Investigating Improvements to Pedestrian Crossings With An Emphasis on The Rectangular Rapid-Flashing Beacon

CHAPTER 3. GATHER DATA ON PEDESTRIAN CRASHES

This chapter documents the results of the review and analysis of pedestrian crash databases. The analysis of crash datasets had the following two key objectives:

• Review specific potential data sources to assess their suitability for safety evaluations that may be conducted in the research.
  
  
• Review the characteristics, circumstances, and contributing factors for crashes at midblock pedestrian crossings.
  
  

REVIEW OF CRASH DATABASES

A review of available crash datasets was conducted to assess which datasets were most promising for characterizing the attributes, circumstances, and contributing factors for crashes at midblock pedestrian crossings. National and State databases were considered in this assessment. The following crash databases were reviewed:

• National Highway Traffic Safety Administration (NHTSA) Fatality Analysis Reporting System (FARS).
  
  
• NHTSA General Estimates System (GES).
  
  
• NHTSA National Motor Vehicle Crash Sampling Survey (NMVCSS).
  
  
• FHWA Highway Safety Information System (HSIS) data for California, Illinois, Minnesota, North Carolina, Ohio, and Washington, as well as similar data for Texas.
  
  

The databases were selected for review because of their broad scope (especially the national databases), their reputation for quality, and their accessibility. These databases represent complete national or statewide datasets that are made available to researchers. Other agencies are often reluctant to release complete datasets.

An initial screening of the candidate data sources was performed to assess their suitability for analysis of midblock pedestrian crossings. This screening was conducted primarily with documentation rather than with the actual data, although portions of each dataset were given an initial review. The screening of the crash data sources addressed the following questions:

• Can crashes involving pedestrians be distinguished from other crashes?
  
  
• Can midblock pedestrian crashes be distinguished from intersection pedestrian crashes?
  
  
• Can crashes involving a pedestrian traveling along the road be distinguished from crashes involving a pedestrian crossing the road (or waiting to cross the road)?
  
  
• Can crashes involving a pedestrian in the traveled way be distinguished from crashes involving a pedestrian on a roadside or shoulder?
  
  
• Can crashes involving a pedestrian in a marked midblock crosswalk be distinguished from other crashes involving pedestrians in the traveled way?
  
  
• Can the type of traffic control device present at a midblock pedestrian crossing be determined?
  
  
• Can vehicle–vehicle crashes at midblock pedestrian crossings be distinguished from other vehicle–vehicle crashes?
  
  

The assessments of these questions are presented in table 3 for the national and State databases reviewed.

Table 3 shows that all of the crash datasets reviewed can distinguish pedestrian crashes from other crash types. Also, all of the crash datasets reviewed, except the NMVCSS, can distinguish midblock pedestrian crashes from intersection pedestrian crashes. NMVCSS was dropped from further consideration because of this deficiency.

The table shows that none of the databases distinguish clearly between crashes involving a pedestrian traveling along the road and crashes involving a pedestrian crossing the road. In other words, the available datasets generally lack a data element that clearly identifies the pedestrian action underway prior to the crash. Lack of data at this level of detail generally limits the application of the FHWA Pedestrian and Bicycle Crash Analysis Tool unless supplementary data were obtained from the review of hard-copy police crash reports.

Most datasets, other than the Washington State crash data, can distinguish crashes involving a pedestrian in a marked midblock crosswalk from other crashes involving pedestrians in the traveled way. It should be noted that only marked crosswalks at midblock locations are considered midblock crosswalks. There is no implied crosswalk at any midblock location unless markings and/or signs identify the location as a crossing. In the Washington data, midblock pedestrian crashes that occur at a signal, stop sign, or yield sign were presumed to occur at or near a crosswalk.

The table shows that only some crash datasets can identify the type of traffic control present at a midblock pedestrian crossing. Roadway characteristics files do not generally identify the locations of midblock crossings or the types of traffic control present at such crossings. However, in most States, the type of traffic control at a midblock crossing can be inferred from the traffic control information in crash data for crossings at which crashes have occurred.

Finally, the table shows that there are only limited cases in which a vehicle–vehicle collision at a midblock crossing (e.g., a rear-end collision with a vehicle stopped at the crossing) can be clearly distinguished from similar collisions at other midblock locations.

Based on the results shown in table 3, a decision was reached to proceed with analysis of all of the datasets tabulated, with the exception of the NMVCSS dataset.

Table 3. Assessment of screening questions for national and State databases.

^aNationwide data for fatal crashes only.
^bBroad-based sample of nationwide crashes of all severity levels from specific primary sampling units.
^cA special-purpose sample of crashes of all severity levels whose locations are not identified.
^dCrashes on State-maintained highways only for all severity levels.
^eIncludes crashes of all severity levels on nearly all roads statewide, including State-maintained and local facilities.
^fIncludes crashes of all severity levels the majority of the State-maintained road system; many State-maintained roads in North Carolina are equivalent to county roads or city streets in other States.
^gYes, in some, but not in all cases.
^hCan be presumed indirectly from traffic control devices present.
ⁱYes, but only in the crash data for locations where crashes occur.
^jNo, except to the extent this can be inferred from the location type.
^kYes, to some extent.
— Unknown.

SUMMARY OF RESULTS

This section presents a summary of the results of an analysis of several national and State crash databases. The analysis focused on the characteristics, circumstances, and contributing factors for crashes at midblock pedestrian crossings.

NHTSA FARS

The FARS database contains data on all fatal crashes that occur on public roads in the United States. FARS crash data were reviewed for the years 2005 through 2009, inclusive. During this period, 22,892 pedestrian fatalities occurred in the United States, and 73.0 percent of the pedestrian fatalities occurred at midblock locations. The following statistics apply to the pedestrian fatalities that occurred at midblock locations:

• 1.3 percent occurred at marked midblock crosswalks.
  
  
• 30.7 percent occurred not in, but near, a marked midblock crosswalk.
  
  
• 68.0 percent occurred at locations that were not near marked midblock crosswalks.
  
  

Other highlights from the analysis of FARS data include the following:

• Most midblock pedestrian crashes occur at locations with no traffic control.
  
  
• For midblock crashes that occurred in or near a crosswalk, 18.6 percent were classified as “improper crossing.” However, this FARS pedestrian-related factor may be a bit ambiguous because at least two other categories—“failure to yield/obey” and “dart/run into roadway” —appear to be forms of “improper crossing.”
  
  
• The data show a general increasing trend in crash frequencies with increasing pedestrian age. Particularly notable is the high proportion of fatal pedestrian crashes for pedestrians of age 70 or older at midblock crosswalks.
  
  
• The proportion of male and female pedestrians killed at locations within midblock crosswalks is relatively even. However, for midblock locations near, but not within, crosswalks, and for midblock locations away from crosswalks, more than 70 percent of the victims were male. This suggests greater risk-taking behavior on the part of male pedestrians.
  
  
• FARS data for FHWA’s 13 Pedestrian Safety Focus States and 5 Pedestrian Safety Focus Cities were reviewed for fatal midblock pedestrian crashes. Collectively, the 13 Pedestrian Safety Focus States experience 59 percent of total U.S. fatal pedestrian crashes at or near midblock crossings, while the 5 Pedestrian Safety Focus Cities represent 6 percent of total U.S. fatal pedestrian crashes at or near midblock crossings.
  
  

NHTSA GES

In a review of a nationwide sample of crash data for all crash severity levels from GES, the database includes 10,079 crashes involving a pedestrian from 2005 to 2009, inclusive. Nearly half of the pedestrian crashes occurred at midblock locations. However, only 2.5 percent of the midblock pedestrian crashes were explicitly identified as midblock crossing crashes.

Other highlights from the analysis of GES data include the following:

• The GES data, which include pedestrian crashes of all severity levels, indicate that midblock crossing crashes are typically less severe than pedestrian crashes elsewhere on the midblock roadway.
  
  
• “No control” was the selected traffic control device for 65 percent of midblock crosswalk crashes and 91percent of the pedestrian crashes that occurred elsewhere on the midblock roadway.
  
  
• The GES variable called “pedestrian action” is similar to the FARS “pedestrian-related” factor. For midblock crosswalk crashes, 4.9 percent were classified as “improper crossing,” while a substantial number of other midblock crosswalk crashes were classified in categories closely related to “improper crossing”: dart/run into roadway, inattentive, and “playing, working, sitting, lying, etc., in the roadway.”
  
  
• When crashes for a range of severity levels were included, there was a large proportion of crashes (more than 20 percent) for young pedestrians in the age range from 11 to 20 (i.e., primarily teenagers).
  
  
• As with the FARS data, there was a greater gender balance in pedestrians crossing at midblock pedestrian crosswalks (actually more females than males), but a substantially higher proportion of male pedestrians crossing at non-crosswalk locations.
  
  

FHWA HSIS

California

HSIS data for California include crash data only for the State highway system, consisting of approximately 15,520 mi of highways. Data analyzed for this report include the years from 2006 to 2008, inclusive. During the study period, 3,944 pedestrian crashes occurred on the California State highway system. Nearly 70 percent of the pedestrian crashes occurred at midblock locations. Only 2.6 percent of the midblock crashes were classified as occurring at midblock crosswalks.

Other highlights from the analysis of California pedestrian crash data include the following:

• As observed in the GES data, crash severities appear to be lower for midblock crosswalk crashes than for other pedestrian crashes on the midblock roadway.
  
  
• Approximately 75 percent of pedestrian midblock crashes occurred on an undivided roadway.
  
  
• Most pedestrian crashes, including midblock crosswalk crashes, occurred at locations with no traffic control present.
  
  

Minnesota

HSIS data for Minnesota include crash data for nearly all crashes statewide, including those that occurred on both State-maintained and local-agency-maintained road systems. Data analyzed for this report include the years 2003 to 2007, inclusive. During the study period, 8,271 pedestrian crashes occurred in Minnesota. Approximately 29 percent of the pedestrian crashes occurred at midblock locations. Only 3.0 percent of the midblock crashes were classified as occurring at midblock crosswalks.

Other highlights from the analysis of Minnesota pedestrian crash data include the following:

• In contrast to other States, the proportion of fatal and incapacitating injury crashes is higher for midblock crosswalks than for other midblock locations in Minnesota, but the sample size for midblock crosswalk crashes is so small that this may not be a valid comparison.
  
  
• The data for midblock roadways include a substantial number of pedestrian crashes at either signals or stop signs; these crashes must either have occurred at nonintersection signals or stop signs (i.e., driveways) or the basic intersection versus nonintersection classification of the crashes is incorrect.
  
  

North Carolina

HSIS data for North Carolina include crash data for approximately 62,000 mi of the 77,000mi of roadway on the State-maintained highway system. Data analyzed for this report include the years from 2005 to 2008, inclusive. During the study period, 3,847 pedestrian crashes occurred on the North Carolina State highway system. Nearly 85 percent of these pedestrian crashes occurred at midblock locations. Only 2.7 percent of the midblock crashes were classified as occurring at midblock crosswalks.

Ohio

HSIS data for Ohio include crash data only for the State highway system, consisting of approximately 19,500 mi of highways. Data analyzed for this report include the years from 2005 to 2008, inclusive. During the study period, 4,127 pedestrian crashes occurred on the Ohio State highway system. Approximately, 45 percent of these pedestrian crashes occurred at midblock locations. Only 1.2 percent of the midblock crashes were classified as occurring at midblock crosswalks.

Texas

Data for Texas includes crash data for crashes both on and off the State highway system. Data analyzed for this report include the years 2003 to 2009, inclusive. During the study period, 3,134,365 crashes were included in the Texas crash database. Of these, 39,993 (1.3 percent) were pedestrian crashes. Nearly 50 percent of the pedestrian crashes occurred at midblock locations. Only 136 crashes (0.7 percent of the midblock crashes) were classified as occurring at midblock crosswalks.

Washington

HSIS data for Washington include crash data only for the State highway system, consisting of approximately 7,193 mi of highways. Data analyzed for this report include the years from 2005 to 2008, inclusive. During the study period, 1,573 pedestrian crashes occurred on the Washington State highway system. Nearly 40 percent of these pedestrian crashes occurred at midblock locations. Only 5.0 percent of the midblock crashes were classified as occurring at midblock crosswalks.

RESULTS FROM ANALYSIS OF CRASH DATASETS

This section of the report presents the results of the analysis of all the crash datasets reviewed in the preceding section. In the tables presented in this section, unless otherwise stated, all percentages shown in parentheses are column percentages. Crashes classified as “unknown” for a given data element are not included in the percentages for that data element.

The crashes classified as “midblock crosswalk” crashes involved collisions with a pedestrian at or near a marked midblock crosswalk. Only the FARS dataset formally distinguishes between crashes for pedestrians in a crosswalk or near a crosswalk. The crashes classified as “midblock roadway” crashes involve collisions with a pedestrian removed from (i.e., not near) a crosswalk. The “midblock crosswalk” and “midblock roadway” crashes include only nonintersection crashes (i.e., not at or related to an intersection) that occurred in the traveled way or on the roadway shoulder; crashes that occurred outside the shoulder (i.e., on the roadside or on a sidewalk) have been excluded.

Some datasets show a substantial number of “midblock roadway” crashes that occur at signals or stop signs. Such crashes either represent signalized or stop-controlled driveways on the midblock roadway or the basic intersection vs. nonintersection categorization is incorrect.

The first two datasets analyzed are national datasets, the first a census of fatalities and the second a sample of crashes for all severity levels. The remaining datasets are for either the State highway system or a broader set of roads in individual States.

NHTSA FARS

The FARS database contains data on all fatal crashes that occur on public roads in the United States. These data can be queried on a number of crash, vehicle, and person-level variables that are commonly available in crash reports. Queries of several variables related to pedestrian fatalities such as pedestrian age, gender, and action; time of day and day of week; speed limit; number of lanes crossed; weather, lighting and road surface conditions; and presence of a median are shown in the following tables.

Nationwide FARS Data

Table 4 summarizes nationwide data for fatalities from the FARS dataset. The table shows that of the 198,708 fatalities that occurred in the United States in the years 2005 through 2009, inclusive, 22,892 (12.1 percent) involved pedestrians. Of those pedestrian fatalities, 16,700 (73.0percent) occurred at midblock locations; the remaining 27.0 percent of pedestrian fatalities occurred at intersections. Only 223 (1.3 percent) of the pedestrian fatalities at midblock locations occurred at marked midblock crosswalks. Another 5,129 (30.7 percent) of pedestrian fatalities at midblock locations occurred not in, but near, a marked midblock crosswalk. Thus, 5,352 of pedestrian fatalities at midblock locations occurred in or near a crosswalk. The remaining 11,348(68.0 percent) of pedestrian fatalities at midblock locations occurred at locations that were not near marked midblock crosswalks.

Table 4. Summary of FARS pedestrian fatalities data (2005–2009).

Table 5 presents nationwide data from FARS for the type of traffic control present at the locations of fatal midblock pedestrian crashes. The table shows clearly that most midblock pedestrian crashes occur at locations with no traffic control. For locations at or near marked midblock crossings, 85 percent of fatal pedestrian crashes occurred at locations with no control, while 15 percent occurred at locations with some positive control. Clearly, “no control” in this database is being interpreted to mean no control other than pavement markings or signing, because such controls must be present at a midblock crossing. Table 5 also illustrates some other classification issues in the data. The fatal pedestrian crashes with a “midblock roadway” location presumably occurred at locations removed from midblock pedestrian crossings and intersections, yet 2.1 percent of the crashes appear to have occurred at signals or at stop signs, including 0.1percent at locations with pedestrian signals. Possibly, these crashes occurred at signalized or stop-controlled driveways; for purposes of this analysis, these might better be classified as intersection crashes.

Table 6 shows nationwide data for a FARS variable called “pedestrian-related factor” that describes the action taken by the pedestrian(s) involved in the crash. For midblock crashes that occurred in or near a crosswalk, 18.6 percent were classified as “improper crossing,” while 8.0percent were classified as “dart/run into roadway,” and 6.5 percent were classified as “failure to yield/obey.” Both of the latter categories appear to be forms of “improper crossing.” Nearly 3percent of midblock crosswalk crashes are classified as “walk, etc., in roadway.” This category is ambiguous, because it could imply improper crossing or it could imply that a pedestrian was walking along the roadway (i.e., traveling in a longitudinal direction) in the traveled way and happened to be struck by a motor vehicle at the location of a midblock crosswalk. The “walk, etc., in roadway” crashes in the “midblock roadway” column clearly imply that the pedestrian was walking along the roadway.

Table 5. FARS midblock pedestrian fatalities by traffic control device (2005–2009).

RR = Railroad.
— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 6. FARS midblock pedestrian fatalities by pedestrian actions (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 7 presents nationwide data for fatal pedestrian crashes at midblock locations by pedestrian age. The table shows a general increasing trend in crash frequencies with increasing pedestrian age. Particularly notable is the high proportion of fatal pedestrian crashes for pedestrians of age 70 or older at midblock crosswalks. There is an unexplained trend in all columns of the table showing that pedestrians in the 61 to 70 age group had a lower proportion of crashes than the age groups on either side.

Table 8 presents nationwide data for fatal pedestrian crashes at midblock locations by pedestrian gender. The table shows a relatively even proportion of male and female pedestrians killed at locations within midblock crosswalks (56 and 44 percent). However, for midblock locations near, but not within crosswalks, and for midblock locations away from crosswalks, more than 70percent of the victims were male. This suggests greater risk-taking behavior on the part of male pedestrians.

Table 9 summarizes the fatal pedestrian crash data by the number of lanes crossed. While interesting, these data are not terribly meaningful without exposure data on the number of lanes crossed by pedestrians who were not killed in crashes. The same is true for table 10, which categorizes fatal pedestrian crashes by the presence of a median, and thus a pedestrian refuge area, on the roadway crossed and table 11, which categorizes crashes by the posted speed limit at the crash location. Table 12 through table 16 present nationwide FARS data for fatal pedestrian crashes at midblock locations classified by weather condition, road surface condition, light condition, hour of the day, and day of the week, respectively.

Table 7. FARS midblock pedestrian fatalities by pedestrian age (2005–2009).

Column percentages may not appear to add to 100 percent due to rounding.

Table 8. FARS midblock pedestrian fatalities by pedestrian gender (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 9. FARS midblock pedestrian fatalities by number of lanes crossed (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 10. FARS midblock pedestrian fatalities by presence and type of median
(2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 11. FARS midblock pedestrian fatalities by posted speed limit (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 12. Midblock pedestrian fatalities by weather condition (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 13. FARS midblock pedestrian fatalities by road surface condition (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 14. FARS midblock pedestrian fatalities by light condition (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 15. FARS midblock pedestrian fatalities by hour of day (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 16. FARS midblock pedestrian fatalities by day of week (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

FARS Data by State

Table 17 presents the frequencies of fatal pedestrian crashes at midblock locations by State. Clearly the fatal crash frequencies are strongly influenced by State population; however, the extent of walking activity and the pedestrian age distribution also likely have an important role. For example, Florida has the highest frequency of fatal pedestrian crashes at midblock crosswalk locations in the United States and nearly as high a frequency of total midblock crashes as California, the most populous State.

The data in table 17 also suggest inconsistencies in the classification of midblock “in crosswalk” versus “near crosswalk” crashes, because the frequency of crashes in midblock crosswalks in many States is so small.

Table 17. FARS midblock pedestrian fatalities by State (2005–2009).

FARS Data for FHWA Pedestrian Safety Focus States and Cities

Table 18 summarizes the FARS data for fatal midblock pedestrian crashes in FHWA’s 13 Pedestrian Safety Focus States and 5 Pedestrian Safety Focus Cities, because they were designated during the period reported (2005–2009). Collectively, the 13 Pedestrian Safety Focus States experience 59 percent of the total U.S. fatal pedestrian crashes at or near midblock crossings, while the 5 Pedestrian Safety Focus Cities represent 6 percent of total U.S. fatal pedestrian crashes at or near midblock crossings.

Table 18. FARS midblock pedestrian fatalities for FHWA pedestrian safety focus States and cities (2005–2009).

NHTSA GES

GES is a database of a statistical sample of crashes from across the United States. These data are gathered in established primary sampling units (i.e., selected geographic areas) throughout the United States. The crash frequencies available in the GES database are shown in this section. The crash frequencies have not been inflated by sampling weights.

Table 19 summarizes the nationwide sample of crash data for all crash severity levels from GES. The database includes 680,316 crashes in the period from 2005 through 2009, inclusive. Of these crashes, 10,079 (1.5 percent) involved a pedestrian. Nearly half of the pedestrian crashes (46.7percent) occurred at midblock locations, while the rest occurred at intersections. Only 127(2.5 percent) of the 4,707 midblock pedestrian crashes were explicitly identified as midblock crossing crashes.

Because GES includes pedestrian crashes of all severity levels, table 20 presents the distribution of midblock pedestrian crashes by crash severity level. The table shows that midblock crossing crashes are typically less severe than pedestrian crashes elsewhere on the midblock roadway.

Table 21 presents GES data for pedestrian crashes by the type of traffic control device present. There was no control present for 65 percent of midblock crosswalk crashes and 91percent of pedestrian crashes elsewhere on the midblock roadway. As in the FARS data, a small percentage of “midblock roadway” crashes are shown as having traffic signals, stop signs, or yield signs present. This may represent traffic control at driveways.

Table 19. Summary of GES pedestrian crash data (2005–2009).

Table 20. GES midblock pedestrian crashes by injury severity (2005–2008).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 21. GES midblock pedestrian crashes by traffic control device (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 22 shows data for the GES variable called “pedestrian action,” which is similar to the FARS “pedestrian-related” factor. For midblock crosswalk crashes, 4.9 percent were classified as “improper crossing,” while a substantial number of other midblock crosswalk crashes were classified in categories closely related to “improper crossing”: dart/run into roadway, 13.1percent; inattentive, 2.5 percent; and “playing, working, sitting, lying, etc., in the roadway,” 0.8 percent. The categories related to “improper crossing” taken together constitute 21.3 percent of midblock crosswalk crashes.

Table 22. GES midblock pedestrian crashes by pedestrian action (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 23 presents the distribution of midblock pedestrian crashes by pedestrian age. When crashes for a range of severity levels are included, there is a large proportion of crashes (more than 20 percent) for young pedestrians in the age range from 11 to 20 (i.e., primarily teenagers). In addition, as with the FARS data, there are relatively fewer crashes in the pedestrian age group from age 61 to 70 than in the age ranges on either side.

Table 23. GES midblock pedestrian crashes by pedestrian age (2005–2009).

Column percentages may not appear to add to 100 percent due to rounding.

Table 24 presents the distribution of midblock pedestrian crashes in the GES data by pedestrian gender. As with the FARS data, there is a more even balance in pedestrian gender at midblock crossings (actually more females than males), but a substantially higher proportion of male pedestrians involved at noncrosswalk locations.

Table 24. GES midblock pedestrian crashes by pedestrian gender (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 25 summarizes the GES pedestrian crash data by the number of lanes crossed. While interesting, these data are not very meaningful without exposure data on the number of lanes crossed by pedestrians who were not killed in crashes. The same is true for table 26, which categorizes pedestrian crashes by the presence of a median, and thus a pedestrian refuge area, on the roadway crossed and table 27, which categorizes crashes by the posted speed limit at the crash location.

Table 25. GES midblock pedestrian crashes by number of travel lanes crossed (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 26. GES midblock pedestrian crashes by presence of median (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 27. GES midblock pedestrian crashes by posted speed limit (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 28 through table 31 present GES data for pedestrian crashes at midblock locations classified by weather condition, road surface condition, hour of the day, and light condition.

Table 28. GES midblock pedestrian crashes by weather condition (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 29. GES midblock pedestrian crashes by road surface condition (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 30. GES midblock pedestrian crashes by hour of day (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 31. GES midblock pedestrian crashes by light condition (2005–2009).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

California

HSIS data for California include crash data only for the State highway system, consisting of approximately 15,520 mi of highways. While some city streets and many small town “main streets” and suburban arterials are State highways, many of the areas with the highest concentrations of pedestrians in California are away from the State highway system. Data analyzed for this report include the years from 2006 to 2008, inclusive. Earlier years of data were not used because of changes in the coding of a key variable.

Table 32 shows that, during the study period, 526,898 crashes occurred on the California State highway system. Of these, 3,944 (0.7 percent) were pedestrian crashes. Nearly 70 percent of the pedestrian crashes occurred at midblock locations. Only 76 crashes (2.6 percent of the midblock crashes) were classified as occurring at midblock crosswalks.

Table 32. Summary of pedestrian crash data for California State highways (2006–2008).

Table 33 shows the relative frequencies of California crashes by midblock crosswalks and other midblock roadway locations.

Table 33. Midblock pedestrian crashes on California State highways by pedestrian location (2006–2008).

Percentages in this table are row percentages.

Table 34 classifies midblock pedestrian crashes in California by crash severity level. As observed in the GES data, crash severities appear to be lower for midblock crosswalk crashes than for other pedestrian crashes on the midblock roadway.

Table 34. Midblock pedestrian crashes on California State highways by injury severity (2006–2008).

Column percentages may not appear to add to 100 percent due to rounding.

Table 35 classifies midblock pedestrian crashes by whether they occurred on a divided or undivided roadway. Approximately 75 percent of pedestrian midblock crashes occurred on an undivided roadway.

Table 35. Midblock pedestrian crashes on California State highways by presence of median (2006–2008).

— Indicates count not included in calculation of percentage.
Column percentages may not appear to add to 100 percent due to rounding.

Table 36 classifies midblock pedestrian crashes in California by whether traffic control devices were present and functioning. Most pedestrian crashes, including midblock crosswalk crashes, occurred at locations with no controls present. Presumably, this refers to traffic signals or stop signs, because clearly, crosswalk markings (by definition) and signing were present at marked crosswalks.

Table 36. Midblock pedestrian crashes on California State highways by traffic control device operating (2006–2008).

Column percentages may not appear to add to 100 percent due to rounding.

Table 37 through table 43 present California data for pedestrian crashes at midblock locations classified by collision factors (contributing circumstances), weather condition, vehicle type at fault, road surface condition, light condition, hour of the day, and day of the week.

Table 37. Midblock pedestrian crashes on California State highways by collision factor (2006–2008).