APPENDIX B.

Detailed Summary of Descriptive Statistics

This appendix provides a detailed summary of descriptive statistics for pedestrian fatalities with specific characteristics. For each characteristic, we present the total number of pedestrian fatalities and proportion of these fatalities that occurred in darkness.

Our dataset shows that the total number of US pedestrian fatalities increased by 51% between 2010 and 2020 (4,302 to 6,516 fatalities per year) (Table B1). During the same period, pedestrian fatalities during darkness increased by 63% (3,030 to 4,951 fatalities per year). These trends resulted in the proportion of pedestrian fatalities that occurred during darkness increasing from 71% to 76% between 2010 and 2020.

The following subsections examine these yearly data for individual variables in the following categories: geographic, temporal, and weather; location; roadway; vehicle; movement; behavior; pedestrian demographic characteristics; and driver demographic characteristics.

B.1. Temporal and Weather

Temporal and weather variables provide important context about characteristics that may be associated with pedestrian fatalities in darkness.

There is a great deal of seasonal variability associated with pedestrian fatalities in all light conditions, but especially in darkness (Table B2). The winter months, especially, are associated with the highest fatality risk for pedestrians, likely related to their greater hours of darkness in the northern hemisphere. In particular, December has the greatest number of pedestrian fatalities, 6,392, and also the highest proportion occurring

B.2. Location

Differences in pedestrian fatalities in darkness are associated with intersection versus non-intersection locations. More than 80% of pedestrian fatalities with known lighting conditions and locations occur outside of intersection locations, and 76% of these fatalities happen in darkness (Table B4). Intersection locations have a lower proportion of fatalities in darkness (64%). Intersections are more likely to have lighting, signals, crosswalks, slower vehicle speeds, and increased driver attentiveness, all of which may be associated with their lower proportion of fatalities in darkness.

The data also suggest different outcomes depending on the leg of the intersection on which a crash occurs. Although fewer fatalities overall occurred between 2014 and 2020 at near side intersection legs than far side legs (3844 and 5398, respectively), nearside legs were associated with a significantly higher proportion of fatalities in darkness (73%) than far side legs (61%) (Table B5). This data element was not known for all crashes and that non-intersection fatalities are included in the not applicable category (Table B5). Further study of the differences between near side and far side intersection legs, including driver interactions with

B.3. Roadway

Several roadway design characteristics are associated with pedestrian fatalities during darkness. Locations with treatments designed to indicate a preferred crossing location or draw attention to pedestrians are associated with fewer fatalities in darkness. Nearly 90% of pedestrian fatalities with known lighting conditions and locations occur outside of marked crosswalks (Table B7). The increasing proportion of fatalities outside of crosswalks occurring in darkness between 2010 and 2020 roughly reflects the overall increase in nighttime pedestrian fatalities over the study period. During this same period, the proportion of fatalities within marked crosswalks in darkness was consistently underrepresented relative to overall fatalities in darkness, suggesting that marked crosswalks are relatively safer than other crossing locations at night.

Beginning in 2014, FARS included a data element indicating whether a marked crosswalk was present at the crash location (Table B8). Approximately 83% of pedestrian fatalities with known lighting conditions between 2014 and 2020 occurred at locations where a marked crosswalk was not present. More than 77% of these fatalities occurred in darkness, and these locations were consistently overrepresented at night relative to overall fatalities.

The statistics presented in Table B7 and Table B8 both suggest that the presence of a marked crosswalk, is associated with fewer fatalities in darkness. It is possible that changes in both driver and pedestrian behavior at these locations contribute to safer outcomes. Marked crosswalk locations may also be more likely to have better lighting and other safety treatments that may reduce nighttime pedestrian risk.

The greatest number of pedestrian fatalities occur on roadways with posted speed limits between 35 mph and 50 mph (Table B9). However, higher-speed roadways have a greater proportion of fatalities occurring in darkness, with more than 81% of fatalities on roadways with posted speeds higher than 50 mph occurring in darkness. While roadways with speed limits of 30 mph or less have the lowest numbers of fatalities overall and the lowest percentage of fatalities in darkness, they experienced the greatest increase in the proportion of fatalities during darkness over the study period.

Roadways with more lanes were more likely to experience pedestrian fatalities in darkness than roadways with fewer lanes (Table B10). Approximately 79% of pedestrian fatalities on roadways with four or more lanes were at night, while 67% of pedestrian fatalities on roadways with three or fewer lanes were at night. The differences in these categories may reflect distribution between urban and rural areas – i.e., that rural areas are more likely to have two-lane roadways that are higher-speed, making the crashes harder to avoid and more likely to be fatal if and when they occur in daylight. The proportion of fatalities that occurred during darkness increased slightly over the study period in all lane categories.

Closely associated with higher posted speed limits and more lanes, arterial roadways and freeways experience the highest proportion of pedestrian fatalities in darkness (Table B11). As the sites of nearly 69% of all pedestrian fatalities – an average of 76% of which occurred in darkness, arterials are the most dangerous roadway type for pedestrians. While the proportion of fatalities occurring in darkness remained relatively consistent on local roads and freeways over the study period, the proportion occurring on arterials increased steadily from approximately 73% in 2010 to over 78% in 2020.

B.4. Vehicle Characteristics

Passenger cars and pickup trucks, vans, and sport utility vehicles accounted for a nearly identical percentage of pedestrian deaths between 2010 and 2020 (Table B12). A greater percentage of pedestrian fatalities involving a passenger car, nearly 77%, occur in darkness than any other vehicle class presented here. Approximately 71% of fatalities caused by pickup trucks, vans, and sport utility vehicles occur in darkness. The proportion of fatalities occurring in darkness has increased over time for all vehicle classes. Shares of the overall vehicle fleet in the United States and exposure are important considerations when interpreting these statistics. A substantial number of commercial vehicles fall within the pickup truck, van, and sport utility vehicle class, and FARS does not differentiate these from privately owned vehicles. Thus, while the proportion of non-passenger cars has increased over the last decade, the fact that these commercial vehicles may be less likely to be active at night may simultaneously mitigate the crash risk they pose to pedestrians at night while also underrepresenting that risk. This phenomenon would also apply to commercial trucks and heavy vehicles.

B.5. Movement

Across all lighting conditions, nearly 84% of pedestrian fatalities with known vehicle maneuvers immediately prior to the crash are caused by vehicles traveling straight (Table B13). Vehicles traveling straight are also the most overrepresented in terms of their proportion of nighttime fatalities, with approximately 79% occurring in darkness. Vehicles making right turns are responsible for the fewest fatalities in terms of absolute numbers, followed by vehicles making left turns. These maneuvers are also least likely to cause a pedestrian fatality in darkness, with 29% of right turn and 28% of left turn fatalities occurring at night. This does not necessarily reflect the number of crashes caused by these maneuvers. Vehicles are typically moving at slower speeds or starting from a complete stop when executing these turns, which may reduce the likelihood of a fatality even in the event of a collision.

Another way to understand the circumstances surrounding a fatality is to consider the interaction between pedestrians and vehicles at intersections (Table B14). The interactions presented here are derived from the initial directions of the pedestrian and vehicle involved in the crash and are only available for pedestrian fatalities at intersections in FARS after 2014. Of the 7,022 pedestrian fatalities with known lighting conditions and interactions, slightly over 65% occurred in darkness. The interaction statistics suggest that a pedestrian intersecting a vehicle path left to right (i.e., crossing a roadway from the driver’s left in front of an approaching vehicle), is the most dangerous interaction. This is true both in terms of the absolute number of fatalities and the percentage occurring in darkness (78%). Slightly fewer pedestrians were struck while intersecting the vehicle’s path from right to left, or crossing from the passenger side in front of a vehicle, but a smaller proportion of these fatalities occurred in darkness (70%). Interactions with pedestrians and vehicles traveling in the same or opposite directions involve pedestrians being struck in the far side crosswalk when a vehicle was turning left or right. These interactions were involved in a much smaller number of fatalities overall, with a considerably smaller percentage occurring in darkness. Fatalities involving pedestrians traveling in the same direction as the vehicle (e.g., a northbound pedestrian being struck by an initially northbound vehicle that was turning left to go west) were more common at night (32%) than pedestrians traveling in the opposite direction (e.g., a southbound pedestrian being struck by an initially northbound vehicle that was turning left to go west) (22%).

We conducted further analysis of these interactions by driver and pedestrian movement and intersection leg. Pedestrian fatalities involving vehicles traveling straight had approximately the same likelihood of occurring during darkness when they were in the near side (77%, or 2,013 of 2,609) and far side crosswalks (76%, or 2,029 of 2,663). Fatalities involving pedestrians crossing from left to right in front of the driver were more likely to be at night in both the near side crosswalk (80% from left to right; 73% from right to left) and the far side crosswalk (80% from left to right; 72% from right to left).

Regarding the pedestrian-vehicle interaction statistics: The sample sizes for all interactions are relatively small, and there are a substantial number of unknown values for either pedestrian or vehicle initial direction of travel in FARS. The proportion of fatalities with known lighting condition and pedestrian-vehicle interaction occurring in darkness is consistently several percentage points lower than the overall proportion of pedestrian fatalities at night, which suggests that directional data is reported less frequently in darkness than in other lighting conditions. This may be related to the high proportion of hit-and-runs occurring at night or more difficulty reconstructing crashes that occurred during darkness. It also suggests that the percentages in Table B14 may be biased lower than the true percentages of fatalities with each type of interaction that occurred during darkness.

B.6. Behavior

Drivers who are drinking are frequently associated with nighttime fatalities (Table B15). Alcohol involvement was reported for drivers in 4,433 pedestrian fatalities, of which 82% occurred in darkness. The proportion of fatalities in which alcohol was not involved on the part of the driver approximately mirrors the overall proportion of fatalities occurring in darkness. These results likely reflect greater alcohol consumption during evening and nighttime hours in general. There are thousands of fatalities in which alcohol involvement was either not reported or reported as unknown. It is possible that many of these fatalities are related to hit-and-run crashes, which may be more likely to occur when a driver has been drinking.

The vast majority of pedestrian fatalities that involve the pedestrian using alcohol occur in darkness. Approximately 93% of fatalities with police-reported pedestrian alcohol involvement occur in darkness (Table B16). Similarly, approximately 88% of fatalities with police-reported pedestrian drug involvement occur in darkness (Table B17).

Although the absolute number of fatalities with alcohol or drugs involved has increased since 2010, the percentage of pedestrian fatalities involving police-reported pedestrian alcohol usage has declined steadily from 2010 to 2020 (39% to 32%, respectively; percentages not shown in the table). In contrast, the percentage of pedestrian fatalities involving police-reported pedestrian drug usage has increased steadily in that same period, from 17% to 27% (with variation; percentages not shown in the table), respectively, with 2019 and 2020 showing the biggest increases. These trends should continue to be monitored to help us to understand if recent years indicate a new trend or an anomaly.

Also, there are a large number of unknown values for alcohol and drug involvement. Further, the variable in FARS that indicates alcohol involvement (“DRINKING”) or drug involvement (“DRUGS”) for a pedestrian or driver is subjective, based on the judgment of the law enforcement officer investigating the fatality. The “DRINKING” variable also includes the presence of opened or unopened bottles of alcohol in a vehicle. Police-reported alcohol or drug involvement does not necessarily indicate that alcohol or drugs contributed to the collision. Despite these limitations, we think that “DRINKING” and “DRUGS” provide the best representation of alcohol or drug involvement in pedestrian fatalities. We did not use the blood test result variables in our analysis because they have many more missing values than the “DRINKING” and “DRUGS” variables.

Table B17. Pedestrian Fatalities in Darkness by Police-Reported Pedestrian Drug Involvement

FARS data do not suggest that reckless, careless, rage, and distracted driving are associated with the increase in nighttime pedestrian fatalities from 2010-2020 (Table B18). All of these factors are underrepresented relative to the overall proportion of fatalities in darkness and do not exhibit an increasing

B.7. Pedestrian Demographic Characteristics

Several pedestrian characteristics are associated with differences in nighttime fatalities. Different pedestrian age groups have very different outcomes (Table B20). Young pedestrians below age 15 and older than age 64 are considerably less likely to suffer a fatality in darkness than the overall population. Conversely, pedestrians between the ages of 15 and 64 are more likely to suffer a nighttime fatality, with ages 20-24, 25-34, and 35-44 experiencing the highest rates of fatalities in darkness (86%, 85%, and 83%, respectively). Much of this variation may be explained by differences in exposure. Children and older adults are less likely to be active outdoors and walking in darkness, while adults aged 20-34 may be relatively likely to be outside in darkness. However, the especially high proportion of fatalities in darkness for young adults between the ages of 20 and 34 may also suggest that there are behavioral differences which warrant further study to understand.

Males and females also experience highly disparate outcomes (Table B21). Male pedestrians are more likely to be killed in darkness than female pedestrians, with 76% and 68% of fatalities in darkness on average, respectively. However, while the proportion of fatalities in darkness for both sexes increased between 2010 and 2020, the proportion for female pedestrians grew more than that for male pedestrians. Given that male pedestrians are overrepresented overall, representing nearly 70% of fatalities in all lighting conditions, these statistics suggest that there may be pedestrian exposure or behavioral differences between the sexes.

Pedestrian race/ethnicity is also associated with differences in fatality risk (Table B22). American Indian or Alaska Native (86%) and Black pedestrians (80%) experience a disproportionately high percentage of fatalities in darkness. Hispanic/Latino pedestrians (75%) are slightly overrepresented in pedestrian fatalities at night, while Pacific Islander (74%) and White (72%) pedestrians are slightly underrepresented. Asian pedestrians experience the lowest proportion of nighttime fatalities, with approximately 56% of fatalities occurring in darkness.

The disparities in pedestrian fatalities between different races/ethnicities at night could potentially be due to differences in pedestrian activity levels at night (e.g., people of certain races/ethnicities being more likely to walk or take the bus to third-shift jobs). These disparities might also be due to differences in roadway and traffic conditions at night as well as nighttime social environments near where people live and work, nighttime behaviors, or other factors, as indicated in other research on pedestrian fatalities and race (Sanders and Schneider 2022).

People who are White and Asian are underrepresented in their share of pedestrian fatalities in all lighting conditions relative to their proportion of the United States population per the 2020 census. Indigenous, Black, and Hispanic/Latino groups have higher rates of fatalities.

Although they are not included in pedestrian fatality data within FARS, it is important to consider how darkness impacts the fatality risk for wheelchair users. In all lighting conditions, wheelchair users accounted for 159 fatalities between 2010 and 2020, of which 52% occurred in darkness (Table B23). There is considerable variation from year to year, likely due to the small sample sizes, yet wheelchair users consistently represent a smaller proportion of nighttime fatalities than other pedestrians.

This result does not indicate that wheelchair users are safer than other pedestrians during darkness. This result may instead be due to wheelchair users being proportionately less likely to travel at night than during the daytime than other pedestrians, since approximately 23% of people with travel-limiting disabilities reported restricting their travel to the daytime as a compensating strategy (Brumbaugh 2018). However, we do not have sufficiently detailed exposure estimates for wheelchair users or pedestrians as a whole by time of day to draw a conclusion. The small sample sizes also make meaningful statistical analysis of crash characteristics relative to other pedestrians difficult, and further research to understand the challenges and risks for wheelchair users would be valuable.

Wheelchair users are not included in NHTSA’s FARS pedestrian fatality data. Currently, wheelchair users are a subset of Person Type 8 (Person on Personal Conveyances) in FARS. This category also includes users of motorized or human powered scooters, roller skates, skateboards, and ridable toys. Wheelchair users would be more appropriately categorized as pedestrians who require an assistive device, much like pedestrians using a cane or a crutch. These individuals use the same facilities designed for other pedestrians and considering them as such would allow for better analysis and understanding of pedestrian fatalities.

B.8. Driver Demographic Characteristics

Driver characteristics are also associated with differences in nighttime pedestrian fatalities. In terms of driver age, drivers older than age 64 are less likely to be involved in a pedestrian fatality in darkness and are underrepresented relative to drivers of all ages (Table B24). Drivers between the ages of 20 and 34 are most likely to be involved in a nighttime pedestrian fatality. Although they are not responsible for the highest number of fatalities in absolute numbers, drivers in the 20-24 category have the highest percentage of fatalities in darkness (an average of 76% over the study period). However, the percentage of fatal pedestrian crashes occurring in darkness increased over the study period for all age groups (although there is year-to-year variation within each group). Like the findings for pedestrian demographic characteristics, the findings regarding groups who are more or less likely to be involved in a pedestrian fatality at night may be explained by a variety of factors, such as exposure. For example, older drivers may be aware of their own biological limitations, such as not being able to see as well in darkness and having slower reaction times, so they may be more likely to avoid driving at night than younger drivers. Behavioral differences and driving experience of different age groups may also affect their involvement in pedestrian fatalities at night (e.g., older drivers may drive slower and more deliberately at night).

Driver sex is a characteristic with pronounced differences in pedestrian fatalities across all light conditions (Table B25). Roughly the same proportion of pedestrian fatalities involving male drivers (73%) and female drivers (71%) occur in darkness, but males are much more likely to be involved in a pedestrian fatality overall (72% versus 28% for females, on average). These findings suggest that males drive more at all times of day, there is something inherently dangerous to pedestrians about the behavior of male drivers, or some combination of the two. It is well established that males, in particular, young males, take more risks while driving, which may underlie these statistics. Darkness does not appear to have a notable impact on the disparity in male and female drivers involved in pedestrian fatalities.