Scholarly article on topic 'Study of Relation between Actual and Perceived Crash Risk'

Study of Relation between Actual and Perceived Crash Risk Academic research paper on "Civil engineering"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"Actual risk" / "Perceived risk" / "Indian city" / "Vulnerable road users" / "Ranking correlation"}

Abstract of research paper on Civil engineering, author of scientific article — Pankaj Prajapati, Geetam Tiwari

Abstract Fatality rates in Indian cities have increased manifolds in the recent years as cities continue to expand. The actual crash risk observed on different infrastructures is different and it depends upon the crash rate and the exposure on these infrastructures. This analysis uses crash data from police reported fatal crashes in the urban limits of the city to decide actual risk. The fatal crashes are accurately reported in police records, while minor and major crashes are under reported in India. The study considered risk to road users from six commonly used modes of transportation; walking, bicycling, riding motorized two- wheeler, auto-rickshaw, car, and bus. The perceived risk has derived for all these modes on different infrastructures; mid- block, signalized intersection, un-signalized intersection, and rotary intersection from household survey. This study examined the trend of fatal crashes and the relationship between actual and perceived crash risk on different infrastructures.

Academic research paper on topic "Study of Relation between Actual and Perceived Crash Risk"

Available online at www.sciencedirect.com

ScienceDirect

Procedia - Social and Behavioral Sciences 104 (2013) 1095 - 1104

2nd Conference of Transportation Research Group of India (2nd CTRG)

Study of Relation between Actual and Perceived Crash Risk

Pankaj Prajapatia1, Geetam Tiwarib

aResearch Scholar, Indian Institute of Technology Delhi, New Delhi, 110016, India and Associate Professor, The M S University of Baroda,

Vadodara, 390001, India bProfessor, Indian Institute of Technology Delhi, New Delhi, 110016, India

Abstract

Fatality rates in Indian cities have increased manifolds in the recent years as cities continue to expand. The actual crash risk observed on different infrastructures is different and it depends upon the crash rate and the exposure on these infrastructures. This analysis uses crash data from police reported fatal crashes in the urban limits of the city to decide actual risk. The fatal crashes are accurately reported in police records, while minor and major crashes are under reported in India. The study considered risk to road users from six commonly used modes of transportation; walking, bicycling, riding motorized two-wheeler, auto-rickshaw, car, and bus. The perceived risk has derived for all these modes on different infrastructures; mid-block, signalized intersection, un-signalized intersection, and rotary intersection from household survey. This study examined the trend of fatal crashes and the relationship between actual and perceived crash risk on different infrastructures.

© 2013 The Authors. Published by Elsevier Ltd.

Selectionandpeer-reviewunder responsibilityoflnternationalScientificCommittee. Keywords: Actual risk; Perceived risk; Indian city; Vulnerable road users; Ranking correlation

1. Introduction

The perceived traffic crash risk may play role in how people choose their mode of transport from the available modes to them. Therefore it is important for the transportation planner to know the perceived risk with respect to different transport infrastructure like mid-block, signalized, un-signalized, and rotary intersection. On other hand, the knowledge of actual traffic crash risk to different road users on the above infrastructures can help the transportation planner to improve the road safety. Fatality rates in Indian cities have increased manifolds in the recent years as cities continue to expand. The actual crash risk observed on different infrastructures is different and it depends upon the crash rate and the exposure on these infrastructures. The analysis in this study uses crash data from police reported fatal crashes in the urban limits of the Vadodara city to decide actual risk. The fatal

* Pankaj Prajapati. Tel.: +91-9427547621; fax: +91-0265-2423898. E-mail address: mailpankajs@yahoo.com

1877-0428 © 2013 The Authors. Published by Elsevier Ltd.

Selection and peer-review under responsibility of International Scientific Committee. doi: 10.1016/j.sbspro.2013.11.205

crashes are accurately reported in police records, while minor and major crashes are under reported in India. The study considered risk to road users from six commonly used modes of transportation; walking, bicycling, riding motorized two-wheeler, auto-rickshaw, car, and bus. The perceived risk has derived for all these modes on different infrastructures; mid-block, signalized intersection, un-signalized intersection, and rotary intersection from household survey. This study examined the trend of fatal crashes and the relationship between actual and perceived crash risk on different infrastructures. There are several studies conducted to identify such relations and they are found correlated; their correlations were sometime positive and sometime negative.

2. Literature Review

Klobucar & Fricker (2007) assumed in his study that bicyclists make decisions on the basis of perceived safety and travel distance. They presented two tools as most commonly used to quantify the perceived safety of a bicycle facility are the bicycle compatibility index (BCI) and the bicycle level of service (BLOS). Each evaluation tool was developed by using stated perceptions of the conditions faced by a bicyclist on various facilities and by using the properties of the facility and its environment to fit a linear regression to predict these perceptions. Based on this safe length have been calculated. It has been concluded that improvement in the total network path safe length indicates an improvement in the perceived safety of the bicyclists.

Parkin, Wardman, & Page (2007) developed two models of perceived risk; based on non-linear least squares, and a model of acceptability, based on the logit model, have been estimated for whole journeys based on responses from a sample of 144 commuters (2002) to video clips of routes and junctions. The risk models quantify the effect of motor traffic volumes, demonstrate that roundabouts add more to perceived risk than traffic signal controlled junctions and show that right turn manoeuvres increase perceived risk. The acceptability model confirms the effect of reduced perceived risk in traffic free conditions and the effects of signal controlled junctions and right turns.

Cho, Rodriguez, & Khattak (2009) examined how perceived and actual crash risks are related with each other and with respect to built environmental characteristics. Perceived risks for pedestrians and bicyclists were measured with a questionnaire including perceived neighborhood safety developed by the Neighborhood Quality of Life Study. Actual pedestrian and bicycle crash data were provided by the National Study Center for Trauma and EMS at the University of Maryland reported by police between January 2000 and December 2002. The secondary GIS data and primary data were collected through a street audit to characterize the built environment. The analytical approach was first to conduct an exploratory factor analysis of built environment variables to reduce them to a lower number of variables representing their underlying dimensions. Then, the predicted factor scores were included in a non-recursive path analysis mode. The results showed that residents who live in low density-single residential neighborhoods are more likely to perceive their neighborhood as dangerous relative to residents of compact, mixed-use neighborhoods even though the latter exhibited higher actual crash rates. The results of path analyses confirmed that a simultaneous but opposite relationship exists between perceived and actual crash risks. The results indicate that higher actual crash risk increases perceived crash risk, while higher perceived crash risk is negatively associated to actual crash rates. Consequently, low density and non-mixed land uses increase individuals perception of crash risk, and increased perception of risk and unfriendly environment for pedestrian and bikers reduces actual crash rates as a result of behavioral changes.

Noland (1995) examined behavioural responses to perceived risk in the mode choice for daily commute trips. The probability of an accident is measured by eliciting the respondent's perceived likelihood that an accident will occur. He stated that the advantage of using risk perceptions is that it is the relevant behavioural response variable to which people react. This allows the survey respondents to consider their own judgments about risky situations. Various other information was elicited by perceived travel time, cost, comfort of alternative modes and demographic information. The results of this research give evidence that increases in the perceptions of the risk of using a given transportation mode may reduce the probability of commuting by that mode.

Elvik and Bjornskau (2005) probed the extent to which the public accurately perceived differences in transport risks. Comparisons have been made between estimates of the fatality rate per billion kilometres of travel and four different summary measures of perceived risk. All these comparisons show high correlations between statistically estimated risk and perceived risk.

DeJoy (1992) presented results that provide some insight into the hypothesized relationship of optimism to the excess involvement of young males in traffic crashes. A particularly volatile combination appears to exist, in that, young males, relative to their female counterparts, possess an exaggerated sense of their own driving competency and they perceive less risk in a variety of dangerous driving behaviors. While the males in this study tended to be generally more optimistic than the females, the most pronounced differences occurred in the ratings of relative driving skill.

Finn and Bragg (1986) mentioned in their study that young drivers are significantly overrepresented among all drivers involved in traffic accidents and fatalities. Excessive risk taking by young drivers appears to be largely responsible for this disproportionate involvement. This excessive risk taking could be due to (1) being more willing to take risks than older drivers are, (2) failing to perceive hazardous situations as being as dangerous as older drivers do or (3) both causes. It was also concluded that young male drivers are overrepresented in traffic accidents at least in part because they fail to perceive specific driving situations as being as risky as older drivers perceive them.

Hayakawa, Fischbeck, & Fischhoff, B (2000) have studied the differences in the traffic risk and their perceptions in two countries. Research has found that traffic accidents are significantly more dreaded in Japan than in the US. As seen in analysis, the majority of traffic-accident deaths in Japan are not among car users. It was hypothesized that the more one-sided nature of fatal accidents in Japan (with cars killing 'vulnerable' motorcyclists, bicyclists, and pedestrians) is more likely to cause these feelings of dread than in the US where drivers kill themselves or equally-protected other drivers. Thus, it seems plausible that objective differences in risk environments combine with cultural influences to produce cross-national differences in risk perceptions.

Massie, Campbell, & Williams (1995) noticed elevated crash rates were observed for drivers aged 16-19 and 75 and over. The oldest drivers had the highest fatal involvement rate, while the youngest drivers had the highest rate of involvement in all police-reported crashes. Men had a higher risk than women of experiencing a fatal crash, while women had higher rates of involvement in injury crashes and all police-reported crashes.General guidelines for the preparation of your text

3. Data and MethodThe fatal crashes used in the analysis are crashes occurred on the urban roads other than local streets. The exposure has been calculated using self-digitized map of all arterial roads and urban highways lying within municipal limits of the Vadodara city. The fatal crash details have been derived from first investigation reports (FIR) collected from all police stations of the city by personal visit. The proportion of fatal crashes on arterial road, urban highway, and street is 56%, 37.2%, and 6.8% for total 643 crashes during study period of 6 years (2005-2010).The risk on major roads; i.e. arterial roads and urban highways at various infrastructures like mid-block, signalized, un-signalized, and rotary intersection is described by the following charts.

Figure 1. Impacting versus Victim type of crashes on major road signalized intersection

Only vulnerable road users are victim at signalized intersections. Bicyclists and motorized 2-wheeler occupants were killed by trucks and no pedestrian by trucks.

17% motorized 2-wheeler occupants were killed in single-vehicle crash. 17% pedestrians were killed by motorized 2-wheeler.

Car has killed 4% pedestrians and not any other type of road users.

In 26% cases motorized 2-wheeler was the impacting vehicle and in 61% cases they were victim also.

Other/ Unknown M2W I Auto I Car

■ Bus

I Truck

■ Single-vehicle

Victim type

Figure 2. Impacting versus Victim type of crashes on major road un-signalized intersection

8% motorized 2-wheeler occupants, 3% truck occupant, 2% auto-rickshaw occupants and less than 1% car and bus occupants were killed in single-vehicle crash.

40% victims were motorized 2-wheeler occupants, 8% from them were killed in single-vehicle crash, while 15% were by trucks.

Car killed 12% vulnerable road users (5.5% M2W, 4% pedestrian, 2% bicyclists).

Figure 3. Impacting versus Victim type of crashes on major road rotary intersection

• 20% out of 23% single-vehicle crash involved M2W, while 17% M2W occupants were killed by truck at rotary.

• 17% M2W occupants were killed by truck.

• Bus has killed 10% pedestrians, while 3% pedestrians were killed by M2W, auto-rickshaw, and truck each.

Other/ Unknown M2W l Auto I Car Bus I Truck

I Single-vehicle

Victim type

Figure 4. Impacting versus Victim type of crashes on major road mid-block segment

In 75% cases M2W occupants (39%) and pedestrians (36%) were victim.

Truck killed 18% M2W occupants, 8% pedestrian and 5% bicyclists.

About 11% crashes observed with car as impacting vehicle to vulnerable road users.

8.2% out of total crashes on major roads were reported as single-vehicle crash of M2W. The fatalities in

such crash generally resulted because of head injury. It was also observed in the city that M2W

occupants not use helmet. Strictly wearing of helmet should be enforced to improve safety of motorized 2-wheeler occupants.

• Fatal crashes of vulnerable road users at signalized intersection show the jumping traffic signals by most of the road users.

• Majority of road users were killed by truck, which shows high presence of heavy vehicles in the city. Restriction should be imposed in terms of time to reduce this number.

• High number of crashes with M2W and pedestrian at rotary shows their risky maneuverings and overtaking behavior.

• Buses and cars have imposed high risk to vulnerable road users followed by truck.

3.1. Actual Risk Determination

The following section explains actual risk to all modes of transportation, followed by perceived risk to all modes, and then comparison between these two risks at different road infrastructures like mid-block, signalized, un-signalized, and rotary intersections. The actual risk to traffic crash has been worked out for six transportation modes prevailing in the case study area, i.e. walk, bicycle, motorized two-wheeler, auto-rickshaw (also include any other 3-wheeler motorized vehicles like tempo), car, and bus.

Exposure calculation of different infrastructure:

Exposure for mid-block is considered for per kilometre road segment length. To make actual risk comparable between mid-blocks and intersections, various exposure levels and equivalent length factors have been considered in the study. The signalized and un-signalized road intersections were considered equivalent to 400 and 300 m mid-block exposure (as most of the signalized intersections are 4-arm intersections and un-signalized intersections are 3-arm intersections), while rotary intersections have treated separately. The exposure of rotary intersections has calculated considering average radius of 10 metre in addition to 400 m length for intersection. Thus, the effective exposure for rotary intersection becomes (400m + 63 m) 463 m. For the same traffic, the exposure levels used for mid-block, signalized, un-signalized, and rotary intersections are 1, 4, 3, and 4 respectively.

Table1. Mode-wise road users killed in fatal crashes on different road infrastructures in six years (2005-2010)

Fatalities of road users

Infrastructure Total

Pedestrian Bicyclist M2W M3W Car Bus

Mid-block 103 29 118 10 14 2 276

Signalized intersection 7 2 15 0 0 0 24

Un-signalised intersection 99 47 125 12 13 2 298

Rotary intersection 10 5 17 0 0 0 32

Total fatalities 219 83 275 22 27 4 630

Table 2. Mode-wise road users killed per year in fatal crashes on different road infrastructures

Infrastructure

Fatalities of road users Pedestrian Bicyclist M2W M3W

Exposure (km)

Mid-block 17.17 4.83 19.67 1.67 2.33 0.33 46.00 250.454

Signalized intersection 1.17 0.33 2.50 0.00 0.00 0.00 4.00 32

Un-signalised intersection 16.50 7.83 20.83 2.00 2.17 0.33 49.67 1551.6

Rotary intersection 1.67 0.83 2.83 0.00 0.00 0.00 5.33 64.792

Total fatalities 36.50 13.83 45.83 3.67 4.50 0.67 105.00 1898.846

In the absence of volume data for the various roads for the case study area, exposure has calculated based on the road length in kilometre as given below:

The actual risk is determined as Relative Risk at different infrastructure by following formula:

Road users killed per year of victim type i on infrastructure j Exposure on infrastructure j Total road users killed per year of victim type i on all infrastructure Total exposure on all infrastructure

where RR j is actual relative fatality risk to road user type i on road infrastructure type j with given exposure (km). Thus, actual relative risk to pedestrian on mid-block is given by

Road users killed per year of victim, type i on infrastructure j

RRij — Total road users killed per year of victim type i on all infrastructure

Total exposure on all infrastructure

Relative Risk to pedestrian on mid — block =

17.167 250.454 36.5

1898.846 = 3.57 per kilometre per year This means that pedestrians at mid-block have 3.57 times higher risk than the risk to pedestrians on all other infrastructures.

Table 3. Risk matrix showing actual risk for different victim types

Infrastructure (j) Relative Risk per kilometre per year to road user type (i) Pedestrian Bicyclist M2W M3W Car Bus

Mid-block 3.57 2.65 3.25 3.45 3.93 3.79

Signalized intersection 1.90 1.43 3.24 0.00 0.00 0.00

Un-signalised intersection 0.55 0.69 0.56 0.67 0.59 0.61

Rotary intersection 1.34 1.77 1.81 0.00 0.00 0.00

3.2. Perceived Risk

To identify perceived risk, 1171 respondents have been asked to rate the accidental risk assuming that they were using each mode of transportation considering traffic in your neighbourhood as per the table given below:

Table 4. Perception of risk for different modes (rating) to be done by respondents

Sr No Select one for all modes_Walk Cycle Motorized two- wheeler Car Bus Auto rickshaw

1 Almost certain not to have an accident

2 Somewhat unlikely

3 50% chance of having an accident

4 Somewhat likely

5 Almost certain to have an accident

The respondent have been asked to rate the probability of accident for other infrastructures; signalized intersection, un-signalized intersection and rotary intersection within one year period. These perceived risk have been rescaled according to scale from 0.01 to 0.99 (Scale - 0.01, 0.25, 0.50, 0.75, 0.99) to represent the probability of accident. The probability score may also not be adequate as a linear scale (Noland 1995).

The perceived risk for every mode for different road infrastructure is represented by the weighted mean values of the perceived risk (rescaled values).

Table 5. Weighted mean values for perceived risk

Perceived risk of crash per year Infrastructure Auto-

Pedestrian Bicyclist M2W rickshaw Car Bus

mid-block 0.732 0.250 0.704 0.257 0.059 0.043

signalized intersection 0.259 0.081 0.592 0.049 0.048 0.069

un-signalised intersection 0.478 0.469 0.253 0.254 0.065 0.051

rotary intersection 0.706 0.497 0.481 0.259 0.065 0.049

The highest crash risk has perceived by the pedestrians at mid-block amongst all modes over all infrastructures.

On mid-block, pedestrians have perceived highest risk (0.732) followed by motorized two-wheelers (0.704). The least crash risk has perceived by the bus users (0.043) and car users (0.059) on mid-block. The bicyclists and auto-rickshaw users have perceived same crash risk (0.250 and 0.257) on mid-block.

3.3. Relation between Actual and Perceived Risk

To establish the correlation between actual risk and perceived risk, spearman's ranking correlation in SPSS has been used. Both the risks have been arranged in order from lower to higher risk. The rank order has been modified for equal risks as a requirement of correlation analysis in SPSS.

Table 6. Ranking of actual and perceived risk to various modes at different infrastructure

Sr. No Road user and infrastructure Actual Risk Rank Perceived Risk Rank

1 Pedestrian at mid-block 22 23

2 Bicycle at mid-block 18 17

3 M2W at mid-block 20 20

4 M3W at mid-block 21 2

5 Car at mid-block 24 9

6 Bus at mid-block 23 10

7 Pedestrian at signalised intersection 17 15.5

8 Bicycle at signalised intersection 14 7.5

9 M2W at signalised intersection 19 3.5

10 M3W at signalised intersection 3.5 24

11 Car at signalised intersection 3.5 11

12 Bus at signalised intersection 3.5 22

13 Pedestrian at un-signalised intersection 7 12

14 Bicycle at un-signalised intersection 12 18

15 M2W at un-signalised intersection 8 13

16 M3W at un-signalised intersection 11 21

17 Car at un-signalised intersection 9 3.5

18 Bus at un-signalised intersection 10 15.5

19 Pedestrian at rotary intersection 13 19

20 Bicycle at rotary intersection 15 5

21 M2W at rotary intersection 16 7.5

22 M3W at rotary intersection 3.5 14

23 Car at rotary intersection 3.5 6

24 Bus at rotary intersection 3.5 1

The spearman's ranking correlation coefficient comes insignificant at 95% confidence interval. That is actual and perceived crash risks are not correlated.

4. Discussion

The proportion of road users killed observed as per their age in the study is 10%, 81%, 8%, and 1% for minor (below 18 years), young and middle-age (18-60), an elder (above 60) age group, and unknown age group crashes respectively. McGwin and Brown (1999) have presented an overview of the characteristics of traffic crashes among young, middle-aged and older drivers. The results suggest that the youngest and the oldest drivers were more likely to be considered at-fault. With respect to crash characteristics, older drivers were less likely to have crashes involving driver fatigue, during the evening and early morning, on curved roads, during adverse weather, involving a single vehicle, and while traveling at high speeds. Conversely, older drivers were over-represented in crashes at intersections and: or involving failure to yield the right of way, unseen objects, and failure to heed stop signs or signals. Crashes occurring while turning and changing lanes were also more common among older

drivers. Alcohol was less likely to be a factor in traffic crashes involving older adults. Synthesizing these results led to the conclusion that the primary problem with the young is risk-taking and lack of skill. The strength of older drivers lies in their aversion to risk, but perceptual problems and difficulty judging and responding to traffic flow often counterbalance this attribute. Similar other study by Massie, Campbell, & Williams (1995) also presented similar results. The result of this study only differs for an elder age group. The involvement of an elder age group is quite low compared to other studies. This may be because of their lower travel exposure of elder people of medium sized Indian urban city compared to other countries. An Average fatalities per million persons is 74 for the urban area of Vadodara city. More than 84% fatalities involved pedestrian, bicycle users and M2W. Thus, vulnerable road users were at the highest risk in road crashes in the city.

The study by Elvik & Bjornskau (2005) observed in their study that the people of study area correctly perceived which modes of transportation are the safest and the least safe. While the perceived and actual risks are different and not correlated in this study.

References

DeJoy, D.M. 1992. An examination of gender differences in traffic accident risk perception. Accident Analysis & Prevention, 24, (3) 237-246 Finn, P. & Bragg, B.W.E. 1986. Perception of the risk of an accident by young and older drivers. Accident Analysis & Prevention, 18, (4) 289-298

Hayakawa, H., Fischbeck, P.S., & Fischhoff, B. 2000. Traffic accident statistics and risk perceptions in Japan and the United States. Accident Analysis & Prevention, 32, (6) 827-835

Massie, D.L., Campbell, K.L., & Williams, A.F. 1995. Traffic Accident involvement rates by driver age and gender. Accident Analysis & Prevention, 27, (1) 73-87

McGwin, J. & Brown, D.B. 1999. Characteristics of traffic crashes among young, middle-aged, and older drivers. Accident Analysis & Prevention, 31, (3) 181-198

Klobucar, M. & Fricker, J. 2007. Network Evaluation Tool to Improve Real and Perceived Bicycle Safety. Transportation Research Record: Journal of the Transportation Research Board, 2031, (-1) 25-33

Parkin, J., Wardman, M., & Page, M. 2007. Models of perceived cycling risk and route acceptability. Accident Analysis & Prevention, 39, (2) 364-371

Cho, G., Rodriguez, D.A., & Khattak, A.J. 2009. The role of the built environment in explaining relationships between perceived and actual pedestrian and bicyclist safety. Accident Analysis & Prevention, 41, (4) 692-702

Noland, R.B. 1995. Perceived risk and modal choice: Risk compensation in transportation systems. Accident Analysis & Prevention, 27, (4) 503-521

Elvik, R. & Bjornskau, T. 2005. How accurately does the public perceive differences in transport risks?: An exploratory analysis of scales representing perceived risk. Accident Analysis & Prevention, 37, (6) 1005-1011