percent of Class A weekday accidents took place between 3 p.m. and 7 p.m., the highest proportion of any class to occur within the afternoon peak. On the weekends, these accidents were more likely to occur between 10 a.m. and 3 p.m. (44 percent, similar to the 47 percent share of all weekend accidents occurring during this period).

Class A accidents were among the most likely to result in the most serious category of non-fatal injury ("visible signs of injury, such as bleeding wounds or distorted member; or had to be carried from scene).

4) Class H (Revised): Other Accidents (MAPC Rev: 13.8%; MAPC Class G: 23.6%; Cross NF: 11.2%; Missoula: 15.6%)

Class H involves accident types which do not fit into the other classes. It thus differs from Classes A through G by the lack of commonality among the types included in this class.

As noted in the introduction to this section, Class H has been revised from the original Class G by removing two types which have been placed in the new Class G, "Slowed or Parked Cars" (Type 27, "Bicyclist Overtaking" and Type 35, "Motorist Drive-out From On-Street Parking").

Within this class, the most frequent types of accidents are:

• Type 25: Accident occurs at uncontrolled intersection
• Type 26: Vehicles collide head on, wrong-way cyclist

Type 25 accidents include those which occur at uncontrolled intersections, and where failure to yield is not apparent from the accident report. In the Cross study, all accidents occurring at uncontrolled intersections were included in this type (even where "fault" was assignable), and the MAPC share (7 percent) of accidents of this type using this definition was much greater than in the Cross or Missoula studies. Undoubtedly, this resulted from the larger proportion of accidents at uncontrolled intersections in the MAPC study (36 percent of all intersection accidents versus 10 percent in the Cross study).

5) Class B: Bicycle Ride-Out At Intersection (MAPC Rev: 11.9%; MAPC: 9.4%; Cross NF: 17.0%; Missoula: 10%)

Class B accidents involve bicyclists emerging from one leg of an intersection, and colliding with a motorist emerging from the orthogonal leg of the intersection. Wrong-way cyclists were involved in 18 percent of Class B accidents, compared to their 24 percent share of all accidents.

Unlike Class A accidents, which involved bicycle ride-out from mid-block locations, Class B accidents occur among a slightly older population. Over 21 percent of these accidents occurred among bicyclists over 25 years of age (approximately the same proportion in which this age group is represented in the study sample). None of these accidents occurred to cyclists between 19 and 25 years of age, whereas over 40 percent occurred among those between 15 and 18 years. In fact, the 19-25 year age group seemed remarkably exempt from accidents. Twenty-one percent of Class B accidents occurred among cyclists between the ages of 12 and 14, and sixteen percent among those under 11 years of age.

Class B weekday accidents occurred with a greater frequency during both the morning peak hours (24 versus 15 percent) and afternoon peak hours (51 versus 41 percent) than did other accident classes. This was also true on weekends (20 percent, morning peak; 50 percent, afternoon peak). They were less likely than other accident classes to occur during mid-day, particularly on weekdays (8 percent versus 29 percent).

Class B accidents were slightly over-represented among the accidents involving serious injuries.

The most frequent type among Class B accidents was an unnumbered type "-", "Bicyclists entering intersections on a red light.'' The 6.5 percent of accidents occurring in this type was higher than both the Cross and Missoula studies, which showed 1.2 percent and 0 percent, respectively, of this type of accident. This discrepancy may in part be due to coding; Cross indicates in his narrative that he was only likely to include an accident in this type if the cyclist emerged well after the light turned red. The MAPC coder generally placed an accident in this type whenever the cyclist entered on the red.

6) Class G: Slowed or Parked Cars (MAPC Rev: 11.3%; MAPC: NA; Cross NF: 2.07%; Missoula: 3.3%)

This class, which was created for the MAPC study, includes accidents in which a bicycle collides with a motor vehicle that is slowed or stopped in traffic, entering or exiting on-street parking, or that has a door opening to let the driver out. Comparison with the percentages for the Cross and Missoula studies of the aggregates of these three types of accidents show that the MAPC region is much higher in the relative frequency with which these accidents occur. This may be due to the greater congestion and narrower widths of the major urban thoroughfares in the MAPC study area.

Only 10 percent of Class G accidents involved wrong-way cyclists, compared to 24 percent of all accidents in the study.

Class G accidents are more common among older bicyclists, with 87 percent occurring among bicyclists 15 years of age and over. Over 64 percent of these accidents occur among bicyclists over 18 years of age.

Class G accidents are unusual in that, unlike all other classes except for Class B, they occur with a greater relative frequency during the morning peak hours (between 6 a.m. and 10 a.m.), both on weekdays and weekends.

Class G accidents are somewhat less likely to occur during the afternoon peak hours (29 percent versus 41 percent of all accidents occurring during the afternoon peak).

In spite of the fact that these accidents involve a slowed or stopped motor vehicle, they are as likely to result in serious injury as the other accidents studied.

The most frequent type represented in this class is Type 41, "Cyclist strikes open door on driver's side of parked car," containing 5.3 percent of all accidents. This type accounted for only .8 percent of all accidents in the Cross study, and they were negligible enough in the Missoula study to be classified as Type 36, "Weird." Again, further investigation is needed to explain this higher relative frequency, but it is reasonable to guess that the Boston area's narrow streets and traffic congestion are significant factors in causing this accident.

7) Class E: Bicyclist Unexpected Turn/Swerve (MAPC Rev: 8.9%; MAPC 8.8%; Cross NF: 14.2%; Missoula 8.9%

Class E accidents involve a bicyclist turning into the path of a motorist approaching from behind or ahead.

Wrong-way cyclists were involved in 21 percent of these accidents, close to the 24 percent of all accidents involving wrong-way cyclists.

Like Class A accidents, Class E accidents occurred among a younger population - 42 percent among bicyclists age 11 and under. Cyclists between 15 and 18 years of age were also overrepresented in this age group, with 35 percent of Class E accidents.

Class E accidents occurred more frequently during the weekday evening hours (7 pm-1 am) than did the sample as a whole (29 percent versus 19 percent). They were last likely to occur during the afternoon peak (39 percent ). On weekends, they were twice as likely as the average accident to occur during the afternoon peak, (14 percent compared to 7 percent).

Class E accidents distributed themselves among the various injury levels in approximately the same proportion as the overall sample.

Type 18 accidents - "Bicyclist unexpected left turn with auto approaching from same direction" - accounted for the greatest proportion of Class E accidents.

8) Class D: Motorist Overtaking/Overtaking Threat (MAPC Rev: 3.40; MAPC 8.3°%; Cross NF: 10.5%; Missoula: 13.3%)

Class D accidents involved a motorist striking a bicycle from behind or aside of the bicyclist. As with the Cross study, this was the least frequent class in the study. The difference between the revised MAPC percentage and the MAPC Cross classification scheme percentage is the removal of parked car door accidents from this class and their placement in Class G.

Wrong-way riding contributed to only 7 percent of these accidents.

Class D accidents were most likely to occur among cyclists 15 years of age and over (67 percent).

Class D accidents were over-represented among evening and mid-day accidents (both 27 percent, compared to 19 percent and 24 percent for the sample). They occurred with greatest frequency during the afternoon peak (36 percent). All of the weekend Class D accidents occurred between 7 p.m. and 6 a.m.

Class D accidents were the least likely among all accident classes to result in no reported injuries, but, unlike the Cross study, they were more likely to cause minor injuries rather than the severe or fatal injuries. Given the smaller number of cases in this class, the one fatality that occurred comprised a higher proportion of Class D accidents (6.7 percent) than they did of any other accident class.