DISTINGUISHING CHARACTERISTICS OF YOUNGER AND OLDER DRIVER CRASH FATALITIES IN TEXAS

Transcription

1 DISTINGUISHING CHARACTERISTICS OF YOUNGER AND OLDER DRIVER CRASH FATALITIES IN TEXAS V. J. Pezoldt Texas Transportation Institute Texas A&M University System This report was prepared in cooperation with The Texas Department of Transportation and The National Highway Traffic Safety Administration February 3 THE TEXAS A & M UNIVERSITY SYSTEM COLLEGE STATION, TEXAS

2 ACKNOWLEDGEMENTS The author wishes to acknowledge the contributions of Dr. Michael Manser, formerly of the Center for Transportation Safety (CTS) at TTI, who initiated the project upon which this report is based. Thanks also to Mr. Danny Morris, TTI, for his considerable contribution in developing the SAS programs used for extracting and compiling the data reported here from five years of Texas Accident Data files and to Dr. Lindsay Griffin, CTS, from whose prior work the relative likelihood index used here was adopted and who provided statistical consultation. While their assistance is greatly appreciated any errors in calculation or interpretation are solely the responsibility of the author. DISCLAIMER The conclusions and opinions expressed in this document are those of the author, and do not necessarily represent those of the State of Texas, the Texas Department of Transportation (TxDOT) or any political subdivision of the State or Federal government. i

3 TABLE OF CONTENTS PAGE List of Tables... iii List of Figures... iv Introduction...1 Variables Examined...3 Measures Employed, Analysis Methods and Data Presentation...5 Data Summaries and Discussion...1 References...4 ii

4 LIST OF TABLES PAGE Table 1 Percent of Texas licensed drivers and driver fatalities by age group....3 Table 2.Crash variables (in bold) and values of the variables evaluated (bullets)...4 iii

5 LIST OF FIGURES PAGE Figure 1. Example of part (a) of the data presentations...6 Figure 2. Example of part (b) of the data presentations...7 Figure 3. Example of part (c) of the data presentations...9 Figure 4. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DAY OF WEEK within three age groups...12 Figure 5. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DAY OF MONTH within three age groups...13 Figure 6. Number (a), percent (b) and relative likelihood (c) of driver fatalities by TIME OF DAY within three age groups...14 Figure 7. Number (a), percent (b) and relative likelihood (c) of driver fatalities by LIGHT CONDITION within three age groups...15 Figure 8. Number (a), percent (b) and relative likelihood (c) of driver fatalities by WEATHER within three age groups...16 Figure 9. Number (a), percent (b) and relative likelihood (c) of driver fatalities by SURFACE CONDITION within three age groups...17 Figure 1. Number (a), percent (b) and relative likelihood (c) of driver fatalities by NUMBER OF PERSONS IN VEHICLE (in addition to driver) within three age groups...18 Figure 11. Number (a), percent (b) and relative likelihood (c) of driver fatalities by TOTAL NUMBER OF VEHICLES INVOLVED within three age groups...19 Figure 12. Number (a), percent (b) and relative likelihood (c) of driver fatalities by ROAD CLASS within three age groups... Figure 13. Number (a), percent (b) and relative likelihood (c) of driver fatalities by POPULATION GROUP within three age groups...21 Figure 14. Number (a), percent (b) and relative likelihood (c) of driver fatalities by INTERSECTION RELATED within three age groups...22 Figure 15. Number (a), percent (b) and relative likelihood (c) of driver fatalities by ROADWAY ALIGNMENT within three age groups...23 Figure 16. Number (a), percent (b) and relative likelihood (c) of driver fatalities by PART OF ROADWAY within three age groups...24 Figure 17. Number (a), percent (b) and relative likelihood (c) of driver fatalities by TRAFFIC CONTROL within three age groups...25 Figure 18. Number (a), percent (b) and relative likelihood (c) of driver fatalities by RESTRAINING DEVICE within three age groups...26 Figure 19. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DRIVER LICENSE STATUS within three age groups...27 Figure. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DRIVER DEFECT within three age groups...28 iv

6 PAGE Figure 21. Number (a), percent (b) and relative likelihood (c) of driver fatalities by CITATION within three age groups...29 Figure 22. Number (a), percent (b) and relative likelihood (c) of driver fatalities by LIABILITY INSURANCE STATUS within three age groups...3 Figure 23. Number (a), percent (b) and relative likelihood (c) of driver fatalities by ALCOHOL/DRUG TEST RESULT within three age groups...31 Figure 24. Number (a), percent (b) and relative likelihood (c) of driver fatalities by VEHICLE MOVEMENT/MANNER OF COLLISION within three age groups...32 Figure 25. Number (a), percent (b) and relative likelihood (c) of driver fatalities by FIRST HARMFUL EVENT within three age groups...33 Figure 26. Number (a), percent (b) and relative likelihood (c) of driver fatalities by CONTRIBUTING FACTOR 1 within three age groups...34 Figure 27. Number (a), percent (b) and relative likelihood (c) of driver fatalities by CONTRIBUTING FACTOR 2 within three age groups...35 Figure 28. Relative likelihood (RL) of younger driver fatalities for all crash data variables with RLs significantly different from 1. (p<.5)...36 Figure 29. Relative likelihood (RL) of older driver fatalities for all crash data variables with RLs significantly different from 1. (p<.5)...37 Figure 3. Relative likelihood (RL) of younger driver fatalities for all crash data variables with significant RLs <.5 or > Figure 31. Relative likelihood (RL) of older driver fatalities for all crash data variables with significant RLs <.5 or > v

7 DISTINGUISHING CHARACTERISTICS OF YOUNGER AND OLDER DRIVER CRASH FATALITIES IN TEXAS INTRODUCTION What differentiates crashes in which older and younger drivers die from fatal crashes in which drivers not typically assigned to either of those somewhat arbitrary groups die? How are they the same? To what extent does Texas crash experience, as reflected in the police accident reports that comprise the Texas Accident File, confirm or refute commonly held assumptions about the circumstances in which older and younger drivers find themselves that are more or less likely to result in fatal crashes? These are a few of the questions that prompted the compilation and review of the five years of Texas fatal crash experience summarized in this report. Multiple (5) years of data are pooled here both to increase the sample size available for examining crash variables that may not have high frequencies of occurrence in any one year and to lessen the influence of conditions which may be specific only to a particular year, e.g., extreme weather conditions or the introduction of new legislation or enforcement efforts that could skew crash data for a relatively short period. For purposes of this analysis, younger drivers are defined as all drivers from 14 1 through years of age who died in a motor vehicle crash during the five-year period. Older drivers are all drivers over 65 years of age who died in crashes during the study years. There is no single accepted definition for either of these sub-populations of drivers. It has been argued that the lower bound of older as used here is too low; see, for example, Skinner and Stearns (1999). They suggest that although 65 is often recognized as the beginning point of the aging population,... the risks associated with motor vehicle operation increase exponentially at age 75 and older (p4). We have elected to include the lower bound (65) in the older category to increase the number of crashes available with full recognition that certain crash variables may be appear less dominate as a proportion of older driver crashes. Similarly, the upper bound of younger drivers as used here may be considered too low. This cut off point was employed primarily because it is consistent with other definitions, as, for example, the age at which alcohol consumption is legal in Texas. Recognition of how younger and older driver fatal crashes are different from each other and from those involving other drivers (21-64 years old) may assist traffic safety practitioners in the development, selection, and targeting of crash countermeasures for specific groups and serve to engender consideration of new approaches to addressing the unique characteristics of different driver populations. Interest in the relationship between traffic crashes and driver age is certainly not a recent phenomenon. Treat et al (1977) in the classic Indiana Tri-level study of accident causation, for 1 Although the minimum age for driving legally in Texas is 15, fatally injured 14 year old drivers were included in the data reviewed because of an initial plan to use data from hard-copy accident reports for, among other purposes, verifying driver ages recorded in the computerized data file. Data from the computerized accident files were already accessed before the unavailability of sufficient hard-copies was ascertained. The year-olds included in the data set constitute 1% of the younger driver crashes examined.

8 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 2 example, found drivers over the age of 65 to be twice as likely to have been involved in crashes due to improper lookout than were accident-involved drivers in general. Shinar (1978) attributes this largely to the reduced field of vision common among older drivers. Thus their improper lookout may be the result of a failure to see rather than a failure to look. Based on a reanalysis of the accidents investigated in-depth in the Tri-level study, Shinar et al (1978) also report inexperienced drivers (typically, though not always, younger drivers), were almost seven times as likely to cause an accident because of inadequate directional control of the vehicle than were drivers not noted as inexperienced. Aging can be viewed as a progressive and variable combination of ontogenetic, historical and life events, resulting in a non-uniform set of behavioral and physiological changes with age. [Baltes, Reese and Lipsitt (198), cited in Smith, Meshkati and Robertson (1993, p453)]. As Smith et al and many others have noted, the effects of aging are highly variable, both within any age group and within individuals. A substantial literature is available that addresses ageassociated characteristics relevant to driving at both ends of the age spectrum. Evans (1991, p41), remarks that the over involvement of young, especially male, road users is one of the most consistently observed phenomena in traffic throughout the world; its prevalence and magnitude suggesting that it must involve much more than a mere lack of driving experience. In addition to Smith et al, cited previously, Triggs and Smith (1996), Staplin et al (1997), Staplin, Lococo and Byington (1998) and Dewar and Olson (2), among others, summarize the characteristics of young and/or old drivers that are associated with increased crash risk or increased severity of injury given a crash. Attributes and characteristics associated with young and old drivers or that are attributed to the aging process encompass physical, sensory, cognitive, and emotional domains. Among older drivers, physical changes include reduction in the range of motion and joint flexibility, strength and force capabilities, and an increased incidence of a vast variety of medical conditions that may negatively affect driving acutely or chronically. As we age, sensory and perceptual changes are evidenced by increased glare sensitivity, decreased static and dynamic visual acuity, and contrast sensitivity, color and visual field decrements. Cognitive changes can include slowing of information processing and response times and memory deficits. Voluntary or socially imposed modifications in driving adopted by many older drivers (e.g., reduction in night driving, total miles driven, long trips, rush hour and bad weather driving) serve to offset some of the problems encountered by older drivers, as perhaps most importantly, do the significant benefits of experience enjoyed by many older drivers. Younger drivers may well be near their peak in terms of physical and sensory capabilities, but the interaction of inexperience, less than fully matured judgment, and their often-cited elevated risk-taking and feelings of invulnerability contribute to the overrepresentation of young drivers in crash statistics. The inexperience of younger drivers may be reflected both in a lack of basic driving skill and in deficiencies in recognizing hazards and the appropriate responses to them. The data compiled and reviewed here are offered to provide an indication of the extent to which the known differences among drivers in different age groups translate to readily discernible differences in the characteristics of crashes as reported in the Texas accident files.

9 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 3 Though we can make rough estimates of the number of older and younger drivers in Texas, a significant limitation of using the accident database for comparing the crash experience of different sub-populations of drivers is the paucity of exposure data. While many aspects of crashes involving populations of interest can be well characterized from crash data, such data do not readily allow comparisons to be made on the bases of miles driven or exposure to particular environmental, traffic or roadway conditions or specific road types. From 1995 through 1999, a total of 1,62 drivers over the age of 14 died in crashes in Texas. Of these, 7,216 were between 21 and 64 years of age, 1,374 were 14 through years old and 1,472 were 65 or older. Table 1 provides these data as the percentage of drivers who died in crashes in Texas in each of the five years by age group. Also shown are the percentages of licensed drivers in the state for each of those years by age group. Table 1. Percent of Texas licensed drivers and driver fatalities by age group. Percent of Driver Crash Fatalities Percent of TX Licensed Drivers Year 5-Yr 5-Yr Age Avg. Avg. < Recognizing that neither all drivers who die in Texas crashes are Texas residents nor are they all licensed to drive in any state, the data in Table 1 are consistent with the view that younger and older drivers may be over represented among crash fatalities. Over the five year period driver crash fatalities among those less than 21 years old accounted for 13.6 % of all driver fatalities, but only 7.6% of the licensed drivers. Older drivers, too, though to a lesser extent, are represented in a higher proportion of the fatalities than their proportion in the licensed population would suggest (14.6 vs. 11.9%). Despite the lack of good exposure data, it is possible to examine the relative incidence of crash characteristics among different groups. As is described below, this is accomplished here through the use of a relative likelihood index to asses the frequency of crash characteristics associated with younger and older driver crashes relative to the frequency of the same characteristics in crashes in which year old drivers died. VARIABLES EXAMINED Each of the 24 variables selected from the Texas Accident File for review are listed in Table 2. Each year, all accident reports completed by peace officers at crash scenes in all jurisdictions in the state are filed with the Department of Public Safety and subsequently coded into three primary files that incorporate data pertinent to the accident, driver and vehicle, and occupants ; the A, B, and C records, respectively. The variables indicated in Table 2 were selected to provide information that would assist in characterizing crashes in a manner that would allow determination of how older and younger fatal crashes are distinguishable from those

10 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 4 Table 2. Crash variables (in bold) and values of the variables evaluated (bullets) Day of Week Sunday - Saturday Month of Year January - December Time of Day 4 AM-8 AM 8 AM-Noon Noon-4 PM 4 PM-8 PM 8 PM-Midnight Light Condition Daylight Dawn Dark - Not Lighted Dark - Lighted Dusk Weather Condition Clear (Cloudy) Raining Fog Sleet or Snow All Other Surface Condition Dry Wet Muddy Snowy/Icy Total in Vehicle - in addition to driver 1 2 >2 Total Number of Vehicles Involved 1 2 >2 Road Class Interstate US & State FM County Rd City St All Other Population Group Rural Towns <5, population Towns 5-25, Cities 25-1, Cities >1, Intersection Related Intersection Intersection related Driveway access Non-intersection Roadway Alignment Straight Curve Part of Roadway On Roadway On Shoulder Beyond Shoulder Traffic Control Center stripe or divider Stop & go signal Stop sign None shown Yield sign No passing zone Flashing red light Other traffic control Warning sign Turn marks Officer, flagman RR gates or signal Driver Restraining Device No restraint Belt, shoulder or both Air bag deployed Unknown or Not Applicable Driver License Status Licensed Unlicensed Unknown Driver Defect None Fatigued/Asleep Ill Eyesight defect Other phys defect Other handicap Mentally defective Hearing defect Limbs missing Citation Citation No citation Unknown Liability Insurance Status Yes No Not shown Driver Alcohol/Drug Test Results No test/refused Result unknown BAC.1+ BAC<.1 Drug test Negative Drug test Pos Not released Contaminated Vehicle Movements/Manner of Collision Single Vehicle 2 Vehicles - approaching at angle 2 Vehicles - going same direction 2 Vehicles - going opposite directions 2 Vehicles - other First Harmful Event Another MV in Transit Fixed Object Overturned RR Train Parked Car Animal Other Non-collision Other object Pedestrian Contributing Factor (1) None applies Speed (limit & unsafe) Fail to yield ROW Wrong side, not passing Disregard stop sign/light & signal All other specified factors Contributing Factor (2) None applies Under influence of alcohol or drugs Other unspecified factor Illegal/unsafe passing Follow too closely Other specified factors

11 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 5 crashes in which drivers in the year old range were involved. The values associated with each of the variables are also indicated in the table. For many of the variables, all possible values the variable can take are noted. For example, the variable Surface Condition, referring to the condition of the roadway on which the crash occurred, can take one and only one of the values indicated, i.e., dry, wet, muddy, or snowy/icy. For some variables, however, values were combined as appropriate to reduce the number of possible values to a useable magnitude and provide more general categories. Thus, for example, although Total Number of Vehicles Involved reports the actual number of vehicles involved in each crash, it has been reduced to 1, 2 or more than 2 vehicles. While the majority of multi-vehicle fatal crashes involve two vehicles and a substantial number involve three, higher numbers of vehicles were involved in some crashes up to 18 in one instance. Similarly, the variable Vehicle movements/manner of collision can take one of more than 4 values, most of which are variations on the movement of two vehicles. There are, for example, ten classifications each of 2 Vehicles - approaching at angle, 2 Vehicles - going same direction, and 2 Vehicles - going opposite directions. These have been combined to the higher order value. Several additional variables were initially considered for inclusion in the final data set but were ultimately excluded either because they did not provide additional information for enough crashes or the information they did provide was judged irrelevant to the driver-age issues of interest here. MEASURES EMPLOYED, ANALYSIS METHODS and DATA PRESENTATION The basic data for the information provided in this report originates in accident reports for all crashes in which drivers died in Texas from 1995 through Each crash has been assigned to one of three groups based on the driver s age at the time of the crash: 14- years old, hereafter referred to as younger drivers, years old, and 65 years old and older, identified here as older drivers. For each variable, three measures of crash involvement, magnitude, proportion and relative likelihood, are summarized as a function of driver age: a. Magnitude: The absolute number of drivers within each age group that died during the five year period in crashes with the characteristics specified in the accident report. b. Proportion: The proportion all driver fatalities within each age group that exhibited the crash characteristics identified, and c. Relative Likelihood: The likelihood of younger and older driver fatal crashes exhibiting the selected crash characteristics relative to the likelihood that crashes in which year old drivers were fatally injured exhibit those characteristics. Each of the three measures is applied to all values of each of the 24 variables. They are summarized in a series of 24 three-part figures, one for each of the crash variables selected from the Texas Accident Files.

12 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 6 A description and explanation of each component of the three-part figures, using one of the variables as an exemplar, is provided here to facilitate understanding and interpretation of the data presentations. (a) Magnitude The first part of each figure, identified as (a), an example of which is shown in Figure 1, is very straightforward. It provides a simple table indicating the total number of driver fatalities from within each of the three age groups, and for all drivers combined, for each of the values reported for the crash data variable of interest. In the example, the variable Light condition can take one of five mutually exclusive values: Daylight Dawn Dark Not Lighted Dark Lighted, or Dusk. As seen in the table, during the five-year period from 1995 through 1999, 598 drivers 14- years old died in daylight crashes compared to 3,128 drivers from years old and 1,9 drivers 65 years old and older. This portion of each figure provides an indication of the absolute magnitude of the given level of each variable as a factor in driver fatalities for each age group.. (a) Driver Fatalities by Light Condition Light Condition ALL Daylight 4, ,128 1,9 Dawn Dark - Not Lighted 3, , Dark - Lighted 1, , Dusk Figure 1. Example of part (a) of the data presentations Direct comparison of the number of drivers in each age group, of course, provides only limited information since it does not account for either the large differences in the total number of drivers in each age group, the total number of fatally injured drivers in each group, or the exposure of any of the age groups to the various light conditions. These data are useful, however, in suggesting answers to questions of the type: How commonly are older drivers killed in crashes on dark, unlighted roadways? (b) Proportion The second part of each of the data figures (b) (see Figure 2 for an example) is derived from (a) and provides additional information that assists in comparing the three age groups. As shown in the example below, the proportion of all deceased drivers within each age group who died in crashes classified by each value of the variable of interest is provided graphically, expressed as a percent of all fatally injured drivers who died in crashes exhibiting that level of the crash variable. Computationally, this measure is simply defined as:

13 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 7 Proportion = a /( a + b) where: a = the number of drivers in each age group (younger, older, 21-64) killed in crashes exhibiting the value of the crash variable of interest. b = the number of drivers in each age group (younger, older, 21-64) killed in crashes not exhibiting the value of the crash variable of interest. As is evident, the expression (a + b) equals the total number of driver fatalities in an age group. The total for each age group equals 1% (within rounding error). Thus, for example, among younger drivers, 44% died in daylight crashes, 37% in unlighted dark crashes, 16% in lighted dark crashes, and 2 and 1% in crashes at dawn and dusk respectively. For this variable, Light Condition, it is clear from part (b) that older drivers who were killed in crashes during the period from were much more likely to have been killed in crashes under daylight conditions than were either younger drivers or drivers ranging in age from years. (b) Percent of Driver Fatalities by Light Condition Within Driver Age Group 1 Percent Driver Age Daylight Dawn Dark - Not Lighted Dark - Lighted Dusk 1 Figure 2. Example of part (b) of the data presentations (c) Relative Likelihood While part (b) of each figure provides a means for comparing the three age groups proportionally, it does not offer a convenient assessment of the relative occurrence of crash experience. The final portion of each figure (c) is offered both to provide such comparisons and to indicate how confident we can be that the observed differences are likely to represent true differences. Part (c) casts the crash data in terms of the relative likelihood that each value of the crash variable of interest was associated with younger and older driver fatal crashes more (or less) than with drivers in the year old group. Relative likelihood is nothing more than the ratio of two proportions, i.e., the proportion of crashes in which an older (or younger) driver died that exhibits a particular characteristic, divided by the proportion crashes in which a year old driver died that exhibits the same characteristic. Computationally, relative likelihood (RL) is defined as:

14 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 8 a /( a + b) RL = c /( c + d) where: a = the number of older (or younger) drivers killed in crashes exhibiting the value of the crash variable of interest. b = the number of older (or younger) drivers killed in crashes not exhibiting the value of the crash variable of interest. c = the number of year old drivers killed in crashes exhibiting the value of the crash variable of interest. d = the number of year old drivers killed in crashes not exhibiting the value of the crash variable of interest. To illustrate, again using the Light Condition data shown in Figure 1 (a), we can calculate the likelihood of older driver fatalities occurring under daylight conditions relative to the likelihood of year-old driver fatalities in daylight as: 19 /( ) RL = = /( ) A RL of 1.89 indicates that crashes in which older drivers died were 1.89 times more likely to have occurred during daylight than were the crashes in which year old drivers died. Similarly, again for older drivers, the relative likelihood of crashes during dawn light conditions is: /( ) RL = = /( ) Therefore, in this case, it can be seen that the crashes in which older drivers died were.68 times - about 7% - as likely to have occurred during dawn light than the crashes in which year old drivers died. The two examples of RL above illustrate the general case for interpreting Relative Likelihoods. If RL equals 1, the two age groups being compared (i.e., crashes of older or younger vs year-old drivers) can be said to be equally likely to exhibit the same condition or characteristic. Values of RL ranging from to 1 represent conditions that are less likely for the older or younger drivers than for year-olds. Values of RL greater than 1 (ranging from 1 to ) correspond to conditions in which the older or younger drivers exceed the year-olds. The magnitude (a), proportion (b), and relative likelihood (c) statistics all pertain to the specific data set examined. As such, they are useful in describing and characterizing the crashes of older and younger drivers that occurred from They do not, however, allow us to consider in any quantitative way how confident we should be that the same results would be observed in an other data set; for example comparable data from a different year. Therefore, in order to provide a measure of generalizability to the present results, 95% confidence intervals (95% CI) have been constructed around each RL reported to test their statistical significance. For a detailed explanation of the calculations used to construct the CIs and additional discussion on the

15 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 9 use and computation of relative likelihoods, see Griffin and DeLaZerda () and Sahai and Khurshi (1996). The 95%CIs for the RLs used in the examples above, 1.89 and.68, are and , respectively. In practical terms, this means that we can be 95% certain that the true value for RL=1.89 lies between 1.83 and 1.96 and is, therefore, statistically, significantly greater than 1 at p<.5. When the CI for any RL includes 1, we cannot be confident that the true value is different from 1 and therefore the RL cannot be said to be statistically significant. This holds without regard to the numerical size of the RL. Thus, the RL of.68 in the second example above cannot be said to differ significantly from 1. The results of all of the calculations of RL and their associated 95% CI, are shown graphically in part (c) of each figure for each value of the variable in question for older (O) and younger (Y) drivers. The examples cited above are illustrated in Figure 3. Shown in this example is the likelihood of Younger and Older drivers dying in crashes for each of the values reported in the crash data file variable Light condition relative to drivers years old dying in such crashes. The RL designated by the vertical bar ( ) indicates that older drivers were nearly twice (1.89) as likely as year old drivers to have died in crashes under daylight conditions. The 95 percent confidence interval, shown by the horizontal line around each RL, indicates if the value is significantly different from the likelihood of year-old drivers deaths under the same conditions. This statistical significance is also indicated on the figure by an asterisk () adjacent to the vale of the variable. Note that the is presented in the figures on a log scale, centered on the equally likely balance point of 1. (c) Younger (Y) and Older (O) Driver Fatalities by Light Condition Daylight: Y Dawn: Y Dark - Not Lighted: Y Dark - Lighted: Y Dusk: Y Daylight: O Dawn: O Dark - Not Lighted: O Dark - Lighted: O Dusk: O RL=1.89 CI = RL=.68 CI = Figure 3. Example of part (c) of the data presentations.

16 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 1 DATA SUMMARIES and DISCUSSION This section includes the 24 figures summarizing the magnitude, proportion and relative likelihood of younger and older driver fatalities in comparison to driver fatalities in the year-old group and a brief discussion of the data therein. Though statistical significance of the relative likelihoods shown in Figures 4-27 is indicated, the figures include all RLs calculated without regard to significance. Following the individual summaries for each of the crash data variables are four additional figures two each for younger and older driver crash fatalities. Figures 28 and 29, for younger and older drivers respectively, provide, in a single view, only those RLs for each of the values of the variables that reached statistical significance. Statistical significance can be viewed as a necessary, but not sufficient indicator of practical importance. With large enough sample sizes, virtually any magnitude of difference will meet the test of significance. Magnitudes of difference that are important depend on the use to which the data are put and on the judgment and knowledge of the user. The final two figures (Figures 3 and 31), again for younger and older drivers, respectively, are offered to point out those variables, which stand out among the others. They include only those variables for which the relative likelihoods are both statistically significant and which are either less than.5 or greater than 2.. The only variables included in these final figures, therefore, are those that indicate characteristics of crashes in which older or younger drivers died that are less than 5% or more than twice as frequent than in crashes in which year old drivers died. Younger Driver Fatal Crashes Inspection of the data summaries reveal the most evident 2 differences in crash characteristics between crashes in which young drivers died and those which were fatal to year-old drivers include roadway surface condition, the number of persons in the vehicle in addition to the driver, the drivers license status and reported driver defects. Specifically, younger drivers who died in crashes were less likely than others to have been in fatal crashes in snowy/icy conditions (RL =.33) and were much less likely to have been reported as ill at the time of the crash (RL =.12). Younger drivers were significantly more likely to have died in crashes with at least two other persons in the vehicle with them, almost exactly twice as likely as year-olds. Younger drivers were also nearly three times as likely to have been unlicensed (RL = 2.9) at the time of the crash than year-olds. Older Driver Fatal Crashes The Texas Accident Files reveal more differences in the characteristics of crashes in which older drivers died relative to year old victims than are observed for younger drivers. As noted, all significantly different RLs are shown in Figure 29 and those significant RLs, which indicate 2 most evident refers here to meeting the criteria for inclusion in Figure 3, i.e. RL is significant and <.5 or >2..

17 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 11 crash characteristics less than.5 or more than twice as frequently relative to year-olds are provided in Figure 31. Among the most striking differences revealed in Figure 31, relative to year-olds, older drivers who die in crashes are: Nearly two and a half (2.4) times more likely to have disregarded a stop sign/light or signal. Almost six times (5.9) more likely to have failed to yield right of way. About three times (2.9) as likely to have died in crashes in which two vehicles were approaching at an angle. Consistent with what is known about the driving habits of many older drivers, relative to their year old counterparts, older drivers were also: More than twice as likely to have been killed in crashes in the morning and early afternoon (8am noon, RL = 2.3, noon 4 PM, RL = 2.4). Much less likely to been in crashes between 8 PM and midnight (RL =.32) or midnight and 4 AM (RL =.15). Less likely to have died in crashes in inclement weather (fog or sleet and snow). Finally, older drivers were reported to have exhibited some driver defect much more frequently than year-olds: Sixteen times more frequently noted as having an eyesight defect. Twelve times more likely to have had some other physical defect reported. While true for all of the crash characteristics reported in these data, but especially for reports involving issues such as driver defects, one should exercise caution in attributing the cause of the fatal crash solely to any one of the characteristics identified. None-the-less, judicious use of these data can provide clues or in some cases confirmation to the relationship between crash characteristics and what is known about older and younger drivers. Increased understanding of these relationships will, it is hoped, assist traffic safety professionals and others in the development, selection, and targeting of crash countermeasures for younger and older drivers..

18 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 12 (a) Driver Fatalities by Day of Week Day of Week ALL Sun 1, ,19 18 Mon 1, Tue 1, Wed 1, Thu 1, Fri 1, , Sat 1, ,459 4 (b) Percent of Driver Fatalities by Day of Week Within Driver Age Group 25 Percent Driver Age Sun Mon Tue Wed Thu Fri Sat (c) Younger (Y) and Older (O) Driver Fatalities by Day of Week Mon: Y Tue: Y Wed: Y Thu: Y Fri: Y Sat: Y Sun: Y Mon: O Tue: O Wed: O Thu: O Fri: O Sat: O Sun: O Figure 4. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DAY OF WEEK within three age groups.

19 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 13 (a) Driver Fatalities by Month of Year Month ALL January February March April May June July August September October November December (b) Percent of Driver Fatalities by Month of Year Within Driver Age Group 14 Percent Driver Age Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec (c) Younger (Y) and Older (O) Driver Fatalities by Month of Year Jan: Y Feb: Y Mar: Y Apr: Y May: Y Jun: Y Jul: Y Aug: Y Sep: Y Oct: Y Nov: Y Dec: Y Jan: O Feb: O Mar: O Apr: O May: O Jun: O Jul: O Aug: O Sep: O Oct: O Nov: O Dec: O Figure 5. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DAY OF MONTH within three age groups

20 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 14 (a) Driver Fatalities by Time of Day Time ALL Midnight-4AM 1, , AM-8AM 1, AM-Noon 1, Noon-4PM 1, , PM-8PM 2, , PM-Midnight 1, , (b) Percent of Driver Fatalities bytime of Day Within Driver Age Group 5 Percent Driver Age Midnight-4AM 4AM-8AM 8AM-Noon Noon-4PM 4PM-8PM 8PM-Midnight (c) Younger (Y) and Older (O) Driver Fatalities by Time of Day Midnight-4AM: Y 4AM-8AM: Y 8AM-Noon: Y Noon-4PM: Y 4PM-8PM: Y 8PM-Midnight: Y Midnight-4AM: O 4AM-8AM: O 8AM-Noon: O Noon-4PM: O 4PM-8PM: O 8PM-Midnight: O Figure 6. Number (a), percent (b) and relative likelihood (c) of driver fatalities by TIME OF DAY within three age groups.

21 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 15 (a) Driver Fatalities by Light Condition Light Condition ALL Daylight 4, ,128 1,9 Dawn Dark - Not Lighted 3, , Dark - Lighted 1, , Dusk (b) Percent of Driver Fatalities by Light Condition Within Driver Age Group 1 Percent Driver Age Daylight Dawn Dark - Not Lighted Dark - Lighted Dusk 1 (c) Younger (Y) and Older (O) Driver Fatalities by Light Condition Daylight: Y Dawn: Y Dark - Not Lighted: Y Dark - Lighted: Y Dusk: Y Daylight: O Dawn: O Dark - Not Lighted: O Dark - Lighted: O Dusk: O Figure 7. Number (a), percent (b) and relative likelihood (c) of driver fatalities by LIGHT CONDITION within three age groups.

22 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 16 (a) Percent of Driver Fatalities by Weather Weather ALL Clear (Cloudy) 9,82 1,26 6,473 1,349 Raining Fog Sleet or Snow All Other (b) Percent of Driver Fatalities by Weather Within Driver Age Group 1 Percent Driver Age Clear (Cloudy) Raining Fog Sleet or Snow All Other (c) Younger (Y) and Older (O) Driver Fatalities by Weather Clear (Cloudy): Y Raining: Y Fog: Y Sleet or Snow: Y All Other: Y Clear (Cloudy): O Raining: O Fog: O Sleet or Snow: O All Other: O Figure 8. Number (a), percent (b) and relative likelihood (c) of driver fatalities by WEATHER within three age groups.

23 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 17 (a) Percent of Driver Fatalities by Surface Condition Surface Condition ALL Dry 8,89 1,218 6,298 1,293 Wet 1, Muddy Snowy/Icy (b) Percent of Driver Fatalities by Surface Condition Within Driver Age Group 1 Percent Driver Age Dry Wet Muddy Snowy/Icy (c) Younger (Y) and Older (O) Driver Fatalities by Surface Condition Dry: Y Wet: Y Muddy: Y Snowy/Icy: Y Dry: O Wet: O Muddy: O Snowy/Icy: O Figure 9. Number (a), percent (b) and relative likelihood (c) of driver fatalities by SURFACE CONDITION within three age groups.

24 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 18 (a) Driver Fatalities by Total in Vehicle (in addition to driver) Total in Vehicle ALL , ,235 1, , , > (b) Percent of Driver Fatalities by Number of Persons in Vehicle (In addition to driver) Within Driver Age Group 1 Percent Driver Age >2 (c) Younger (Y) and Older (O) Driver Fatalities by Number of Persons in Vehicle (In addition to driver) : Y 1: Y 2: Y >2: Y : O 1: O 2: O >2: O Figure 1. Number (a), percent (b) and relative likelihood (c) of driver fatalities by NUMBER OF PERSONS IN VEHICLE (in addition to driver) within three age groups.

25 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 19 (a) Driver Fatalities by Total Number of Vehicles Involved Number of Vehicles ALL , , , , > (b) Percent of Driver Fatalities by Total Number of Vehicles Involved Within Driver Age Group 8 Percent Driver Age >2 (c) Younger (Y) and Older (O) Driver Fatalities by Total Number of Vehicles Involved 1: Y 2: Y >2: Y 1:O 2: O >2: O Figure 11. Number (a), percent (b) and relative likelihood (c) of driver fatalities by TOTAL NUMBER OF VEHICLES INVOLVED within three age groups.

26 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas (a) Driver Fatalities by Road Class Road Class ALL Interstate 1, , US & State 4, , FM 2, , County Rd City St 1, All Other (b) Percent of Driver Fatalities by Road Class Within Driver Age Group 6 Percent 55 Driver Age Interstate US & State FM County Rd City St All Other (c) Younger (Y) and Older (O) Driver Fatalities by Road Class Interstate: Y US & State: Y FM: Y County Rd: Y City St: Y All Other: Y Interstate: O US & State: O FM: O County Rd: O City St: O All Other: O Figure 12. Number (a), percent (b) and relative likelihood (c) of driver fatalities by ROAD CLASS within three age groups.

27 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 21 (a) Driver Fatalities by Population Group Population Group ALL Rural 5, , Towns <5, Towns 5-25, Cities 25-1, Cities >1, 2, , (b) Percent of Driver Fatalities by Population Group WIthin Driver Age Group 7 Percent 62 Driver Age Rural Towns <5, Towns 5-25, Cities 25-1, Cities >1, (c) Younger (Y) and Older (O) Driver Fatalities by Population Group Rural: Y Towns <5,: Y Towns 5-25,: Y Cities 25-1,: Y Cities >1,: Y Rural: O Towns <5,: O Towns 5-25,: O Cities 25-1,: O Cities >1,: O Figure 13. Number (a), percent (b) and relative likelihood (c) of driver fatalities by POPULATION GROUP within three age groups.

28 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 22 (a) Driver Fatalities by Intersection Related Intersection Related ALL Intersection 1, Intersection related Driveway access Non-intersection 7,115 1,57 5,381 1,57 Total Driver Fatalities 1,62 1,374 7,216 1,374 (b) Percent of Driver Fatalities by Intersection Related WIthin Driver Age Group 1 Percent Driver Age Intersection Intersection related Driveway access Non-intersection (c) Younger (Y) and Older (O) Driver Fatalities by Intersection Related Intersection: Y Intersection related: Y Driveway access: Y Non-intersection: Y Intersection: O Intersection related: O Driveway access: O Non-intersection: O Figure 14. Number (a), percent (b) and relative likelihood (c) of driver fatalities by INTERSECTION RELATED within three age groups.

29 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 23 (a) Driver Fatalities by Roadway Alignment Alignment ALL Straight 8,158 1,74 5,744 1,34 Curve 1,94 3 1, (b) Percent of Driver Fatalities by Roadway Alignment Within Driver Age Group 1 Percent Driver Age Straight Curve (c) Younger (Y) and Older (O) Driver Fatalities by Roadway Alignment Straight: Y Curve: Y Straight: O Curve: O Figure 15. Number (a), percent (b) and relative likelihood (c) of driver fatalities by ROADWAY ALIGNMENT within three age groups.

30 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 24 (a) Driver Fatalities by Part of Roadway Part of Roadway ALL On Roadway 5, ,746 1,131 On Shoulder Beyond Shoulder 4, , (b) Percent of Driver Fatalities by Part of Roadway Within Driver Age Group 1 Percent Driver Age On Roadway On Shoulder Beyond Shoulder (c) Younger (Y) and Older (O) Driver Fatalities by Part of Roadway On Roadway: Y On Shoulder: Y Beyond Shoulder: Y On Roadway: O On Shoulder: O Beyond Shoulder: O Figure 16. Number (a), percent (b) and relative likelihood (c) of driver fatalities by PART OF ROADWAY within three age groups.

31 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 25 (a) Driver Fatalities by Traffic Control Traffic Control ALL Center stripe or divider 6, , Stop & go signal Stop sign 1, None shown Yield sign No passing zone Flashing red light Other traffic control Warning sign Turn marks Officer, flagman RR gates or signal (b) Percent of Driver Fatalities by Traffic Control Within Driver Age Group 8 Percent Driver Age Center stripe/div. Stop sign Yield sign Flashing red light Warning sign Officer, flagman Stop/go signal None shown No passing zone Other traf control Turn marks RR gates/signal (c) Younger (Y) and Older (O) Driver Fatalities by Traffic Control Center stripe or divider: Y Stop & go signal: Y Stop sign: Y None shown: Y Yield sign: Y No passing zone: Y Flashing red light: Y Other traf control: Y Warning sign: Y Turn marks: Y Officer, flagman: Y RR gates or signal: Y Center stripe or divider: O Stop & go signal: O Stop sign: O None shown: O Yield sign: O No passing zone: O Flashing red light: O Other traf control: O Warning sign: O Turn marks: O Officer, flagman: O RR gates or signal: O Figure 17. Number (a), percent (b) and relative likelihood (c) of driver fatalities by TRAFFIC CONTROL within three age groups.

32 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 26 (a) Driver Fatalities by Driver Restraining Device Driver Restraining Device ALL No restraint 4, , Belt, shoulder or both 3, , Air bag deployed 1, , Unknown or Not Applicable (b) Percent of Driver Fatalities by Driver Restraining Device Within Driver Age Group 5 Percent Driver Age No restraint Belt, shoulder or both Air bag deployed Unk or NA (c) Younger (Y) and Older (O) Driver Fatalities by Driver Restraining Device No restraint: Y Belt, shoulder or both: Y Air bag deployed: Y Unk or NA: Y No restraint: O Belt, shoulder or both: O Air bag deployed: O Unk or NA: O Figure 18. Number (a), percent (b) and relative likelihood (c) of driver fatalities by RESTRAINING DEVICE within three age groups.

33 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 27 (a) Driver Fatalities by Driver License Status License Status ALL Licensed 9,359 1,16 6,752 1,447 Unlicensed Unknown (b) Percent of Driver Fatalities by Driver License Status Within Driver Age Group 98 1 Percent 94 Driver Age Licensed Unlicensed Unknown (c) Younger (Y) and Older (O) Driver Fatalities by Driver License Status Licensed: Y Unlicensed: Y Unk: Y Licensed: O Unlicensed: O Unk: O Figure 19. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DRIVER LICENSE STATUS within three age groups.

34 Distinguishing Characteristics of Younger and Older Driver Crash Fatalities in Texas 28 (a) Driver Fatalities by Driver Defect Driver Defect ALL None 8,919 1,5 6,46 1,38 Fatigued/Asleep Ill Eyesight defect Other phys defect Other handicap Mentally defective Hearing defect Limbs missing (b) Percent of Driver Fatalities by Driver Defect Within Driver Age Group 14 Percent Driver Age Fatigued/Asleep 2 Ill 1 Eyesight defect Other handicap Hearing defect Other phys defect Mentally defective Limbs missing (c) Younger (Y) and Older (O) Driver Fatalities by Driver Defect None: Y Fatigued/Asleep: Y Ill: Y Eyesight defect: Y Other phys defect: Y Other handicap: Y Mentally defective: Y Hearing defect: Y Limbs missing: Y None: O Fatigued/Asleep: O Ill: O Eyesight defect: O Other phys defect: O Other handicap: O Mentally defective: O Hearing defect: O Limbs missing: O Figure. Number (a), percent (b) and relative likelihood (c) of driver fatalities by DRIVER DEFECT within three age groups.