U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

 
REPORT
This report is an archived publication and may contain dated technical, contact, and link information
Back to Publication List        
Publication Number:  FHWA-HRT-13-098    Date:  January 2014
Publication Number: FHWA-HRT-13-098
Date: January 2014

 

Human Factors Assessment of Pedestrian Roadway Crossing Behavior

Factor Specific Analyses

This results section evaluates each of the recorded crossing variables in detail. Although there are too few data collection sites in phase 1 to use inferential statistics to compare one site with another, general trends are examined.

Crossing Location

Table 13 summarizes the percentage of pedestrians, by location, who crossed at marked intersections, unmarked non-intersections, and unmarked intersections. The percentage of each type of crossing is presented at each location. For example, at Location 1, 75.13 percent of the total crossings took place in the marked intersection during the walk phase. When all locations are combined, the mean percentage of pedestrians who crossed at the marked intersection is 85.35. However, the Location 3 marked intersection percentage of 50.9 is 2.27 standard deviations below the mean, making it an outlier.

Further examination of Location 3 shows that the relatively low number of pedestrians crossing at the marked intersection is the result of pedestrians crossing at the unmarked intersection. In fact, pedestrians were about equally likely to cross in the marked intersection (50.88 percent) as the unmarked intersection (44.65 percent).

If Location 3 is removed, the mean percentage of pedestrians who crossed at the marked intersection in the remaining seven locations is 90.27. Here, the Location 8 marked intersection value of 76.8 percent is 2.05 standard deviations below the mean, making it an outlier. The Location 7 marked intersection crossing value of 97.0 percent is 1.02 standard deviations above the mean. While this is not an outlier, it is the most extreme high value. As a result, this is explored further in the subsequent discussion section.

Because of the third crossing area choice at Location 3 (i.e., the unmarked intersection), it is worthwhile to further explore the percentage of pedestrians who crossed at unmarked non-intersection areas. Not surprisingly, the Location 8 unmarked non-intersection value of 28.2 percent is 2.32 standard deviations above the mean and consequently is classified as an outlier. If the Location 8 value is removed, the mean percentage of pedestrians crossing at an unmarked non-intersection at the remaining seven locations is reduced to 7.06. No further outliers exist. However, Location 1 and Location 7 remain the two most extreme values at 1.63 standard deviations above the mean and 1.34 standard deviations below the mean, respectively.

Further exploration for the possible reasons for these differences is presented in more detail in the phase 1 discussion section.

Table 13. Percentage of pedestrians at each crossing area in each data collection location.

 

Location

1

2

3

4

5

6

7

8

Marked Intersection

Walk

75.13

87.91

44.56

67.98

93.43

90.63

75.46

74.21

+traffic flow change

.72

1.22

.38

.71

.37

.82

1.30

.64

Don’t Walk

8.47

1.75

5.73

17.89

.49

.92

18.69

1.33

+traffic flow change

3.60

.52

.21

4.59

.05

.90

1.53

.64

Overall in Intersection

87.93

91.30

50.88

91.17

94.33

93.26

96.98

76.81

Unmarked Non-Intersection

With Traffic

5.23

6.24

1.00

5.38

5.19

4.67

1.81

16.58

+traffic flow change

.45

.30

.15

.17

.14

.27

.01

.64

Against Traffic

5.77

1.60

3.15

2.77

.33

1.11

1.20

5.46

+traffic flow change

.63

.56

.15

.52

.01

.68

.51

Overall in Unmarked Non-Intersection

12.07

8.70

4.46

8.83

5.67

6.74

3.02

23.19

Unmarked Intersection

 

 

 

 

 

 

 

 

With Traffic

18.34

+traffic flow change

1.62

Against Traffic

24.06

+traffic flow change

.63

Overall in Unmarked Intersection

44.65

— Indicates no actions of this type were recorded.

Pedestrian Yielding

Table 14 summarizes the percentage of pedestrians who yielded to vehicles within each crossing type. As an example, at Location 1, of the 834 pedestrians who crossed during the walk phase in the marked intersection, 1 pedestrian yielded to a vehicle. This is equivalent to .12 percent. Also at Location 1, of the total 976 crossings in the marked intersection, only 1 pedestrian yielded to a vehicle. This is the equivalent of .10 percent of the total crossings in the marked intersection.

Overall, the mean percentage of pedestrians who yielded to vehicles was .53. If all of the locations are compared with one another, the Location 3 value of 1.67 percent was 2.15 standard deviations above the mean and is considered an outlier. If Location 3 is removed from these values, no other outliers remain. These overall yielding behaviors do not provide much information about where pedestrians are yielding to vehicles. As a result, yielding behaviors by crossing area are examined next.

The mean percentage of pedestrians who yielded to vehicles in the marked intersection was .19. If all of the locations are compared with one another, the Location 7 value of .67 percent was 2.42 standard deviations above the mean and is considered an outlier. If Location 7 is removed, no further outliers exist.

Next, the mean percentage of pedestrians who yielded to vehicles in unmarked non-intersections was examined. Overall, the mean percentage of yielding was 2.19. No outliers exist here. In other words, when looking at these eight locations alone, none are significantly different with regard to the proportion of pedestrians who yielded to vehicles in unmarked non-intersection areas.

A t-test was performed to determine whether the proportion of pedestrian yielding behaviors differed between the marked intersection and the unmarked non-intersection.

A significant difference between the two locations was found, t(7) = -2.40, p = .048. Crossings where the pedestrian crossed entirely during the walk phase in the marked intersection were compared with all other crossings made where the pedestrian yielded. There was no significant difference in the percentage of pedestrians who yielded to vehicles during the walk phase in the marked intersection (M = .18 percent) compared with those pedestrians who yielded to vehicles in all other situations (M = 1.68 percent), t(7) = -2.11, p > .05.

Table 14. Percentage of pedestrians yielding to vehicles in each crossing area at each data collection location.

 

Location

1

2

3

4

5

6

7

8

Marked Intersection

Walk

.12

.12

.18

.03

.15

.74

.09

+traffic flow change

1.92

2.12

4.14

Don’t Walk

0

.17

+traffic flow change

25.00

.50

4.76

.51

Overall Yielding in Intersection

.10

.28

.17

.03

.19

.67

.08

Unmarked Non-Intersection

With Traffic

3.45

2.08

.14

.18

.65

.43

+traffic flow change

20

7.14

2.22

10

Against Traffic

12.16

5.49

2.86

8.20

.65

+traffic flow change

14.29

34.62

25.00

4.50

Overall Yielding in Unmarked Non-Intersection

2.99

6.20

4.74

.09

.33

2.35

.51

.27

Unmarked Intersection

 

 

 

 

 

 

 

 

With Traffic

.31

+traffic flow change

3.57

Against Traffic

5.36

+traffic flow change

9.09

Overall Yielding in Unmarked Intersection

3.27

Grand Percentage

.45

.80

1.67

.16

.05

.34

.66

.13

— Indicates no actions of this type were recorded.

Vehicle Yielding

Table 15 summarizes the percentage of vehicles that yielded to pedestrians within each crossing area at each location. As an example, at Location 1, there were three instances of vehicles yielding to pedestrians with traffic at the unmarked non-intersection. This is 5.17 percent of the 58 total crossings in this area.

The overall percentage of crossings that involved a vehicle yielding to a pedestrian was compared across locations. The mean percentage of crossings that involved vehicle yielding was 9.77. The Location 7 value of 38.70 percent was 2.15 standard deviations above the mean and considered an outlier. If the Location 7 value is removed, Location 4 becomes an outlier at 2.05 standard deviations above the mean. If Location 4 is removed from these values, no other outliers remain. These overall yielding behaviors do not provide much information about where vehicles are yielding to pedestrians. As a result, yielding behaviors by crossing area are examined next.

The mean percentage of vehicles that yielded to pedestrians in the marked intersections was 10.21. If all of the locations are compared with one another, the Location 7 value of 39.80 percent was 2.11 standard deviations above the mean and is considered an outlier. If Location 7 is removed, Location 4 becomes an outlier at 2.05 standard deviations above the mean. If Location 4 is removed from these values, no other outliers remain.

Next, the mean percentage of vehicles that yielded to pedestrians in unmarked non-intersections was examined. Overall, the mean percentage of yielding was 1.25. No outliers exist here. In other words, when looking at these eight locations alone, none are significantly different with regard to the proportion of drivers who yielded to pedestrians in unmarked non-intersection areas.

A t-test was performed to determine whether the proportion of vehicle yielding behaviors differed between the marked intersections and the unmarked non-intersections. No significant difference between the two locations was found, t(7) = 1.86, p > .05. Next, crossings where the pedestrian crossed entirely during the walk phase in the marked intersection were compared with all other vehicle yielding. Once again, there was no significant difference in the percentage of vehicles that yielded to pedestrians during the walk phase in the marked intersection (M = 12.62 percent) compared with those vehicles who yielded to vehicles in all other situations (M = .97 percent), t(7) = 1.81, p > .05.

Table 15. Percentage of vehicles yielding to pedestrians within each crossing area at each data collection location.

 

Location

1

2

3

4

5

6

7

8

Marked Intersection

Walk

4.08

7.22

29.88

.01

8.74

50.71

.34

+traffic flow change

28.21

.74

16.57

20

Don’t Walk

.62

+traffic flow change

.99

.51

Overall Yielding in Intersection

3.49

6.95

22.34

.12

8.50

39.80

.50

Unmarked Non-Intersection

With Traffic

5.17

.70

.18

.13

2.13

.38

+traffic flow change

6.67

10

Against Traffic

4.69

.61

.27

4.52

2.33

+traffic flow change

Overall Yielding in Unmarked Non-Intersection

4.48

.43

.52

.33

.09

3.07

1.10

Unmarked Intersection

With Traffic

.11

+traffic flow change

3.57

Against Traffic

1.04

+traffic flow change

Overall Yielding in Unmarked Intersection

.73

Grand Percentage

3.60

6.35

.35

20.4

.16

7.93

38.70

.64

          — Indicates no actions of this type were recorded.

Yielding Comparison

It is important to understand overall yielding behavior. Here pedestrian and vehicle yielding are compared. First, yielding behaviors that occurred within the marked intersection were explored. No significant difference in the percentage of pedestrians yielding to vehicles (M = 0.19 percent) and the percentage of vehicles yielding to pedestrians (M = 10.21 percent), was found, t(7) = -2.05, p > .05. Similarly, no significant difference in pedestrian yielding (M =  2.19 percent) and vehicle yielding (M = 1.25 percent) in pedestrian crossings occurring in unmarked non-intersections was found, t(7) = .87, p > .05.

Next, yielding behaviors within the marked intersection that took place entirely during the walk light phase were examined. A significant difference in the percentage of pedestrians who yielded to vehicles (M = .18 percent) and the percentage of vehicles who yielded to pedestrians was not found (M = 12.62 percent), t(7) = -1.95, p > .05. Next, yielding behaviors that took place while the pedestrian crossed outside the marked intersection or in the marked intersection that was not entirely during the walk light phase were examined. A paired comparison revealed that there was no significant difference in the percentage of pedestrians who yielded to vehicles (M = 1.68 percent) and the percentage of vehicles who yielded to pedestrians (M = 0.97 percent),
t(7) = .769, p > .05.

Evasive Pedestrian Actions

Table 16 summarizes the percentage of pedestrian evasive actions in each crossing area at each of the data collection locations. Each of the three types of evasive actions (running/ accelerated walking, abrupt stopping, and directional change) was combined to obtain a better overall perspective on pedestrian evasive actions. As an example, at Location 1, there were 29 instances of running/accelerated walking, 2 abrupt stops, and 1 directional change in the marked intersection during the walk phase. This is 3.72 percent of 834 total crossings in this area.

The overall percentage of crossings that involved a pedestrian evasive action was compared across locations. The mean percentage of crossings that involved a pedestrian evasive action was 4.39. None of the locations had a mean percentage that was more than 2 standard deviations from this mean. In other words, no outliers existed. These overall values do not provide much information about where pedestrians are taking evasive actions. As a result, evasive actions by crossing area are examined next.

The mean percentage of pedestrians who took evasive actions in the marked intersection was 3.44. If all of the locations are compared with one another, none are considered outliers.

Next, the mean percentage of pedestrians who took evasive actions in unmarked non-intersections was examined. Overall, the mean percentage was 10.17. No outliers exist here. In other words, when looking at these eight locations, none are significantly different in regard to the proportion of pedestrians who took evasive actions in unmarked non-intersection areas.

A t-test was performed to determine whether a difference existed in the percentage of pedestrian evasive actions between the marked intersection and the unmarked non-intersection. Indeed a significant difference between the two locations was found, t(7) = -2.85, p = .025. Next, crossings where the pedestrian made the entire crossing during the walk phase in the marked intersection were compared with all other pedestrian evasive actions. Here, once again, there was a significant difference in the percentage of pedestrians who took evasive actions during the walk phase in the marked intersection (M = 2.21 percent) compared with those pedestrians who took evasive actions in all other situations (M = 11.77 percent), t(7) = -3.51, p = .010.

Table 16. Percentage of pedestrian evasive actions in each crossing area at each data collection location.

 

Location

1

2

3

4

5

6

7

8

marked intersection

Walk

3.72

4.96

.22

1.60

1.70

1.83

.24

3.42

+traffic flow change

12.50

25.00

20.0

12.77

74.36

28.15

14.79

80.0

Don’t Walk

28.72

19.75

7.05

4.57

36.54

25.83

8.42

42.86

+traffic flow change

17.50

16.67

1.49

40.0

7.48

.51

70.0

Overall Evasive Actions in Intersection

6.76

5.56

1.13

2.26

2.18

2.36

2.01

5.29

Unmarked Non-Intersection

With Traffic

18.97

4.84

3.85

1.41

2.17

2.30

4.68

13.79

+traffic flow change

20.00

24.44

40.0

Against Traffic

31.25

21.62

9.15

1.92

14.29

11.48

14.84

40.70

+traffic flow change

28.57

26.92

12.50

5.41

12.50

Overall Evasive Actions in Unmarked Non-Intersection

24.63

9.18

7.76

1.46

3.32

5.52

8.70

20.82

Unmarked Intersection

With Traffic

.94

+traffic flow change

7.14

Against Traffic

4.08

+traffic flow change

Overall Evasive Actions in Unmarked Intersection

2.84

Grand Percentage

8.92

5.87

2.19

2.19

2.25

2.57

2.21

8.89

          — Indicates no actions of this type were recorded.

Evasive Vehicle Actions

Table 17 summarizes the percentage of vehicle evasive actions within each crossing area at each of the data collection locations. Each of the four types of evasive actions (abrupt braking—first vehicle, abrupt braking—second vehicle, directional change—first vehicle, and directional change—second vehicle) were combined to obtain a better overall perspective on vehicle evasive actions. As an example, at Location 2, there were four instances of abrupt braking by the first vehicle, two directional changes by the first vehicle, and four directional changes by the second vehicle in the marked intersection during the walk phase. This is .25 percent of the 4,071 total crossings at this area.

The overall percentage of crossings that involved a vehicle evasive action was compared across locations. The mean percentage of crossings that involved a vehicle evasive action was .05. None of the locations had a mean percentage that was more than 2 standard deviations away from the mean. In other words, no outliers existed. These overall values do not provide much information about where vehicles are taking evasive actions. As a result, evasive actions by crossing area are examined next.

The mean percentage of vehicles that took evasive actions in the marked intersection was .04. If all of the locations are compared with one another, the Location 2 value of .24 percent is 2.27 standard deviations above the mean, making it an outlier. If Location 2 is removed, not surprisingly, Location 1 (the only other non-zero value) becomes an outlier at 2.27 standard deviations above the mean.

Next, the mean percentage of vehicles that took evasive actions in unmarked non-intersections was examined. Overall, the mean percentage was .09. Not surprisingly, Location 1 (the only non-zero value) was an outlier at 2.47 standard deviations above the mean.

A t-test was performed to determine whether a difference existed in the percentage of vehicle evasive actions between the marked intersection and the unmarked non-intersection. No significant difference between the two locations was found, t(7) = -.568, p > .05. Next, crossings where the pedestrian completed the entire crossing during the walk phase in the marked intersection were compared with all other vehicle evasive actions. Here, there was no significant difference in the percentage of vehicles who took evasive actions during the walk phase in the marked intersection (M = .03 percent) compared with those vehicles who took evasive actions in all other situations (M = .09 percent), t(7) = -.600, p > .05.

Table 17. Percentage of vehicle evasive actions within each crossing area at each data collection location.

 

Location

1

2

3

4

5

6

7

8

Marked Intersection

Walk

.25

+traffic flow change

Don’t Walk

1.06

+traffic flow change

Overall Evasive Actions in Intersection

.10

.24

Unmarked Non-Intersection

With Traffic

1.72

+traffic flow change

Against Traffic

+traffic flow change

Overall Evasive Actions in Unmarked Non-Intersection

.75

Unmarked Intersection

With Traffic

+traffic flow change

Against Traffic

+traffic flow change

Overall Evasive Actions in Unmarked Intersection

Grand Percentage

.18

.22

            — Indicates no actions of this type were recorded.

Evasive Action Comparison

It is important to understand overall evasive action behavior. Here pedestrian and vehicle evasive actions are compared. First evasive actions that occurred within the marked intersection were explored. A significant difference in the percentage of pedestrians taking evasive actions (M = 3.44 percent) and the percentage of vehicles taking evasive actions (M = .04 percent) was found, t(7) = 4.74, p = .002. Similarly a significant difference in pedestrian evasive actions (M = 10.17 percent) and vehicle evasive actions (M = 0.09 percent) in pedestrian crossings occurring in unmarked non-intersections was found, t(7) = 3.54, p = .010.

Next, evasive behaviors within the marked intersection that took place entirely during the walk light phase were examined. A significant difference in the percentage of pedestrians who took evasive actions (M = 2.21 percent) and the percentage of vehicles that took evasive actions (M = .03 percent) was found, t(7) = 3.78, p = .007. Next, evasive actions that took place while the pedestrian crossed either in the unmarked non-intersection or in the marked intersection at least partially during the don’t walk phase were examined. Again, there was a significant difference in the percentage of pedestrians who took evasive actions (M = 11.77 percent) and the percentage of vehicles that took evasive actions (M = .09 percent), t(7) = 3.83, p = .006.

 

Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101