DRIVER SEX DIFFERENCES IN AUTOMOBILE ACCIDENTS Thesis for the Degree of M. S MICHIGAN STATE UNIVERSITY HERBERT ERECH STOCKMAN 1973 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII \/ DRIVER SEX DIFFERENCES IN AUTOMOBILE ACCIDENTS By Herbert Erech Stockman A THESIS Submitted To Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of criminal Justice 1973 ABSTRACT DRIVER SEX DIFFERENCES IN AUTOMOBILE ACCIDENTS By Herbert E. Stockman It is estimated that by 1975 there will be 125 million licensed drivers and almost 120 million registered motor vehicles in the United States. Each of these vehi- cles will be driven an average of 11,000 miles per year. If present trends continue, this increase in vehicular travel will result in more highway accidents with annual accident costs for 1975 exceeding 11 billion dollars. Although research of quality has been building up piece-by-piece now for almost a half century, we need to knowwa great deal more and do a great deal more before we can expect a significant reduction in traffic accidents. The present study had as its focus a comparison of the characteristics of the automobile accidents of sale and female drivers. The sample for this study consisted of all reports of motor vehicle traffic accidents contained in the Michigan State Police files which occured during the years 1966 and 1971 in Berrien County, Michigan. The data con- sisted of 1,909 single vehicle accidents and “,250 multiple vehicle accidents for 1966, and 2,167 single vehicle acci- Herbert E . St ockmn dents and h,820 multiple vehicle accidents for 1971. The purpose of the study was to determine and explain driver sex differences on variables contained in the accident reports. The hypothesis adopted at the outset, in contradistinction to previous studies, was that all findings were potentially explainable in terms of hypothesized driving exposure differences between the sexes. Many statistically significant differences were found for both single and multiple vehicle accidents. Most were explainable by exposure; there were exceptions, how- ever. Pemales were found to be positively related to the presence of road defects and snow. In combination with other studies, these results indicate that females have more accidents in situations requiring a greater than usual amount of skill. The explanation for this finding was that females tend to drive less frequently under stressful con- ditions, and hence have less Opportunity to learn appropri- ate responses. Male drivers on the other hand, were more likely to have consumed alcohol previous to the accident. They generally traveled faster and were more often ticketed for speeding violations. The explanation given to these re- sults was in terms of cultural roles and driving confidence rather than in terms of differences between the sexes in driving abilities under the influence of alcohol or at high speeds. Based on the results of the study, practical sug- gestions regarding differential educational and training procedures for the two sexes were offered. DRIVER SEX DIFFERENCES IN AUTOMOBILE.ACCIDENTS By Herbert Erech Stockman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of criminal Justice 1973 APPROVED: (Memberl {A ACKNOWLEDGMENTS To all who knowingly or unknowingly assisted in the conception and realization of this thesis, I am heavily indebted. In particular, special thanks is due my chairman, Doctor Robert C. Trojanowicz, and my committee members Doctor Victor G. Strecher, and Mister Adrian H. Koert. And to those members of my family who watched my academic odyssey with apprehension . . . I thank you for your patience and confidence. TABLE OF CONTENTS CHAPTER PAGE I 0 INTRODUCTION 0 O O O O O O C O O O O O O O O 1 e e e e e e e e N Importance of the Study . Review of the Literature . . . . . . . . b Accident Causation . . . . . . . . . . . 15 Exposure 0 O O O O O O O O O C O O O O O 19 Driving Habits of the Two Sexes . . . . . 25"” Specific Predictions . . . . . . . . . . 29 II. THE DATA . . . . . . . . . . . . . . . . . . 3” III. METHOD OF ANALYSIS . . . . . . . . . . . . . 38 IV. THE RESULTS . . . . . . . . . . . . . . . . #2 Time variables . . . . . . . . . . . . . #2 Road Characteristics . . . . . . . . . . 52 Location . . . . . . . . . . . . . . . . 59 Collision Characteristics . . . . . . . . 60 Vehicle Characteristics . . . . . . . . . 65 [/Driver Characteristics . . . . . . . . . 68 I Driver Behavior . . . . . . . . . . . . . 77 . Weather variables . . . . . . . . . . . . 91 V. DISCUSSION . . . . . . . . . . . . . . . . . 102 Time variables . . . . . . . . . . . . . 102 Road Characteristics . . . . . . . . . . 103 Location . . . . . . . . . . . . . . . . 106 111 CHAPTER Collision Characteristics Vehicle Characteristics . ,/) I‘Driver Characteristics (\Dri:er Behavior . ... . . Heather variables . . . . VI. CONCLUSION . . . . . . . . . BIBLIOGRAPHY . . . . . . . . . . . . APPENDIX . . . . . . . . . . . . . . iv PAGE 107 109 110 112 115 117 121 125 TABLE 1. 2. 7. 8. 9. 10. 11. 12. 13. 14. LIST OF TABLES Chi Square Values and Significance Levels for Sex Crossed With All Other Variables . . . . . The Percentage of Accidents for Each Sex Within Each Category of Light Condition . . . . . . . The Percentage of Accidents for Each Sex Within Each Category of Light Condition Within Locations 0 O O O O O 0 O O O O O O O O O O O The Percentage of Accidents for Each Sex Within Each category of Time of Day . . . . . . . . . The Percentage of Accidents for Each Sex Within Rush Hour (7-9.AM a “-6 PM) and Other Daylight Hour8(9AM-'¥PH)............o The Percentage of Accidents for Each Sex Within Each Category of Day of the Week . . . . . . . The Percentage of Accidents for Each Sex Within Each Category of Weekday - Weekend . . . . . . The Percentage of Accidents for Each Sex Within Each category of Highway Classification . . . The Percentage of Accidents for Each Sex Within Each Category of Road Geometry . . . . . . . . The Percentage of Accidents for Each Sex Within Each Category of Road Geometry Within Locations The Percentage of Accidents For Each Sex Within Each category of Road Geometry Within Light Conditiona....o............. The Percentage of Accidents for Each Sex Within Each Category of Road Defect . . . . . . . . . The Percentage of Accidents for Each Sex Within Each category of Intersection . . . . . . . . The Percentage of Accidents for Each Sex Within Each Category of Intersection Within Locations V PAGE “3 #6 u? #8 50 51 52 53 55 56 57 58 59 61 TABLE 15. 16. 17. 18. 19. 20. 21. 22. 23. 2h. 25. 26. 27. 28. 29. 30. The Percentage of Accidents for Each Sex Each Category of Location . . . . . . . The Percentage of Accidents for Each Sex Each Category of Location Within Light conditIOHSeeeeeeeeeeeeeee The Percentage of Accidents for Each Sex Each Category of Direction Analysis . . The Percentage of Accidents for Each Sex Within Within Within Within Each Category of Direction Analysis Within Locations . . . . . . . . . . . . . . . The Percentage of Accidents for Each Sex Each Category of Vehicle Type . . . . . The Percentage of Accidents for Each Sex Each Category of Age . . . . . . . . . . The Percentage of Accidents for Each Sex Each Category of Age Within Locations . The Percentage of Accidents for Each Sex Within Within Within Within Each Category of Age Within Light Conditions . The Percentage of Accidents for Each Sex Each Category of Experience . . . . . . The Percentage of Accidents for Each Sex Within Within Each category of Experience Within Locations . The Percentage of Accidents for Each Sex Each Category of Experience Within Light Goad 1t1°n8 O O O O O O O O O 0 O O O O O The Percentage of Accidents for Each Sex Each Category of Age Within Experience . The Percentage of Accidents for Each Sex Each category of Registration . . . . . The Percentage of Accidents for Each Sex Each Category of Speed Violation . . . . The Percentage of Accidents for Each Sex Each Category of Speed Violation Within Locationseeeeeeeeeeeeeee The Percentage of Accidents for Each Sex Within Within Within Within Within Within Each category of Speed Violation Within Light Conditions . . . . . . . . . . . . . . . Vi PAGE 62 63 6h 66 6? 69 7O 71 72 73 74 75 76 77 79 80 TABLE 31. 32. 33- 3‘5. 35. 36. 37- 38. 39. #0. #1. #2. #3. an. “5. #6. RACE The Percentage of Accidents for Each Sex Within Each category of Right-of-way Violation . . . 81 The Percentage of Accidents for Each Sex Within Each category of Right-of-way Violation Within Locations.................. 82 The Percentage of Accidents for Each Sex Within Each Category of Violation Other Than Speeding or Drinking O O O O O O O O O O O O 0 O O O O 8“ The Percentage of Accidents for Each Sex Within Each category of Alcohol . . . . . . . . . . . 85 The Percentage of Accidents for Each Sex Within Each Category of Alcohol Within Locations . . 87 The Percentage of Accidents for Each Sex Within Each category of Alcohol Within Light Conditions . . . . . . . . . . . . . . . . . . 88 The Percentage of Accidents for Each Sex Within Each Category of Sleep . . . . . . . . . . . . 89 The Percentage of Accidents for Each Sex Within Each Category of Speed . . . . . . . . . . . . 90 The Percentage of Accidents for Each Sex Within Each Category of Speed Within Locations . . . 92 The Percentage of Accidents for Each Sex Within Each Category of Speed Within Light Conditions 93 The Percentage of Accidents for Each Sex Within Each Category of Weather . . . . . . . . . . . 9“ The Percentage of Accidents for Each Sex Within Each Category of Weather Within Locations . . 95 The Percentage of Accidents for Each Sex Within Each category of Weather Within Light Conditions 97 The Percentage of Accidents for Each Sex Within Each category of Surface Condition . . . . . . 98 The Percentage of Accidents for Each Sex Within Each Category of Surface Condition Within Locations 0 O O O O O O O O O O O O O O O O O 100 The Percentage of Accidents for Each Sex Within Each Category of Surface Condition Within Light conditions 0 O O O O I O O O O O O O O O O O O 101 vii CHAPTER I INTRODUCTION The debate between the sexes concerning the superi- ority of each in driving an automobile has dominated many an evening's parlor discussion. Hales are quick to point out the shortcomings of “women drivers", while females are equally quick to counter with harsh criticisms of “inpatient men“. Automotive insurance companies tend to side with the fairer of the sexes, a fact few women overlook'in creating a convincing argument. This is true for even young drivers. There was a time when all drivers under 25 were charged higher rates for insurance premiums; however, insurance companies found the risk to be greater for only young male drivers, and, hence, young female drivers were relieved of at least part of the greater assessment. It is only fair to note, however, that insurance companies are only inter- ested in gross number of accidents and other causes of claims. The well-recognized fact that men on an average drive more and probably during periods of higher potential risk is irrelevant to their purposes. Men, incidentally, are far from unaware of these facts in forming a rebuttal of the feminine interpretation. The reason that this disagreement has been able to sustain discussion for so long is because there is indeed 2 no clear cut answer. Possibly in some situations the aver- age man is a better driver than his wife, while in other situations she might prove superior. A blanket statement that one sex is superior to the other is bound to be an oversimplification. The present study has as its focus a comparison of the characteristics of the automobile accidents of male and female drivers. The specific purpose of the study is to discover and try to explain statistically significant dif- ferences between the sexes on variables which are recorded in accident reports. Included in these variables of in- terest are time and location variables, road, collision, vehicle and driver characteristics, driver behaviors, and weather. Importance of the Study There are, of course, more important reasons for conducting this study than merely ascertaining which sex is the better driver. The great number of accidents which are taking place on our highways constitutes a major prob- lem and the integration of motor vehicles into our way of life has become very costly in terms of fatalities, inJu- ries, and damaged equipment. As miles traveled, passengers carried, and tons conveyed have increased, traffic has be- come more dense. Also speed levels have risen, and there has been an increase in deaths, injuries and monetary costs. 3 In fact, highway safety has become a matter of pressing national concern. According to a public opinion survey conducted by the Michigan State university Highway Traffic Safety Center, Michigan citizens feel that crime and traffic safety are 1 The sheer the most important problems facing the state. magnitude of losses from highway accidents demands system- atic, carefully planned research studies. Although research of quality has been building up piece-by-piece now for al- most a half century, we need to know a great deal more and do a great deal more before we can expect a significant reduction in traffic injuries and deaths. In the state of Michigan alone, a total of 310,015 reported motor vehicle traffic accidents and 2,152 highway fatalities were recorded by the State Police during 1971. This was the eighth consecutive year and the eleventh time in thirty-four years of record keeping that Michigan traf- 2 This certainly fic deaths have exceeded 2,000 annually. leaves no room for complacency nor does it indicate that we are making much progress in the overall struggle for safer highways. 1"HSU Conducts Two Surveys,” Traffic Safety, (May, 1966, p. 25. 2Michigan Department of State Police, Michi an Traffic Accident Facts: 122 , (Lansing, Michigan, 72), p. 1. u If accidents are to be related to highway, vehicle, and human factors in a meaningful fashion it is essential that the relative exposure of people to accidents be meas- ured. One method of accomplishing this is to compare the characteristics of the automobile accidents of male and female drivers. Looking at accidents is a negative approach to the study of drivers to be sure, but it is not an invalid ap- proach. There is no other time in the driving history of an individual when so much data is collected on a small segment of driving. Nor is any aspect of driving as dis- ruptive, harmful, and in need of reduction as accidents. Any insight gained concerning the general etiology of automobile accidents would seem to be worthwhile. Review of the:Lgterature In one of the first studies which addressed itself to the question of sex differences in automobile accidents, M. S. Viteles and H. M. Gardner used District of Columbia taxicab drivers as their subjects. In a sample of approx- imately 200 men and from 35 to no women they found female drivers to be involved in 3.h9 times as many accidents per 1,000 miles than male drivers. When considering only seri- ous accidents they found that women were involved in fewer accidents than men but the women caused more accidents on the part of other drivers through the driving tactics they 5 employed.3 This rather startling result must be tempered by mention of several uncontrolled variables. The female drivers were not allowed to drive at night. They further- more did not have to pass a standard driving test which males did. Also fourteen percent of the women were com- pletely inexperienced when hired and had to be trained by the company. In 1927 and 1928 use of the automobile was not widespread and female drivers were rarely encountered at that time. This leads to the suspicion that in addition to the fourteen percent of the women who were completely inexperienced, most of the rest of the women had rather limited experience. Given these shortcomings, the Viteles and Gardner study seems to have little more than historical value. A. R. Lauer also compared the accident involvement of male and female drivers. His subjects were a sample of 7,692 drivers drawn from the Iowa state records during the years of l9h8 and 19h9. The accurate mileage data available to Viteles and Gardner was not available to Lauer, so he had drivers estimate their annual mileage both during the day and at night. Using these estimates, he found that females did about ten percent of the driving 3M. S. Viteles and H. M. Gardner, "Women Taxicab Drivers: Sex Differences in Proneness to Motor Vehicle Accidents,“ Egzsgnng;_ggggg§;, Vol. 7. (1929), pp. 3h9-355. 6 and had about nine percent of the accidents.“ No differen- tiations were made concerning the severity of the accidents. In terms of accidents per annual mileage, Lauer found male drivers under thirty contributing a disproportionaly large amount to the accident total. Males also appeared to drive for five years before they exhibited any improvement in their driving record, while female drivers began improving immediately.5 Clifford 0. Swanson, Lilliam C. Schwenk, and A. R. Iauer studied the drivers of vehicles involved in all the fatal accidents which transpired in Iowa during 1955 and 1956. Driver fatalities showed almost a nine to one ratio of male to female for 1955 and about a six to one ratio in 1956. When the two sexes were equated for estimated annual mileage (gained from drivers still alive), the ratio for the two years combined reduced to 2.49 to 1, male to female. The investigators also found that the fatal acci- dent rate for both sexes decreased until about age thirty- five, stayed level throughout the middle years, and began increasing again around the age of fifty-five. Males had a higher accident rate, for all ages except twenty-five to “A. R. Lauer, I'Age and Sex in Relation to Acci- dents,” Accident Research Methods and Aggroaches, ed. William Haddon, Edward A. Suchman, and vid K ein, (New York: Harper and How, 196“), pp. 137-138. 51bid., p. 137. 7 twenty-nine and sixty to sixty-four when corrected for average annual mileage.6 Moving ahead in time and to the West coast, the l96h California Driver Record Study also compared male and fe- male drivers on accident involvement, as well as conviction frequency. In analyzing the driving records of over 90,000 males and over 65,000 females, the investigators found fe- male involvement to be relatively less frequent than that of males in accidents and traffic violation convictions. Males had a mean of .260 accidents over the 3 years: while fe- males had a mean of .126 during the same period.7 However, when the accidents were broken down into those reported to the California Highway Department and those not, the difference between males and females was accentuated in the former case and shrunk in the latter, although males still almost doubled females.8 A small portion of this difference could be attributed to an enforcement differ- ential, however, the magnitude of the difference indicates that males are probably involved in more severe accidents. 6Clifford O. Swanson, Lilliam C. Schwenk, and A. R. Lauer, ”Age and Fatal Motor vehicle Accidents,” Highway Research Abstracts, Synopsis Issue, Vol. 27. (Washington, 5. 5.: National Academy of Science - Mational Research Council, December, 1957), p. 69. 7"Accidents, Traffic Citations and Negligent Oper- ator Count by Sex,“ The 1960 California Driver Record Stud , Part II, (Sacramento: California Department of Motor Vehicles, Report No. 20, March, 1965), p. 19. 81bid., p. 15. 8 During the same period of time, males had a mean of 1.103 total convictions, while females had a mean of only .374.9 When mean accident frequency was plotted as a func- tion of age separately for single and married males and females, several interesting relationships were apparent. Up to the age of twenty-five, married males actually had a higher average accident frequency than single males. However, beyond that single males were consistently higher. Both curves for males had negative slapes, dropping sharply through the younger ages and more slowly as the age became greater. Both female curves, on the other hand, remained 10 These results are incon- relatively flat over all ages. sistent with those of Lauer, reported earlier. Perhaps the discrepancy lies in the fact that Lauer did equate for esti- mated annual mileage, and the California study used data fifteen years more recent than Lauer. Married females were consistently lower in accident frequency than single females. In fact, beyond the age of thirty the curves of single fe- males and married males were at almost the same level.11 The results were very similar when mean conviction rate was plotted rather than mean accident rate. The slope for males 91b1dc ’ p0 190 10Driver Record by Age, Sex and Marital Status," The 1264 California Driver Record Study, Part V, (Sacramento: ifornia partment of Motor Vehicles, Report No. 20, June, 1965), pp. “-5. Ibld. 9 was a little more sharply negatively decelerated, and mar- ried males were a little more consistently lower than sin- gle females. Overall, however, the two sets of curves were very similar.12 B. J. Campbell studied 32,387 fatal and injury acci- dents involving only passenger cars and within the juris- diction of the state highway patrol. (Although not cited explicitly, the state was probably New York). Male drivers accounted for eighty percent of this sample. By looking at male and female accident involvement relative to total involvement of each sex respectively, Campbell found the proportion of accidents for females was higher than that for males on weekdays, while males were higher on weekends. When the data was plotted by time of day, male drivers were found to have proportionately more accidents from 6 P.M. to 6 A.M., with female drivers being proportionately higher during the remaining twelve hours.13 Using data from all of the 17,400 accidents con- tained in the 1957 Michigan State Police records, Terrence M. Allen calculated phi coefficients between all pairs of twenty-three dichotomized variables. Included among these variables was sex (female involved or not). Let it be noted that not all the variables were naturally dichotomous Ibid. 133. J. campbell, “Driver Age and Sex Related to Accident Time and Type,” Traffic Safety Quarterly Research Review, Vol. 10, (1966), pp. 36-h3. 10 as is sex. Rather, most had to be artificially dichoto- mized to make the calculation possible. The phi coeffi- cients were then factor analyzed. Sex loaded most highly (-.h5) on factor III (night). Other variables showing high loadings on the same factor were alcohol (.55). daylight (—.77), and rush hour (-.63). The signs of the loadings indicate that females were associated with daylight, rush hour, and absence of alcohol, with males being associated with the converse of these variables. Moderately small loadings for sex were found on factor VI (.21) which Allen named "youth-inexperience”, factor V (-.18) referred to as ”rural”, and factor VIII (-.17), a small factor having vehicle defect as the only variable loading highly}!+ In another study of California drivers, Hugh 8. Penn used 5,203 single vehicle accidents transpiring in September of 1961 and June of 1962 as data. Two previous California studies (unreferenced), one to determine the proportion of each sex in the total driving population and the other to determine the average annual mileage for driving members of each sex, were used to equate males and females on exposure. Unfortunately, the method by which annual mileage was deter- mined was not made explicit. Using this derived index of relative driving distance, the investigators concluded that 1“Terrence M. Allen, “A Factor Analysis of Accident Records," Highway Research Record No. 22, (Washington: National Aca any of Sciences - Nationa Research Council, January, 1965), p. 20. 11 if equated for mileage male drivers would have been involved in 53.6 percent of the accidents studied and female drivers would have been involved in the remaining h6.h percent.15 Each accident was placed into one of eleven catego- ries of ”causes" or, more accurately, "precipitating fac- tors." The accidents of male drivers were found to be sig- nificantly more frequently contained in the categories of speed, drowsiness, and drinking: while female drivers were found more frequently in the categories of faulty driving, adverse driving conditions (emergency situations in the driving environment), and distraction inside of the vehicle. The categories of mechanical failure, distraction outside the vehicle, medical problems, unknown vehicle, and miscel- laneous did not differentiate the sexes at a significant level.16 The results of a study by Leonard Uhr appear consist- ent with Penn's finding of a tendency for female drivers to be more frequently involved in accidents precipitated by ”adverse driving conditions." The brevity of Uhr's article precluded any precise understanding of the design and con- duct of the study. A motor scooter was used to confront drivers with an unusual situation. At the time of his 15Hugh 8. Penn, I'Causes and Characteristics of Single Car Accidents,“ Highway Research Record No. 22, (Washing- ton: National Academy of Sciences -'fiationa1 Research Council, January, 1965), pp. l-l6. 16Ibid., p. 3. 12 study motor scooters were newly legalized, and very un- common. The study was carried out as follows: An auto was first judged to be behaving dangerously toward a motor scooter (by cutting across or into the scooter's path from a stop street or alley so that the scooter driver was forced to brake or swerve his vehicle). Only after this judgement was made, the sex of the auto driver was determined. . . Twenty-five such incidents were accumulated, along with twenty-five comparison incidents . . .17 This behavior was found to be highly related to the sex of the driver. Nineteen of the drivers judged to be behaving dangerously were women, while only six men were so desig- nated: of the ”safe” drivers, twenty-two were men and only three were women. Thus, female drivers were significantly more likely to make an inappropriate and dangerous response in the presence of this situation.18 A preliminary report of a study conducted at Northwestern University by J. Stannard Baker adds further evidence of an adverse driving situation being more likely to terminate in an accident for females than males. Baker found that female drivers have about four times the average likelihood of accidents following flat tires. Females under twenty were found to be twenty-two times as likely, and women between twenty and thirty-five were about five times as likely. Apparently the additional skill needed to avoid 17Leonard Uhr, "Sex as a Determinant of Driving Skills: Women Drivers,“ Journal of Applied Psychology, Vol. “3. (1959). p. 35. Ibid. 13 a collision in such a situation tended to be found more often with male drivers than female drivers.19 In a very ambitious study of accident causation on the Pennsylvania Turnpike, Paul Blotzer, et. al. analyzed over 9,000 State Police Reports compiled during the years of Operation of the turnpike. The investigators adopted the model that ”Accidents are caused in a vast majority of cases by human error, and to a smaller extent, by vehicle error -'both factors influenced by the environmental con- ditions incountered in the driving operation."20 Those accidents judged primarily caused by human error, as op- posed to vehicle error, were classified into one of eight categories of "causes” or “precipitating factors,“ based on what was judged to be the primary cause of the accident. When accidents involved more than one vehicle, only the driver of the vehicle judged at fault was entered into the analysis. The investigators found a significantly greater per- centage of female drivers involved in accidents classified in two human error categories'- 'flailure to cope with road conditions“ and "deficiencies in routine driving skills." 1gllighway Research News Briefs, Hi hwa Research News No. 31, (Washington, D. 0.: National Academy of Sciences, ig way Research Board, Spring, 1968), pp. 8-9. 20Paul Blotzer, Richard L. Krumm, Donald M. Krus, and Donald E. Stark, Accident CausationI- Penns lvania Turn ike Joint Safet Research Erou , (Earrisburg: Westinghouse Air ”' i 5) 11. Brake Co., 95 , p. 1% The ”failure to cope with road conditions“ category con- tained mostly skidding accidents with a few resulting from high winds. The “deficiencies in routine driving skills“ category contained very similar types of accidents. The difference between the two categories is simply that in the case of the former no judgment of specific driver error was made. ”The skidding accidents listed as 'failure to cope with road conditions' would probably be included in the 'deficiencies in routine driving skills' category had suf- ficient information regarding what the driver did or did not do been included in the accident reports for these acci- 21 This result seems consistent with the results of dents.” Penn and Uhr. Since sex differences were not mentioned in the report for the remaining six categories, it must be assumed no significant differences existed, other than within the 'inattentiveness” category, where male drivers were more often found in the ”asleep" subcategory. The investigators isolated four broad environmental variables, ”light conditions", “weather conditions”, "road- way conditions”, and the "roadway element“. Male drivers were found to be involved in relatively more accidents at night. 'Females, of course, were involved in more daytime 22 accidents. No mention was made of sex differences on the 211bid., p. 37. 221bid., p. 9n. 15 other three variables; therefore, again, the assumption must be made that no such differences existed in the data. Accident Causation In reporting these studies, the concept of cause has been introduced without adequate discussion. Allen has pointed out that most people are aware that accidents can- not validly be thought of as having a single cause.23 In speaking of "myths and misconceptions in traffic safety,“ William E. Tarrants stated, "It is evident . . . that acci- dents have multiple causes . . ..2h However, some very thorny definitive and other practical problems in accident research are present when accidents are considered as being caused by more than one factor. Edward A. Suchman defined an accident as ". . . the end product of a sequence of acts or events which result in some 'unanticipated' consequence that is judged as 'undesirable'. . .“25 What in this se- quence of events can be defined as one of the causes, and what can be ignored? The situation is such that changing any one or more of the factors to determine the effect upon 23Allen, op. cit., p. 17. quilliam E. Tarrants, "Myths and Misconceptions in Traffic Safety," Highway Research News No. 21, (Washington, D.C: National Academy of Sciences, Rig way Research Board, Spring, 1968), p. 60. 25William Haddon, Edward A. Suchman, and David Klein, Accident Research_Methods and Approacheg, (New York: Harper and Row, 196E), p. 275. 16 the end product cannot be accomplished. Furthermore, the unexpected nature of an accident combined with the rapidity with which it transpires makes an accurate assessment of the sequence of acts on events difficult if not impossible. Also, statistical analyses are much simpler if each acci- dent is classified into only one category, even though with the development of high Speed computers, this is becoming less of a handicap. Considerations such as these led many investigators to classify each accident according to what they considered the primary cause, while still maintaining a multiple cau- sation model on a theoretical level. For example, Penn dis- cussed the concepts of accident conditions, factors, and causes originally formulated by Baker. Accident conditions are ”. . . environmental or behavioral circumstances which surround the accident, but which may not be a part of the causative process . . . Factors are causative elements, no one of which is usually strong enough to produce an acci- dent. In combination, however, and triggered by some pre- 26 cipitating incident, they will produce a mishap." The "cause" then is the precipitating incident. Blotzer, et. a1. emphasized that accidents are ". . . the result of in- teractions between environmental, vehicular, and human fac- tors."27 They included a list of possible ”antecedent“ 26Penn, 0p. cit., p. h. 27Blotzer, et al., op. cit., p. 35. 1? factors for each of the "precipitating“ factors. Cause within this conceptual scheme became the precipitating fac- tor. However, Tarrants rightly pointed out that ”It may be possible to identify proximate and distal causal factors in a particular accident by going back in time or space from the point of impact. However, to identify one of these factors as the primary cause is a very difficult if not impossible task."28 The assumption of a single or primary cause does not have to be made at all. When an automobile accident occurs, there are many factors present which may be recorded in an accident report without any mention of which of the factors or what factor caused the accident. Accidents can then be studied to determine which of these factors tend to be re— lated, or, in other words, which variables tend to be asso- ciated across accidents and which tend to be independent of each other. The accident can then be analyzed in a manner more congruent with what it is, a complex many-faceted event. Reducing this event to one cause just does not do it justice. For example, the time of day an accident takes place can be an important variable in the over-all understanding of acci- dents. But the time at which an accident took place would hardly be expected to ever attain the distinction of having ”caused” the accident. Likewise the driver judged not at fault in a multiple vehicle accident still played a signif- 28Tarrants, op. cit., p. 59. 18 icant role in that accident and really does not deserve to be ignored completely. For example, the excessive speed of a driver may be judged as the cause of an accident, but had the other driver judged that speed properly and not pulled out, the accident would not have happened. The approach taken in the study by Allen, reported earlier, was that of characterizing each accident by the presence or absence of twenty-three variables. Relations between these variables were sought. Thus, notions of cau- sation were not necessary.29 The shortcoming of that study lay in the necessity for artificial dichotomization of some of the variables: thus limiting the amount of information gained from each variable. Also, because Allen analyzed all accidents at once, he had to place dissimilar accidents into the same categories. For example, single vehicle accidents involving female drivers and multiple vehicle accidents in- volving either one or two female drivers had to be placed into the same category on the "female involved or not" variable. This was true of other variables as well.30 The present study also eliminated any consideration of cause in the accidents studied. The question asked was, how are the two sexes different in terms of the character- istics of the accidents they have? The underlying statisti- cal assumption made was that if the two sexes are not dif- 29Allen, op, cit., p. 18. 3°Ibid., p. 19. 19 ferent on a specific variable, then the proportion of acci- dents for each sex within each category of that variable will be very similar. For example, if male drivers had 60 percent of their accidents in an intersection, while female drivers had 60.1 percent of their accidents in an intersec- tion, then no difference between the sexes on the "intersec- tion or not“ variable can be inferred. On the other hand, if the two relative proportions differed greatly, then some relationship appears to be in evidence. For the example above, a variable was chosen which is naturally dichotomous. This method is not limited to dichotomous variables, but rather is applicable to a larger number of categories within each variable. The present study also gained precision by analyzing the data from sin- gle and multiple vehicle accidents separately. Reasons will be discussed in more detail later. Esme Exposure is an equally important concept in accident theory and research. In order for an accident to occur a person has to place himself or be placed in a situation whereby he can become involved. In other words, he has to be exposed to some risk. It is immediately obvious that one who is never near an automobile will never be involved in an automobile accident. It follows that other things being equal the accident rate of a driver will be an increasing function of the amount he drives. This reasoning was made 20 very explicit by Ross A. McFarland and Roland C. Moore con- cerning their sample of young male and female drivers. “If boys and young men drive three times more than equal-aged members of the opposite sex, they acquire three times as much exposure to the possibility of accidents . . ."31 In much of the accident literature exposure is operationally defined as precisely the number of miles driven per unit of time, usually a year. In studies concerned with sex differ- ences, investigators, recognizing that males tend to drive more than females, have attempted to equate the sexes by multiplying by a factor dictated by annual mileage differ- ences. It must be noted, however, that accurate mileage data are difficult to obtain. Frederick E. vanosdall ob- tained exposure data by having 6,358 drivers from the state of Michigan estimate the number of miles they drove during the average week. Male drivers represented 70.35 percent of the drivers in the study sample and female drivers 29.65 percent. The estimates indicated that male drivers drove 88.8 percent of the total miles and female drivers drove 11.2 percent. Males estimated they did sixty-eight per- cent of their driving in the daytime and thirty-two per- cent at night. Females, on the other hand, indicated they did seventy-six percent of their driving in the daytime and 31Ross A. McFarland, and Roland C. Moore, Youth and the Automobile, (Golden Anniversary White House Conference on Children and Youth, 1960), p. #69. 21 only twenty-four percent at night.32 These findings are consistant with the results of Campbell's study.of male and female accident involvement, reported eariler. Earl Allgaier,33 and Clifford 0. Swanson, et al.,3u obtained overall exposure data by having drivers estimate the total number of miles they had driven per year. This method, although perhaps better than no data at all, leaves much to be desired. Estimates are likely to be in error, for most drivers do not keep accurate records of the amount they drive.35 Little trips, such as are common to women, have a tendency to "add up” mileage much more quickly than is sometimes realized. Like Vanosdall, Siebrecht,36 and Lauer37 went one step farther in having members of both sexes estimate their total daytime and nighttime driving. The data collected 32Frederick E. Vanosdall, “An Introductory Study to Show the Relationships Between Michigan Drivers by Age, Sex and Exposure in Miles of Motor Vehicle Operation,” (unpublished Master's thesis, Michigan State University, 1966). pp. 69-70. 33Earl Allgaier, ”Some Road-User Characteristics in the Traffic Problem," Traffic Quarterly, Vol. h, (1950), PD. 59‘770 3“Swanson, et a1,, op: cit., p. 69. 35H’s.ddon, et a1“ op: cit., p. 138. 36E. Siebrecht, “A Preliminary Report of Accident Characteristics of Iowa Drivers,‘ Iowa Academy of Science, 1953 Proceedings, 60, pp. 552-557. 37Louer, op. cit., pp. 130-138. 22 were perhaps even less accurate than estimates of the total mileage driven. It is unlikely that people have the abil— ity to be sufficiently accurate in estimating daytime and nighttime driving, and such data could be more misleading than enlightening. The inaccuracy of the estimates is not the only rea- son why it is not adequate to consider the sexes equalized on exposure by merely dividing accident frequencies by average annual mileage. There is no basis for concluding that annual mileage directly reflects accident risk inde- pendent of other factors such as where an individual drives and the experience gained while driving under different circumstances. Ross A. McFarland observed that drivers '. . . differ widely in their exposure, even though they have equal records, one may have driven under vastly dif— ferent circumstances.“38 Insofar as there are differences in the amount of driving done by the sexes in factors which have an influence upon risk, there will be differences in the actual exposure to accidents which are not reflected by annual mileage alone. It is well documented, for example, that limited access freeways are less dangerous per mile driven than two-lane roads. Also, per mile driven more “serious“ accidents happen after dark than during the day- light hours. To say, then, that an individual driving one 38Ross A. McFarland, Roland C. Moore, and A. Bertrand Warren, Human Variables in Motor Vehicle Accidents, (Boston: Harvard School of Public Health, 1955), p. 12. 23 hundred miles at night on a two-lane road is exposed to the risk of an accident equally with another individual driving one hundred miles on a freeway in the middle of the day is obviously not true. The discussion of exposure has been actually tangen- tial thus far, since the present study is not particularly interested in total accident involvement but rather sex differences in attendant variables recorded in accident records. Within each sex a great deal of variance is ex- hibited in driving habits leading to overlap on any expo- sure variable. This will be discussed in more detail below. In spite of the overlap, there are believed to be stable differences between the sexes taken as two distinct popu- lations. These differing driving habits lead to differing levels of exposure to accidents of a certain type. For example, as was previously pointed out, in this culture men do a higher proportion of their driving at night than women. Given this information, male drivers would be expected to have a higher prOportion of their accidents at night than female drivers. A difference on this variable in that di- rection is at least partially explainable by the exposure difference. If such a difference was not accompanied by a similar difference in exposure, an alternative explanation, probably concerned with how males and females drive rather than when or where they drive would be warranted. An example of a conclusion which did not take expo- sure into account is found in the California study of sin- 2n gle car accidents. . . . men's aggressiveness, daring, and rebellious- ness make for reckless and often unlawful behavior. As women's psychological make-up embodies the ob- verse of these traits, they are comparatively low in accidents due to speed, drinking and perhaps ths9 aftermath of recreational activities - drowsiness. This explanation was put forth in response to the finding that males had significantly more accidents caused by speed, drowsiness, and drinking. The explanation is in terms of psychological determinants of the way people drive. A wholly more parsimonious explanation, it would seem, is that of exposure. Males tend to drive relatively more in high-speed situations. Likewise males drive relatively more frequently at night, thus leading to more accidents involving drowsiness. Males probably also drive relatively more frequently in the presence of alcohol, not necessarily because of a difference in psychological makeup, but rather because of cultural roles and expectations. More will be said concerning these expected differences in exposure later in the paper. Lilliam C. Schwenk utilized an explanation similar to that of Penn when she discussed her finding of a tend- ency for males at younger ages to be killed than females, both as drivers and pedestrians. She wrote, '. . . this may be due to the masculine characteristic of aggressive- 39"Driver Record by Age, Sex and Marital Status," The l26h California Driver Record Study, Part V, Report No. 0, Sacramento: California Department of Motor Vehicles, June, 1965), p. h. 25 mess.”O This explanation also ignored exposure differ- ences. It is the contention of this researcher that this type of reasoning only clouds the issue. A realization of which differences may or may not be accounted for by con- sideration of exposure is essential'for an understanding of the data. Driving Habits of the Two Sexes It would be good at this point to use past studies to determine the driving habits of males and females. How- ever, to this researcher's knowledge, there are no suitable studies available. As a suggestion for future research, a method whereby this needed information could be accurately obtained is in the form of personal interviews. Partici- pants should be requested to state exactly where and when they drove during the immediate past. The number of past days for which people can accurately remember their driving would have to be determined, and then this limitation could not be exceeded. A large number of interviews, repeatedly with the same subjects or with different subjects, would be necessary, but an accurate and meaningful pattern could be expected to emerge which could then serve as reliable expo- sure data for use in this study and in other similar studies concerned with sex and other biographical variables. uoLilliam C. Schwenk, “Age and Sex in Relation to Fatal Traffic Accidents for 1957 -'A Continuation Study,” Iowa Academy of Science, 1958 Proceedings, 65, p. #25. 26 This study must content itself with reasoning from observed differences in the roles of the two sexes in this culture. The following discussion and the specific predic- tions which follow are of an armchair quality which leaves the reader the choice of accepting the reasoning as logical and congruent with his own observations or rejecting all or parts of the reasoning. The main thrust of the report is based on these suspected differences in the driving habits of the two sexes, and the conclusions reached will be in- valid insofar as the premises are rejected. As was previously mentioned, the driving habits of the two ssxss exhibit a great deal of overlap. In general, although in physical terms sex is almost completely a di- chotomous variable, such is not so nearly the case from a psychological point of view. The distribution of the two sexes‘have heen found to be overlapping on all measurable psychological variables. On any given attribute, some files will be found more "feminine" than some females and vice versa.' This is certain to be true of driving habits. For example, it is very possible for an unmarried female to develop driVing habits which are more similar to a man's than to anotherrmamber of her own sex. She may be going to and coming from work at the same time as files and also using a great deal of driving at night and on the highways. The driving of a certain male salesman, on the other hand, might be characterized by short, daytime trips in the city, a pattern of driving which is more typical of female driv- 2? ers. These overlaps must be kept in mind as discussion turns to suspected differences in exposure and to specific predictions as to how the sexes will differ as a result of these differences. The typical male and the typical female Idll be referred to often, indicating some sort of average or normal member of each group who, in fact, has no real existence. The rule is rather relatively small mean dif- ferences and relatively large amounts of overlap. For this reason, small but significant differences become important, and the underlying relationships must be sought with care. How can the driving habits of male and female driv- ers be characterized? Some differences seem immediately apparent. Men tend to drive to work in the morning, re- turning in the late afternoon. The female, in the meantime, has taken the children to school, gone shopping, and per- formed a myriad of other small chores necessitating short trips. In the evenings, the male of the household is more likely to do the driving, especially as the hour gets later. Certainly if adult members of both sex travel together the male is more likely to be the one driving. If alcohol is present it was probably preceded by a trip to a bar or the home of friends. Masculine escorts usually accompany women in both of these pastimes, in which case the male is more likely to be the one driving if he is able. It is relatively rare for a woman to attend a bar alone or with female companions compared to that activity by males. 28 Though members of both sexes commonly-drive in town, the driving of'a female is probably more typically non- rural, since she tends to stay less far away from home and since the shopping is usually done in highly populated areas. Men, on the other hand, might be expected to do more of the rural driving. Family trips usually find the male doing most of the driving. Hales would seem more likely to be driving On weekends, with females driving rel- atively more on the weekdays while males are working. since female truck drivers are still a rare sight, a' straightforward assumption is that male drivers drive vehi- cles other than passenger cars relatively more than female drivers. In our culture, timidity is more characteristic of females which leads to the assumption that females are less likely to drive under adverse driving conditions, such as snow, rain, etc. As a group they are probably more prone toward waiting for the weather to improve or for a man to do the driving. Among young drivers an automobile is much more im- portant to the status of a male than a female. Young males own their own vehicles in more cases than young females, as well as more frequently borrowing the family car. Once behind the wheel, it seems reasonable that boys would tend to drive greater distances than girls. Certainly it is a strong American custom for the male to do the driving on dates. Boys probably also tend to begin driving earlier than girls. The sixteenth birthday of the average boy is 29 marked by increasing eagerness until he can get his driv- er's license, while a girl is more likely to patiently postpone her license while letting her “boyfriend" do the driving. Specific Predictions The reader can probably think of other cultural sex differences which lead to suspected differences in driving habits. Given these basic assumptions about the nature of the sex differences in driving habits, what relationships between sex and the variables of interest can be hypothe- sized? To present an answer to the question, each of the classes of variables will be examined separately. Time variables Since it is thought that females drive relatively more in the daytime, and males at night, a relationship between sex and both time of day and light condition is expected, with females having relatively mere accidents in the daylight hours and men having relatively more acci- dents at night. An exception to this is predicted during the rush hours (7 - 9.A.M. and h - 6 P.M.). Men are ex- pected to have relatively more accidents during these times than during the rest of the daylight hours than fe- males. A further prediction is that female drivers have relatively more accidents on weekdays, with men taking precedence on the weekends. 30 Road Characteristics The theory based on exposure predicts that female drivers have relatively more accidents on city streets. Hale drivers, on the other hand, are expected to have rela- tively more accidents on U. S. and State highways. It is not clear what to predict concerning county roads. It could be that many of the short errand-type trips that the female is suggested as frequently making is in large part on county roads. Many suburban streets are actually considered county roads because they exist outside the city limits. As for road geometry, since the more urbanized areas are characterized by a preponderance of straight, level streets and roads as opposed to the more often curved and graded rural highways, it is predicted that females have relatively more accidents on straight, level roads, with the converse being true of male drivers. Location Hale drivers should have relatively more accidents occurring in rural areas according to the above reasoning. Because intersections are more common in urban areas, it is expected that relatively more of the accidents of female drivers are at an intersection. Collision Characteristics It is hypothesized that females are associated with intersections, which leads to the expectation that female 31 drivers are also associated with "entering from angle“ in multiple vehicle accidents. Since same direction and oppo- site direction accidents can take place at locations other than intersections, male drivers might be expected to be slightly higher proportionately on these categories than female drivers. In single vehicle accidents, the pedes- trian category is difficult to predict since pedestrian accidents are probably associated both with night and urban locations; thus leading to opposite predictions. Therefore, no prediction can be made with any certainty. An association between sex and fatal accidents is ex- pected, with males being relatively higher in the fatal cat- egory. This prediction is made on the strength of the ex- pectation that males drive more at night, after having con- sumed alcohol, at higher speeds, and on rural roadways. All these variables contribute to conditions in which a fatal accident is more likely to happen. Vehicle Characteristics It is straightforward to predict that females are involved in relatively more accidents involving passenger vehicles with the opposite, of course, being true of male drivers. Driver Characteristics It would seem that male drivers would tend to have relatively more accidents during the younger years than fe- ‘1 32 males, perhaps up to age twenty-five, according to rea- soning based on assumed exposure differences. It is also expected that inexperienced males have relatively more acci- dents than inexperienced females. In terms of registration, since males are assumed to do relatively more driving on long trips, the prediction follows that they are found rel- atively more frequently in the 'outstate" category. Driver Behavior Since speeding violations can occur any place at any time, no prediction can be made based on exposure for this variable. Bight-of-way violations probably are most common in urban driving; thus, it is predicted that female drivers are ticketed relatively more often for this violation. No prediction is made concerning following-too-close violations, while male drivers are expected to be charged with passing violations relatively more often than female drivers because exposure predicts that males do more rural driving. Again the expectation that females do more urban driving leads to the prediction that they are involved in relatively more turning violations. Consideration of exposure leads to the assumption that male drivers tend to drive at higher speeds, and, hence, to the prediction that they are involved in acci- dents at higher rates of speed. The assumption that males tend to drive relatively more at night leads to the pre- diction that males are guilty of sleeping behind the wheel 33 relatively more often than females. Exposure also directly leads to the prediction that males are relatively more often guilty of having consumed alcohol prior to an acci- dent. Weather variables Because the weather conditions are largely independ- ent of time of day and location, i.e. the rain falls on everybody, and so does the snow, only a prediction of a small relationship between weather and sex, and surface condition and sex can be made. The fact that the feminine role dictates that females be a little more hesitant to drive under adverse conditions, as well as being more often able to wait until the road and weather clears, suggests the prediction that female drivers are relatively less likely to be involved in an accident in rainy, snowy, or foggy conditions. CHAPTER II THE DATA The accident records maintained by the Michigan State Police include all reported motor vehicle traffic accidents occuring in the state excluding the city of Detroit. The sample selected for this study consisted of all reports of motor vehicle traffic accidents contained in the Michigan State Police files which occured during the years 1966 and 1971 in Berrien County, Michigan. The data consisted of 1,909 single vehicle accidents and h,250 multiple vehicle accidents for 1966, and 2,167 single ve- hicle accidents and h,820 multiple vehicle accidents for 1971. The discriptive variables regarding motor vehicle traffic accidents, which were selected as the basis for this study, were determined by their appearance on the standard Michigan accident report form. Prior to this selection, a personal observation of the reports was made at the Michigan State Police accident records section in lensing to verify their completness. Consultation with personnel at the records section provided assurance that the reports offered the most reliable official motor vehicle traffic accident information available. 35 Twenty-nine variables in eight separate classes taken from the accident report form were found to be on record in usable form. These variables are: Time variables Light condition Time of day Day of week Road Characteri_s_t_;g_3_§ Highway classification Road geometry Road surface Road defects Location Intersection Locality Collision Characteristics Directional analysis Fatal Vehicle Characteristics vehicle type Vehicle defects Vision obscured .Qgiver Characteristics Sex Age Experience Registration 36 Driver Behavior Speed violation Bight-of-way violation Following-too-close Passing violation Turning violation Traffic control violation Violation other than speeding or drinking Alcohol Sleep Speed Heather variables Heather Surface condition (A list of the categories within each variable is found in the Appendix.) The data were collected and recorded by the State Police for their own purposes and not for the purposes of this study. Therefore, some of the variables recorded are not as relevant as one might hope, some are not grouped so as to gain maximum information, and some variables which would be of great interest are not included. An example of such a deficiency is that the age variable was grouped so that drivers between the ages of twenty and twenty-five were placed in the same category. This grouping is unfor- tunate since the legal drinking age (during the two sample years) falls right in the center. Thus, alcohol is a large 37 factor in part of the group, a smaller but perhaps not neg- ligible factor in the other part, and there is no way to separate the two. 0n the other hand, the accident records kept by the Michigan State Police contain a great deal of information, and care is exercised in the collection and assessment of those records. This is definitely not second-class data, particularly in relation to that collected by other states. Despite the care which is taken by Michigan law en- forcement organizations, there is a bias which can creep into any accident record. Since only one person fills out the report for a given accident, no reliability check is available. This makes the data in general somewhat sus- pect. Specific to this study, an implicit assumption which the investigating officer may have concerning sex differ- ences in driving abilities or habits can influence the way he fills out his report.”1 For example, the implicit feeling that female drivers are less likely to be under the influence of alcohol can lead to a smaller likelihood of a female being Judged in such a condition than a male driver, other things being equal. This is a very subtle bias, the effect of which can not be assessed. It may very well be that some results which will be reported are merely a result of the working of this bias and have no actual basis in reality. ulfladdon, et al., op. cit., p. 139. CHAPTER III METHOD OF ANALXSIS Chi square contingency table analyses were applied to the data in order to compare the relative proportions of each sex within the categories of each variable. A signif- icant chi square value indicates dependence between the two variables in question, which is equivalent in the present study to saying that the proportions of the male drivers are statistically significantly different from those of the female drivers. Let it be emphasized that the comparisons between the sexes throughout this report are relative to the total number of accidents for each sex. The fact that males had almost three times as many accidents as females makes a straight comparison meaningless. For each variable crossed with sex, separate analy- ses were done for each year. Within years, the data were analyzed separately for single and multiple vehicle acci- dents; and within multiple vehicle accidents, analyses were done separately for driver one and driver two. Thus, in all six separate analyses were done for each variable. Separate analyses were done to permit an examination of the results in terms of replicability. To illustrate, a significant relationship, as indicated by the chi square value, in the data from one year which did not hold up for 39 the other year would be considered suspect, unless a trend in the same direction was evident. On the other hand, com- parable results for both years added confidence to the con- clusion drawn from the data. In all cases, single and multiple vehicle accidents were analyzed and discussed separately. There are several reasons for this. It allows drivers in the case of multi- ple vehicle accidents to be analyzed separately. If all accidents were grouped together, this could not reasonably be done. Also, some categories within variables are appli- cable to only one type of accident. For example, for sin- gle vehicle accidents only the ”pedestrian“ and “single vehicle" categories are relevant in the "direction analy- sis" variable, while for multiple vehicle accidents only the other five categories are relevant. But perhaps most importantly, single and multiple vehicle accidents representtwo distinct types of acci- dents. This is true in terms of what actually happens; i.e. a single vehicle accident can be accomplished by one vehicle; it takes at least two to have a multiple vehicle accident. It is also true in terms of relationships be- tween sex and other variables. The reader will be able to note cases in which the relationship between sex and other variables is not the same for both types of accidents throughout the results to be reported in this study. The data from driver one and driver two of the same year are actually not completely independent since a pair #0 of drivers is involved in each accident. However, credence is added to any conclusion drawn for multiple vehicle acci- dents by considering the data from both driver one and driver two. The dependency within years merely makes com- parisons of the results of both years relatively more im- portant and consistency between drivers of the same year somewhat less convincing than if they were independently sampled. A significant chi square value gives no information about the direction of the relationship. Once a statisti- cally significant relationship has been determined, the data must be further examined to discover the nature of the relationship. In this study, this was done by com- paring proportions for those variables with significant chi square values. Within each sex the number of acci- dents in each category was divided by the total number of accidents in which a member of that sex was the driver. By comparing these proportions for the six (or fewer) separate analyses, conclusions were drawn as to the nature of the relationship between sex of the driver and the variable in question. For some variables, further analyses were accom- plished by looking at the relation of sex and the variable of interest within urban and rural accidents separately in some cases, and within day and night separately in other cases. The reason for doing this was simply to assess the effect of either the location of accidents or the time of #1 the accident upon the relation between sex and the variable of interest. Some differences were expected to diminish while others were expected to accentuate by these further analyses. CHAPTER IV THE RESULTS _ The results of the chi square analyses are found in table 1. Time variables All of the analyses with the "light condition” vari- able were significant. The relative proportions are pre- sented in table 2. In every case female drivers have pro- portionately more accidents in the daytime, while male drivers are relatively higher at night and at dusk and dawn. The light - dark difference is much more pronounced for single than for multiple vehicle accidents. Consid- ering the consistency of these results it can be concluded that the differences found between the sexes on the ”light condition" variable are statistically reliable. When light condition was broken down into rural and urban locations, the proportions of accidents for each sex were very similar to the proportions in the entire sample. Thus, location did not appear to influence the relationship between sex and light condition. These results are found in table 3. The highly related ”time of day” variable exhibited a similar consistency. The results are presented in table h. For single vehicle accidents in both years, male driv- “3 mom.“ use.na sa.a oem.a .ee.a so. sensuous nao.e eaa.fia os.n an.m No. on.H sensuousopca 8383 cc. «a. ma.n «we.» emm.sH HH. upooeoe econ Ha. do. smm.s Ho. so. am. oceans» econ m~.m nm~.oa me. n~.n oHo.nH eam.an nauseous econ os~.- o~a.nd oaa.na omm.HH ooH.~H mm.s codsmodeswmsao sarcasm noapnaaopocacao doom an~.n owe.a eea.:fl ea.~ oam.m emm.nH aaosuomoaao N. sun nma.oa som.nH pne.ma ¢m~.na nmo.ma ona.H~ ”sustenance av emm.~e osm.o~ eso.os pmn.ma eos.aHH em~.oaa sue co case eem.- eme.ma oNH.oH oom.m e-.Hm e~s.~o~ consaeeoo sauna QOHDGAHI> 2.59 N an H as N to H to Haaa mesa uqm¢Hm¢> oHaduasz sawsam gmfimsa E80 Add 29H: Gammomo “an mom wags mozaHmquHm 924 mgqfia flmgom Hmo H 3mg Hm.H pon.e HcHe.~H «NH.s mo. Ho. coHpuHoup aasucoupstm e~o.eH eH~.H~ «as.s eHo.nH em:.o~ e8.5 :oHpuHoHp noose aoHpezom noaHaa mm.H oH.H Na.n Hs.~ aH. oun.oH coHuuuuuHmom u--- ---- ee~.an esH.em u--- eHm.eH oonoHnoauu umm.H~ ~e.a no.3 Hm.s on~.aH ea~.a~ use moHpmHaouoeacso aobHaa mH. nH. mo. mo. so.H neo.e eousomno :oHuH> me. om. mH. m~.H as. on.~ upooeoe oHoHsos eHn.on eam.om esm.an cam.mn e~m.om emH.nm can» oHonos eoHanaoaosacco oHoHao> om. mn.~ Ho.H «a. mm.~ ao.~ Hausa dmm.HH eoe.e~ ems.o~ ooH.aH om.H mH. anaHaee HuaoHpoosHa moHpmHaoposaszo soHnHHHov N an H as N to H as HaoH eemH mum oHaHsHsz oneHm wmqdem<> amuse Add maHz Qmmmomo Nam mom mqm>ma HUZ Monaco: eHo.o~ eo~.~s ens.a~ eeo.mn e~o.~m eem.sa eooam emu.“ eem.0H aoa.= oa.~ emH.- eno.~H noon eno.Hs ene.ss em~.~n eem.sn eNo.anH eas.n~H HoeooH< . . . . . . mstsHac no waHeooae am: e «an a aaH m mm as H cum sH can» sense :oHaaHoH> am.~ aa.~ mo. no., No. so. :oHpoHoHp Hosanna oHeeana HH. me. OH. 5H. am. we. :oHusHoab wsHsaaa em. mH.H aH.H so. Ha.m aH.m :oHpaHoHp wcHaasm am. so. eH. oH. sa.H no. ouoHo-oouumeHsoHHoa m an H as N to H as eemH HamH memH sameHmas oaadaasz oawaum mmademds mmmao 444 maHz Qummomu Nflm mom mqu>mq ”CaduHthuHm 92¢ mmDAdS madSGm Hmo AooasHpaoov H mqmda #6 TABLE 2 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF LIGHT CONDITION Single vehicle 1966 1971 Female Male Female Male Dark 25.0 50.8 33.0 53.5 msk-dam 501‘ 507 305 309 Multiple Vehicle 1966 1971 Female Male Female Male Driver 1 Light 69.9 61.6 70.0 60.1 Dark 25.3 32.8 26.5 3h.l DDSK-dam “08 5.6 305 508 Driver 2 Light 67.1 58.1 71.7 58.9 Dark 29.5 37.7 25.2 37.0 BUSH-dam 30“ ”.2 3.1 ”01 “7 n.n v.3 o.a e.~ Q.“ a.m o.n o.~ masque-no o.~n «.mn n.oH m.e~ o.Hn s.mn m.nH H.a~ sung H.ae «.en o.es m.oa «.me «.mn H.Ho a.oa ageHH m uspHua H.n 0.: o.m o.n a.m o.m «.3 H.“ saueuauan ~.~n n.an ~.e~ e.on m.m~ m.an H.n~ H.cm suns 5.3 Tan 92. 3.3 «.3 «in 0.2. min 3qu H nopHsn Hausa n88 Hausa ears Haas 53.5 Hausa 88.5 oHua oHuaoa oHus oHusoa HamH emaH oHoHso> oHaHuHax a.n n.a m.n n.m a.n m.n N.“ n.m caucus-so ~.an c.ma a.nn m.Hn s.an m.as H.m~ ~.o~ saga H.an n.as e.~e «.mm m.an a.aa a.ne m.aa aemHH Hausa sense Hausa ensue Hausa naps: Hausa cease oHsm aHssoh oHe: eHusom HaaH omoH aflodneb Odwndm monaduod szBH: ZOHBHQZOU HmoHA ho Hmoofladu 304M szBHz Nun mud” mom mequHUO<_ho madfizuumum ”NH n Manda #8 TABLE # THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF TIME OF DAY Single Vehicle 1971 Female 1966 Female Male Male 21.01.40 56 2666 62hh51167 11 1111 (Up/0022300700 O 8608000011.“.1 1 1.1111 85”“36539 o o e 0 0 0 mmmuszlzbé 11 1111 a. 23366 718 O O O O O C O C 0 86997253? 211.. 3 AH MM HHHM AA PPPPm 79 1&69 11 AM 2 1 .-.—.-.- mmuwummmm 2 379101.469 1 1 Multiple Vehicle 1971 Female 1966 Female Male Male 210683009 857707u81 11111 52 £393“ 05 526929689 1111 3?“ 5755014 7“ 7777502 1121 2 392726 91 Driver 2 D4. 731.. 37.51 00/ 856707u81 11111 58 768 7532 320236u78 111111 u1u°u222sdQU 75768 7590’2 111 1 5800056u02 0 0 0 0 0 3100011796 111211 ”MN" ”I" AAAA P? P 37911-4692 1 1 .-..-.- "M!” "H AAAAmPPm 7 . H P 2 911.”.69 1.. 1 “9 ers clearly had proportionately more accidents between the' hours of six P.M. and seven A.H. Female drivers were pro- portionately higher from seven A.M. until four P.M., with the remaining two hours being somewhat uncertain. The pro- portions for multiple vehicle accidents were similar, but not as much difference was shown between the sexes. Very similar conclusions can be drawn as were in the case Of the data from single vehicle accidents, although there are sin- gular exceptions. There does appear to be sufficient con- sistency to conclude a reliable difference between the sexes of the nature reported does, in fact, exist. The proportions of rush hour (7 - 9 A.H., and h - 6 P.H.) accidents to total daytime (7 A.H. to 6 P.H.) acci- dents are found in table 5. In all conditions except 1966 driver two, male drivers had a higher proportion of rush hour accidents than female drivers. All of the analysis proved to be significant when considering each of the seven days of the week as a sepa- rate category. With the data dichotomously grouped into the categories "weekend" (Saturday and Sunday) and ”week- day“ (Monday through Friday), all of the analyses were sig- nificant except driver one in 1966 multiple vehicle acci- dents. Although there are some weekdays in which males had proportionately more accidents than females, the general trend was definitely for males to have a greater proportion of their accidents on weekends than female drivers. Table 6 presents the proportions of accidents for each sex on the 50 TABLE 5 THE PERCENTAGE or ACCIDENTS FOR EACH sax WITHIN RUSH noun (7-9 AM a. me pm) AND OTHER DAYLIGHT nouns (9 AH - a PM) Single Vehicle 1966 1971 Female Male Female Male Rush hour 38.8 “2.1 35.7 02.9 Other daylight hours 61.2 57.9 69.3 57-1 Multiple Vehicle 1966 1971 Female Male Female Male Driver 1 Rush hour 38.9 00.9 36.1 37.1 Other daylight hours 61.1 59.1 63.9 62.9 Driver 2 Rush hour 00.0 37.7 36.9 37.9 Other daylight hours 60.0 62.3 63.1 62.1 51 TABLE 6 THE PERCENTAGE OF.ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY 0F DAY OF THE WEEK Single Vehicle 1966 1971 Female Male Female Male Sunday 13.5 20.3 15.1 17.3 Monday 10.6 11.“ 12.2 12.4 Tuesday 11.8 10.0 16.5 11.6 Wednesday 10.9 10.# 13.5 11.7 Thursday 15.8 10.8 11.7 10.5 Friday 18.1 1h.5 13.0 10.2 Saturday 19.3 22.6 18.0 22.3 Multiple vehicle 1966 1971 Female Male Female Hale Driver 1 Sunday 16.6 19.8 12.9 16.9 Monday 13.1 12.0 15.2 12.5 Tuesday 11.6 9.0 9.3 11.2 Hednesday 11.6 8.9 13.5 11.“ Thursday 1h.6 11.6 12.9 12.3 Friday 11.6 16.2 18.0 15.0 Saturday 20.9 22.6 18.2 20.6 Driver 2 Sunday 10.1 20.3 13.5 16.6 Monday 1309 11.9 11.05 1209 Tuesday 11.2 9.0 10.0 10.8 Rednesday 11.3 9.1 13.2 11.7 Thursday 13.3 11.9 17.5 11.5 Friday 18.0 14.9 10.1 16.0 Saturday 18.2 22.9 17.2 20.5 52 different days. Table 7 presents the relative proportion of each sex for the weekday and weekend categories. TABLE 7 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF WEEKDAY - WEEKEND Single Vehicle 1966 1971 Female Male Female Ma le Weekday 67.2 57.1 66.9 60.4 Weekend 32.8 “209 3301 3906 Multiple vehicle 1966 1971 Female Male Female Male Driver 1 Weekday 62.5 57.6 68.9 62.5 Weekend 37.5 42.4 31.1 37.5 Driver 2 Weekday 67.7 56.8 69.3 62.9 Weekend 32.3 43.2 30.7 37.1 Road Characteristics The “highway classification” variable crossed with sex yielded significant chi square values for all of the analysis except single vehicle accidents in 1966. The dif- ferentiating factor for the multiple vehicle accidents was the tendency for male drivers to have relatively more acci- 53 dents on U. S. highways and, to a lesser extent, state highways, while female drivers tended toward relatively more accidents on city streets and, to a lesser extent, on county roads. These results can be seen in table 8. Although the chi square value for 1971 single vehicle acci- dents was significant, the 1966 single vehicle data contra- TABLE 8 THE PERCENTAGE OF.ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF HIGHWAY CLASSIFICATION Single Vehicle 1966 Female Hale Female Male U. S. 16.7 16.4 17.9 18.6 State 15.7 20.3 16.7 14.? County 60.7 57.2 45.1 51.5 City 6.9 6.1 20.3 15.2 Multiple Vehicle 1966 Female Male Female Hale Driver 1 U. S. 24.1 32.4 24.2 28.2 State 21.8 22.3 24.8 25. County 48.? 40.8 29.9 31.7 Driver 2 U. S. 26.2 32.0 20.0 28.5 State 18.6 22.9 22.3 25.9 County 48.1 40.9 36. 30.3 54 dict this result. Therefore, no inference can be made con- cerning the single vehicle data on thisvariable. It should be noted, however, that the “highway classification“ variable is not as relevant to this study as it is to administrative considerations. The highway classification is often not indicative of the type of road- way in question, e.g., a U. S. highway is often a city street. The relative proportions of each sex for each cate- gory of the "road geometry" variable are found in table 9. Both of the chi square values for single vehicle accidents were significant. However, for multiple vehicle accidents only one analysis was significant. The large chi square value for single vehicle accidents in 1966 came mainly from the contribution of the ”curve" category, where females were far below the eXpected value. Female drivers were also below the eXpected value for the 1971 data but not as much. The results of the two analyses were also consistent for the other categories, with male drivers having rela- tively more accidents in the "grade-curve” category and female drivers having relatively more in the “straight“ and ”grade“ categories. Because only one of the multiple vehi- cle analyses proved significant, no relationship for this type of accident can be inferred. When the data were analyzed separately for rural and urban accidents, the tendency for males to have more single vehicle accidents on curves was accentuated in rural 55 TABLE 9 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF ROAD GEOMETRY Single vehicle 1966 1971 Female Male Female Male Straight 67.4 56.3 64.0 62.7 Grade 17.0 14.6 16.7 12.7 Curve 8.3 20.0 10.7 15.2 Grade-curve 7.3 9.1 8.6 9.4 Multiple Vehicle 1966 1971 Female Male Female Kale Driver 1 Straight 74.6 75.5 80.9 74.8 Grade 18.8 16.3 14.5 18.1 Curve 4.9 5.3 3.0 4.2 Grade-curve 1.7 2.9 1.6 2.9 Driver 2 Straight 75.9 75.2 78.9 75.9 Grade 15.4 16.9 15.8 17.4 Curve 6.0 5.1 4.2 3.8 Grade-curve 2.7 2.8 1.1 2.9 locations. A similar rise in sex differences on curves was exhibited in accidents happening during the night-time. The sex differences remained in evidence within urban locations and during the daylight; however, with reduced magnitude. No new information was in evidence from these further anal- yses for multiple vehicle accidents. These results are found in tables 10 and 11. 56 m.n a.H o.m , o. e.n a.H o.n m.a opasonueuac «.0 e.~ a.o o.: o.n o.n n.» n.: «page n.:~ m.n~ a.a~ s.HH 5.0m o.HH ~.H~ n.» cease ~.ee m.am ~.se e.sm 0.05 m.~m m.ae a.mo unmauuum m ao>aan ~.m m.~ n.e w. a.n a.a o.~ a.“ opaso-oesuc «.0 o.n ~.n ~.~ o.e n.s a.e o.m opasc m.m~ o.sa a.a~ m.HH m.aa ~.~H o.e~ s.oa cease a.~o ~.Hm m.oe s.nm m.oa m.~o n.me «.mm snmaaaam H nopauo Hausa scans Assam 53.5 fleas—m naps: Hahsm song: was» oassmm ass: oassom Hams oema afloano> cacapaax m.Ha H.e a.oH m.o n.oa m.n n.oH n.a opusouoesao «.ma a.m m.HH m.m o.- m.n~ ~.o v.5 opuso s.nH a.HH m.a~ n.aa e.eH m.m o.ma H.0H cease n.mm m.~a o.nn o.na m.om a.mo o.ae o.Hm unwaaaam 53.5 Hanan :33 Hanna some: The :33 can: 330m can: cameo.“ Haofl moma afloano> oneam monedooq szBHz Magomc 03m mo Nmoomadu mug szfiHz Nam mud” mom mBZMDHUod mo mcazmommm ma. 0H mama 57 :.~ n.n o. o.a o.~ ~.m o. m.n opasouoeauo 5.5 s.: m.: 3.: 5.m m.n H.m n.m opuso m.ea 5.5a 5.n~ n.~a 5.5a n.oa n.HH 5.0a means ”.05 v.35 o.o5 n.~m 3.05 5.35 w.om m.e5 passage» 5 papaya n.~ n.n m.~ m.” 5.H m.m n.H H.~ opnso-oeauc o.n 5.: n.~ m.n ~.s a.» m.n H.s opnso e.oH 5.5a o.m~ o.ma 5.5a n.na m.~a m.o~ cacao H.05 H.a5 n.om 5.Hm o.e5 .5.e5 H.5m n.M5 unmdaupm H nopaaa sumo swag saga swan saga swag acne pnwaq can: nausea can: camsom H5ma moms oaoanoa oaaduas: ~.HH m.5 ~.m m.m ~.Ha n.e n.HH n.a opusouoeaac :.5H a.5 m.HH ~.oa ~.5~ o.- o.~a m.5 ouuso e.HH m.nH ~.5H ~.ea m.oH n.5H o.m m.mH cease m.5n c.5e 5.0m m.se 0.0m s.~e 5.me 3.5m unmaaupm anus panda anus arena snag unmaq mesa usage can: camsom can: oaasom H55a mean afloa£o> onCam mZOHBHono BmUHA szaHz Hmfimxomo Q oflaasasz monadqu szBHa onBommmHBZH mo Hmoomnko modm szaHz 8mm mug mom mBZMQHOU< mo $04928me an. +2” mama 62 TABLE 15 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF LOCATION Single Vehicle 1966 Female Male Female Male Urban 30.2 29.6 47.9 44.8 Rural 69.8 70.4 52.1 55.2 Multiple Vehicle 1966 Female Male Female Male Driver 1 Urban 48.7 40.7 75.3 67.1 Rural 51.3 59.3 24.7 32.9 Driver 2 Urban 45.5 41.3 73.8 68.1 Rural 54.5 58.7 26.2 31.9 63 0.00 5.:0 0.00 0.50 0.:0 :.00 0.50 5.50 00000 0.05 0.00 0.00 :.05 0.0: 0.00 0.00 0.0: 0.000 0 000000 0.00 0.00 :.00 0.00 0.00 0.00 5.:: 0.00 00000 :.05 0.00 0.05 0.:5 :.0: 0.50 0.00 0.0: 00000 0 000000 00.0 00000 0000 00000 0000 00000 0000 00000 000: 000000 000: 000900 0000 0000000 00000000 0.00 5.00 0.:0 0.00 0.:5 0.00 0.05 5.00 00000 0.0: 0.0: 0.0: 0.0: 0.00 :.:0 0.00 0.00 0.000 00.0 00000 0000 00000 0000 00000 0000 00000 add: 00000.0 0002 00000.0 0500 0000 0000000 000000 mo Hmoumedu mo; szBHz Nmm 84m mom mezmQHoo< mo mo 0H0: pant mZOHsddoq ZHMBHS mHqua ZOHBOMMHQ mo Hmoomedu 54” 55.5”: Nam mud” mom mazmn—Hbod mo mosazmommm mun. ma mama 67 TABLE 19 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF VEHICLE TYPE Single Vehicle 1966 Female Hale Female Male Car 97.8 89.0 97.2 6.0 Pickup 1.“ “.7 1.5 5.8 Truck .0 3.5 .5 u.“ Other .8 2.8 .8 3.8 Multiple Vehicle 1966 Female Male Female Male Driver 1 (hr 970“ 86.0 9900 505 Pickup .9 4.1 .6 5.9 Truck .3 7.6 .0 7.0 Other 1.“ 2.3 .b 1.6 Driver 2 car 97.0 85.6 98.8 6.1 Pickup 2.1 u.3 .3 “.5 Truck .6 6.8 .0 6.1 Other 03 303 09 3‘3 68 Driver Characteristics The analysis of single vehicle accidents for both years were highly significant when the variable ”age” was crossed with sex of the driver. The proportions for each sex in each category are presented in table 20. It can be seen that the differences between the two sexes were in a similar direction for the two years. Female drivers had proportionately more accidents between the ages of 25 and 64, while male drivers took relative precedence for 28 and under and very slightly for 65 and older. Since only one of the multiple vehicle analyses demonstrated a significant relationship, no conclusion can be inferred from the mul- tiple vehicle data. When age was analyzed within locations, and within light conditions nothing with any consistency was exhibited for multiple vehicle accidents. For single vehicle acci- dents, males between the ages of 20 and 2h consistently had proportionately more accidents than females of the same age. These results are found in table 21 and 22. Unfortunately the accident records from 1971 did not contain data on the driving eXperience of the drivers. However, such data were collected as a part of the 1966 records and were analyzed as a part of this study. All three analyses showed a significant relationship between experience and sex. The results in terms of relative pro- portions are found in table 23. All three tables show a 69 TABLE 20 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF AGE Single Vehicle 1966 Female Male Female Male "' 19 1503 18.“ 18.0 1802 20 - 24 15.6 23.2 16.0 21.6 25 - nu nu.2 “0.7 40.2 39.2 #5 - 6h 21.6 13.7 22.8 16.8 65 - 3.3 “.0 3.0 “.2 Multiple Vehicle 1966 Female Male Female Male Driver 1 - 19 8.8 11.“ 10.7 12.6 20 " 21‘ 15c? 15.]. 1806 15.“ 25 " “a 5105 “6.7 no.“ “101 05 - an 19.2 20.2 25.u 23.9 55 - h.8 6.6 u.9 7.0 Driver 2 - 19 10.9 11.5 10.8 1508 25 ‘- ”a 50.1 “5.2 “8.8 “0.2 “5 " 6b 20.“ 21.3 2303 23.8 65 - ”on 6.9 1.9 509 70 oHodno> oawcdm a.» A.n :.N A.H m.o v.3 H.e w.~ - me a.w~ ~.n~ w.m~ n.- H.- H.o~ ~.- o.u~ am - :3 «.mn «.H: :.mn m.ma o.oa 3.“: «.ma n.:n a: - “N A.ma o.=H H.5H m.aa A.na A.ma :.:H a.nH an - om o.na a.na “.ma m.na A.m ~.nH H.HH n.HH ad - N nopdua o.s H.A n.A o.: :.A n.m n.m m.n - me H.a~ A.n~ ~.:~ :.wm ~.H~ v.0“ n.n~ H.:H so u n: o.H= H.~a o.na m.on 5.:: a.o: o.ma H.5m a: . mm :.AH n.3a ~.aa n.m~ “.ma m.:a o.oa a.nH aw . ow n.o~ m.n~ H.a a.- o.HH a.~a o.» o.o oH . H nopaua 3:5 :33 Hausa 93.5 Hausa 53.5 Hanna 53.5 9nd: wagon can: oadfiom flea” oHodso> oaaauass H.n a.“ a.~ :.a o.n m.a m.~ n.m u mm «.6H A.wa m.ma m.m~ m.nH m.na m.- m.m~ am . ma «.mn o.on o.na a.mn m.on a.na A.~: u.na a: - mu m.n~ n.ofi «.5H w.ma A.m~ m.- m.mH A.HH aw . ow o.oH n.0a m.na m.o~ o.aa 5.9” n.0H m.mH ma . dram 53.5 Hausa 52.5 «9:5 53.5 756 nuts 3nd: 3 mac.— one: oamaom anon mead mo HMOUMHQD mo: ZHZBHz Nam mug mom mBzmQHoo< mo mugzflommm mun. AN an: mzo Hadooq 22.3.; med 71 o.~ 5.5 H. o.~ H.n «.5 n.H :.m u no o.mH ~.e~ o.ma H.3N “.ma w.H~ A.~N o.o~ am u m: o.o: o.o: 0.3m a... n..: n.:= :.N: m.Hm a: - nN m.ma n.nH m.- A.~H m.mH «.ma m.mH n.~H am . om m.o~ o.~a H.:~ m.:a :.NH m.HH o.na m.oH ma . N hmbdhn m.n o.a o. a.m m.m H.m o. m.m - mm ¢.ma m.o~ :.HN o.m~ m.ma «.mm «.ma ~.o~ so u m: 0.:: m.mm H.32 n.mm A.Hm H.:: 0.0m m.w: a: . mm m.AH ~.:H :.HN m.mH nJeH n.3H H.3H m.mH am . om o.mH A.HH H.nfl o.m m.- n.0H :.m H.@ ma . H H258 xamn unmaq gang unwdq xuua pswdq xuna uzwdq ode: 3.28m odd: oadaoh Mama coma oaodso> canapasx H.~ 6.. :.N w.m ~.H v.0 a.~ H.: u we ~.HH ~.~N m.mH m.H~ o.oa :.ma o.AH m.m~ :o u m: A.Ha o.mn :.am m.o: o.~a ~.H: A.Ha 5.“: a: . mm H.n~ o.AH m.o~ 0.:H m.u~ m.AH N.o~ ~.~H am . om m.ma a.ma m.mH o.mH o.o~ n.0a m.AH m.:~ ad . xuua psmdq anus pnmdq mama pgmdq xnmq unmdq can: camaom can: oHuaom aha” mead 0H0a£o> oawcdm mZOHBHono HMOHA ZHmBHz mu< mo Hmoomsdo mo; ZHMHHB xmm $04M mom mazmnHoo< mo MOS—.zmommm mus NN mama 72 strong consistency, with female drivers of less than five years eXperience having a disproportionately large number of accidents compared to male drivers with the same amount of eXperience. It is unfortunate that the relation could not be tested by replication. However, in the face of the consistency of the 1966 data the inference that a true re- lationship as described above does exist seems reasonable. TABLE 23 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF EXPERIENCE Single Vehicle 1966 Female Male Less than one year 8.0 5.0 One to five years 35.9 28.8 More than five years 56.1 66.2 Multiple Vehicle 1966 Female Male Driver 1 Less than one year 3.8 2. One to five years 32.6 19.0 More than five years 63.6 78. Driver 2 Less than one year 6.0 2.7 One to five years 29.1 17.6 More than five years 6h.9 79.7 73 For the 1966 sample of single vehicle accidents, a tendency was shown for inexperienced females to differ from inexperienced males more extensively in urban settings than rural settings. This was not the case in multiple vehicle accidents. These results are found in table 2h. Likewise the same sex difference is more pronounced in the daytime TABLE 2“ THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF EXPERIENCE WITHIN LOCATIONS Single Vehicle 1966 Female _ Hale Urban Rural Urban Rural less than one year 8.1 7.8 h.l 5.3 One to five years 36.b 33.8 25. 30.0 More than five years 55.5 58.8 70.3 6b.? Multiple vehicle 1966 Female Male Urban Rural Urban Rural Driver 1 Less than one year 2.8 5.1 2.5 1.9 One to five years 31.2 32.6 19.3 19.2 lore than five years 66.0 62.3 78.2 78.9 Driver 2 less than one year 6.“ 5.8 3.3 2.3 One to five years 33.6 2b.? 20.6 17.“ lore than five years 60.0 69.5 76.1 80.3 71; than at night for single vehicle accidents. The difference between day and night is much less pronounced in multiple vehicle accidents. These results are presented in table 25. To help clarify the age and experience relationship, proportions of each sex and age level within levels of ex- periences are found in table 26. It can be seen that mid- TABLE 2 5 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF EXPERIENCE WITHIN LIGHT CONDITIONS Single Vehicle 1966 Female Male Light Dark Light Dark Less than one year 9.1 . 6.1 h 6 8.3 One to five years 33.9 “0.0 21.9 3h.3 More than five years 57.0 55 a 72.0 61.h Multiple Vehicle 1966 Female Male Light Dark Light Dark Driver 1 Less than one year 1.9 8.2 2. 2. One to five years 33.0 26. 18.1 20.1 More than five years 65.1 65.3 79.6 77.7 Driver 2 Less than one year 6.2 6.3 2.7 3.1 One to five years 27.9 35.9 17.1 21.5 More than five years 65.9 57.8 80.2 75.h TABLE 26 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF AGE WITHIN EXPERIENCE Single Vehicle 1966 Male Female NONI-'3 e #00“ HWN oomaom .0... Oxééu-i 473' “MN “m e QNQHH {\u—l - 19 20 - 2h 25 - an #5 - 6h 65 - 75 Multiple Vehicle 1966 hale Female Driver 1 OQBO: e 11\wa ' ‘ON m-fimm eeeee O‘VVOG) HM!“ (\HNOO -19 zo-2u 1.5-ea 65- Driver 2 25-1.1. “04‘0““ ounces "\N «who: 0 \OHO\N 3:3 monnn NWNI‘N (I) MRMMN “HM -19 20-2u zs-uu 1.5.5:. 65- 76 dle-aged females have relatively more accidents than middle- aged males within the ineXperienced category. This differ- ence is quite large. This is true for both single and multiple vehicle accidents. The only significant analysis of the variable “reg- istration” crossed with sex was of the data from 1966 sin- gle vehicle accidents. Table 27 presents the proportions of accidents for each year. TABLE 27 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY 0F REGISTRATION Single Vehicle 1966 ' 1971 Female Male Female Hale Instate 98.6 95.2 94.3 9b.? Outstate . 1.9 “.8 5.7 5.3 The relationship in the 1966 single vehicle data shows females to be associated with 'instate'. However, the 1971 single vehicle data contradict this result; hence, no conclusion can be drawn. This likewise is true for the multiple vehicle data, since none of the analyses were significant. 77 Driver Behavior The variable “speed violation” crossed with sex yielded significant chi square values for all six analyses. The results in terms of proportions are presented in table 28. In all cases male drivers have relatively more acci- dents involving a speed violation than female drivers. The relation is not as strong for multiple as for single vehicle accidents, nonetheless it can be assumed to be reliable for both types of accidents. TABLE 28 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF SPEED VIOLATION Single Vehicle 1966 1971 Female Male Female Male Speed violation h1.8 52.5 37.“ h9.h No speed violation 58.2 “7.5 62.6 50.6 Multiple Vehicle 1966 1971 Female Male Female Ma le Driver 1 Speed violation 11.6 19.9 7.3 15.6 No speed violation 88.h 80.1 92.7 85.h Driver 2 Speed violation 8.6 12.7 7.5 15.2 No speed violation 91.“ 87.3 92.5 8h.8 78 The breakdown of the speed violation proportions within the sexes for urban and rural accidents is found in table 29. It can be seen that the relationship between the two variables is unaltered by this analysis. The analysis within light conditions for single vehicle accidents did have the effect of reducing the sex difference, though not eliminating it. The relationships for multiple vehicle ac- cidents were not altered to any degree. These results may be found in table 30. Table 31 contains the proportions of the variable 'right-of-way violation“ crossed with sex of the driver for multiple vehicle accidents. No significant relationship was found between these two variables in the single vehicle data, while all data except driver two for 1971 yielded sig- nificant chi squares in multiple vehicle accidents. On all multiple vehicle analyses there were preportionately more accidents involving a female driver and a right-of-way vio- lation than accidents involving a male driver and the same violation. Thus, a reliable relationship can be inferred for multiple vehicle accidents. When multiple vehicle ac- cidents were examined for urban and rural accidents sepa- rately, the relationship between the two variables remained very similar. These results can be seen in table 32. The analyses of sex crossed with the following vio- lations all yielded non-significant chi square values both for single and multiple vehicle accidents for both years: 79 e.so o.oa o.om e.mm «.mm H.oa «.mm e.no cascades» econ» oz e.na o.oa o.aa a.e m.aa m.m a.aa e.a caucuses» ocean N uo>ano m.am 0.5m a.em 6.5m e.oa 5.55 «.50 H.mm canvases» coon. oz «.ma o.na n.na s.m e.n~ n.ma m.~H m.oa censuses» scene a .8» a .5 Hanan :85 H83 :85 8.5m 93.5 75m :85 can: cannon can: causes Ham” moo” oaoazo> oaaauasz n.5n a.ne H.5m n.aa m.mm ”.00 H.~m m.ea censuses» econ. oz m.ae n.0n m.me a.m~ m.ae a.mn m.as n.n~ caucuses» eooam Adam SBA: Huh—.5 flanks H ahfim £385 H 93m SBHD 3s. 33....“ can: 3an Ham” coma encano> on:«m monaduoq szBHB on9 Qmmmm mo Nmoomedo modm ZHMBHx xmm modfl mom mBzmoHood mo mo oaaapaax a.an n.ao n.3n a.ee n.mn H.5m a.as a.oo coauuaoa> soon» 0: m.~e u.mm n.n: n.nm u.am 0.0: n.mm n.mn ceapeaoab eooam saga unmaq same scene sumo sands suns unmaa east 3 gem and: Shaun weed oaoaso> oawnam mZOHBHono BmoHa szBHI ZOHH¢AOH> Qmmmm mo Hmoomadu mo<fl szBHz Nam mudw mom mszmQHoo< mo m0<fizmommm mma on Mama. 81 TABLE 31 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF RIGHT-OF-WAY VIOLATION Driver 1 Violation No violation Driver 2 Violation No violation Female 77.0 Multiple Vehicle 1966 Female 82 a.:o o.ma o.ma e.aa o.om a.~m o.oa 5.05 ooaouaoao oz m.m~ H.H~ o.- o.- o.aa m.aa o.o~ o.a~ nodooaoap ao3-oouosmam ~ Hoodoo m.nm H.om m.om m.:a o.nm m.~m c.5m o.ma ooaooaoao oz ~.oz o.oa ~.oa «.mN o.mH n.5a «.ma a.e~ soap-Hos» zosuoo-osmam a goofing dram 58.5 255 5.75 :55 :35 The :85 can: oduaom can: ease—em Hams oooa oaoaso> oaoaoaoz anon—”8400...— ZHmBg ZOHBSOE’ mglmOIBmon .mo Hmoomedo mug ZHmBHx “am 8% mom mezmnHUOd mo HUSHZMUmum awn. mm sandy 83 Following-too-close Passing violation Turning violation Traffic control violation Because of the non-significant chi square values, no rela- tion between sex and any of these variables can be inferred. However, it may be faulty to assume that no relationship actually exists since all of the violations were small in number compared to the total number of accidents. Analyses were done by putting all violations other than speeding or drinking into one category, making the other category the absence of any non-alcohol or non- speeding violation associated with the accident. Table 33 contains the preportions of each sex for each category. The analyses for single vehicle accidents were indefinate. The analysis of 1966 data proved significant, but in a di- rection opposite that of the non-significant 1971 data. For multiple vehicle accidents, the differences between the sexes were all in the same direction, with female drivers having a relatively greater preportion of accidents in- volving a violation other than speeding or drinking than male drivers. Even though only three of the four analyses were significant, in the face of this consistency it seems safe to conclude a reliable difference between the sexes of this nature for multiple vehicle accidents does exist. All analyses of the variable “alcohol” crossed with sex were highly significant. The proportions presented in 8# TABLE 33 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF VIOLATION OTHER THAN SPEEDING OR DRINKING Single Vehicle 1966 1971 Female Male Female Male Violation 15.2 9.0 16.0 18.“ No violation 84.8 91.0 8h.0 81.6 Multiple Vehicle 1966 1971 Female Male Female Male Driver 1 Violation 76.1 7h.6 83.0 79.0 No violation 23.9 25.7 17.0 21.0 Driver 2 Violation 80.1 74.6 83.0 77.0 No violation 19.9 25.“ 17.0 23.0 85 table 38 make it very evident that in all cases male driv- ers were found relatively more frequently in both the ”drunk" and “drinking” categories than female drivers. The inference can be made that this is a reliable relationship. TABLE 3“ THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF ALCOHOL Single Vehicle 1966 1971 Female Male Female Male NOt drinking 89.9 6302 87.1 6005 Drinking 9.2 27.7 12.h 29.5 Drunk .9 9.1 .5 10.0 Multiple Vehicle 1966 1971 Female Male Female Male Driver 1 Not drinking 96.2 8h.3 94.h 83.2 Drinking 3.5 11.1 5.2 13.5 Drunk .3 “.6 .u 3.3 Driver 2 Not drinking 95.7 8h.5 96.1 83.8 Drinking 3.6 11.0 3.“ 13.7 Drunk e7 ”05 05 2.5 The relationship remained consistent and strong when accidents were further separated into urban and rural locations, and proportions calculated separately. These 86 results can be seen in table 35. When accidents are di- vided by light condition, males continue to be found more often in the drinking and drunk categories than females; however, the sex difference is considerably stronger in accidents happening at night. These results are found in table 36. The ”sleep” variable crossed with sex yielded sig- nificant chi squares for both analyses of the single vehi- cle data and three of the four analyses of multiple vehicle data. The data in terms of prcpcrticns are presented in table 37. The data from all six conditions are consistent in showing male drivers to be guilty of having relatively more accidents involving sleep than female drivers. It appears safe to conclude that this is a reliable difference between the sexes. The chi square analyses indicate a stronger relationship in the single vehicle data than in the multiple vehicle data. Analyses were not done within light conditions because very few accidents involved sleep in the daytime. All of the analyses of the variable "speed” crossed with sex were significant. The proportions of all six sets of data are presented in table 38. For all data, females have proportionately more accidents at lower speeds, while males take precidence at higher speeds. The significant chi square values, coupled with the strong consistency of the prcpcrtions, leaves little doubt as to the validity of the conclusion that male drivers had a greater propor- 37 H.5 5.« 5. o. 5.« 5.5 5. 5. sense 5.~H 5.5a 5.5 o. 5.5 «.5 5.5 5.5 wosxcaoo 5.55 5.55 5.55 5.555 5.55 5.55 5.55 5.55 55555555 ooz « oopann o.« 5.5 o. 5. H.5 5.5 o. 5. space «.HH 5.5a 5.5 5.5 «.HH 5.5H 5.« 5.5 masseuse 5.55 o.«5 5.55 H.55 5.55 5.55 «.55 5.55 55535555 ooz a .355 :55 :85 . atom 935 :55 :35 3.55 :35 can: canach «an: cannon H55“ 555" oaoaso> oaoaoaoz 5.5 5.5 H.a o. 5.5 5.oa 5. 5.« noose 5.55 5.5« 5.oa 5.«a 5.«5 5.5a H.Ha «.5 masseuse 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.H5 nonsense ooz 1.55 :35 ~55 53.5 3.55 585 3.55 :85 can: cacsom can: nausea a55a 555a oaoano> camoao mZOHB¢UOH ZHmBHI AONOUQ‘ mo NMOUMBdU mo¢fl szHHz Nam modm web mazmQHoo< ho fledflzmumflm was an num4H 88 5.5 5. 5. 5. 5.5 5.5 5. 5. 55555 5.5« 5.5 5.5 5. 5.55 5.5 5.5 5.5 55555555 5.55 5.«5 5.55 5.555 5.55 5.55 5.«5 «.55 55555555 5oz « 5o>5ao 5.5 5.5 5. 5. 5.5 . 5.« 5.5 5. 55555 5.5« 5.5 5.55 o.« 5.55 5.5 5.55 5.5 55555555 5.55 5.«5 5.55 5.55 5.«5 5.55 5.55 5.55 55555555 ooz 5 no>5uo 5555 5:555 5555 55555 5555 5:555 5555 55555 o5¢a o5uaoa 55.: 5555.5 5555 5555 o5o5no> o55555oz 5.5 5.« 5. 5. 5.55 5.5 5.5 5. 5:555 5.55 5.55 5.5« 5.5 5.55 5.55 5.55 5.5 55555555 5.:: 5.55 0.55 5.55 «.55 5.55 «.55 0.55 wsdxa5ao cc: 5555 55555 55.5 55555 55.5 55555 5555 5:555 ode: 0,528.5 one: cacao." 5555 5555 edodno> Ocham monBHono HmoHA szBHz Homoufld ho Hmoumadu undfl sz8H3 Nun mod” mat mezmQHoo< ho modazmommm ”we 55 55559 89 TABLE 37 THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF SLEEP Single Vehicle 1966 1971 Penn 1e Male Female Male 8109]) 107 505 2.0 701 No sleep 98.3 9h.5 98.0 92.9 Multiple Vehicle 1966 1971 Female Male Female Male Driver 1 Sleep 03 102 .6 208 No sleep 99.? 98.8 99.0 97.2 Driver 2 Sleep .0 1.3 .h 2.2 No sleep 100.0 98.7 99.6 97.8 90 TABLE 38 THE PERCENTAGE OF.ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY 0? SPEED Single Vehicle 1966 Female Male Peule Hale 0 - 10 5.9 5.1 16.“ 1200 ll - 20 12.8 7.8 16.7 12.5 21 - 30 2702 16.0 22.8 160‘. 1 - #0 25.5 23.5 17.0 21.0 51 - 60 600 1 09 8.8 1200 61 - 70 1.7 506 10'. 5.“ 71 - .2 1.3 1.1 1.7 Multiple Vehicle 1966 Female Male Female Male Driver 1 O - 10 12.0 9.“ 17.6 1h.5 11 - 20 16.8 1307 2905 Zlho 21 - O 1505 1109 19.6 1500 1 - O 21.8 1708 1601 1706 l - 50 22.“ 22.1 10.7 15.2 51 " 60 7.2 1605 “.5 8.6 61 " 7O “.0 608 108 307 71 "' .3 108 .2 10“ Driver 2 0 - 10 lh.2 13.2 16.0 1u.3 11 - 20 1909 1 .1 28.1 21.8 21 - O Ines 11.1 21.0 1608 31 " 0 2306 1900 1601‘ 17.8 “1 " 50 1 .9 21.6 1007 15.6 51 - 60 8.1 In.“ 503 8.7 61 " 7O ”.1 5.5 2.5 306 71 - 07 101 .0 10h 91 tion of their accidents at higher speeds than female drivers. The speed and sex relationship was further examined within locations and within light conditions. For these analyses, speed was subdivided into three larger categories. The relationship remained of similar magnitude within loca— tions. Within light conditions, however, the difference between the sexes tended to be reduced. Males nonetheless remained more frequently involved in accidents at higher speeds than females in all conditions. These results are found in tables 39 and no. Heather Variables Both analyses of the single vehicle accidents were significant for the variable ”weather". Only one of the multiple vehicle analyses was significant. The results in terms of proportions are presented in table #1. In single vehicle accidents for both years, females demonstrated a relatively greater number of accidents in rainy and snowy weather, while male drivers were relatively higher in the "clear” and "foggy" categories. The inference can be made that a reliable difference exists between the sexes of this nature for single vehicle accidents. No conclusion can be made for the multiple vehicle data. The proportions of both sexes for the various weather conditions within locations are found in table #2. Snow is definately shown to be more prevalent for females 92 5.55 5.5 5.5 5.5. 5.5 5.5 5.5 5. - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 5 5 no>555 5.55 5.5 5.55 5. 5.55. 5.5 5.5 5.5 - 55 5.55 5.55 5.55 5.5 5.55 5.55 5.55 5.55 55 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 u 5 . 5 uo>5uc 5.5.555 553.59 595555 553.5: 51.5555 553.555 5.5.5.5555 83.5 05w: 0588.5 0555. 055935 5555 5555 o5o5so> o55555e: 5.55 5.55 5.55 5.5 5.55 5.5 5.5 .5.5 u 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 5 Henna saga: Haasm saga: amasm scan: Hanan scan: .55: o5eaom 05¢: o5¢aom 5555 5555 o5o5co> o55:55 monBdUS szaHz Omaha ho Hmoomfidu mug szBHI “an MD: mom MBZMQHUO< mo madezgmmm nun. on 3mg 93 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.5 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 5 5 505555 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.5 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 5 5 555555 5555 55555 5555 55555 5555 55555 5555 55555 055: 055305. 055: 05530m 5555 5555 c5o5no> «555555: 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.5 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55 - 5 5555 55555 5555 55555 5555 55555 5555 5:555 and: oHGBoh can: nausea 5555 5555 udoacob oam:«m ho amoum8 5555555: 5.5 5.« 5.5 5. 5.5 5.5 5.5 5.5 555 5.5 5.5 5.55 5.5 5.5 5.5 5.55 5.5 5555 5.5 5.5 5.55 5.5 5.5 5.5 5.55 5.5 5555 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55050 5555a 5595: 5555a 5595: 5555m 5595: Hansm 2595: 055: 555805 055: o5maom 5555 5555 5555555 555555 mZOHeduoq ZHmBHz mmmeduu mo Hmoumedu mvdu ZHmaHz xmm mo<fl mom mazmQHoo< mo m0<fizm0mflm mus N: mqmdfi 96 in rural accidents than in urban accidents, relative to the male proportions in the two settings. This is also some- what true for the rain category. In the multiple vehicle accidents, snow again appears to be more prevalent for fe- males in rural locations. The proportions for the relationship within light conditions are found in table #3. It can be seen that for single vehicle accidents, light condition made little dif- ference. For multiple vehicle accidents, females were more often involved in accidents in snow in the darkness than males in all four analyses. There was very little differ- ence between the sexes in the daylight within the snow cat- egory. None of the other categories showed much consist- ency. Both chi square analyses of the variable ”surface condition” crossed with sex for single vehicle accidents were highly significant. Two of the four multiple vehicle analyses were significant. The results are presented in terms of proportions in table uh. For the single vehicle accidents, female drivers tended to have proportionately more accidents in the presence of snowy road conditions for both years. This is likewise true of the two multiple vehi- cle conditions which exhibited significant relationships. The two sets of data which were not significant show little difference on snow. Since, however, they do not contradict the relation between female drivers and snow, it seems rea- sonable to conclude, though cautiously, that a relationship 97 5.5 5.5 5.5 5. 5.5 5.5 o. 5.5 555 5.5 5.5 5.55 5.5 5.5 5.55 5.55 5.55 5555 5.55 5.55 5.5 5.5 5.55 5.55 5.55 5.5 5555 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55555 5 555555 5.« 5.5 o. 5.5 5.5 5.5 5.5 5.5 555 5.5 5.5 5.5 5.5 5.5 5.55 5.55 5.55 5555 5.55 5.55 5.55 5.55 “.55 5.5 5.5 5.55 5555 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55555 5 555555 5555 55555 5555 5555 5555 55555 5555 55555 055: 055805 055805 5555555 5555555: 5.5 5. 5.5 5.5 5.5 5. 5.5 m. 555 5.5 5.5 5.5 5.5 5.5 5.5 5.5 5.55 5555 5.5 5.5 5.55 5.5 5.5 5.5 5.55 5.55 5555 5.55 5.55 5.55 5.55 5.55 5.55 5.55 5.55 55555 5555 55555 5555 55555 5555 55555 5555 55555 055: 055505 055905 0.3555055 Odwcam mZOHBHono BSOHA szBHz Ea mo Hmoomedu modm szBHJ Nflm IOdM mom mezmQHUod mo mo<fizmommm 0: Hana nus 98 TABLEM THE PERCENTAGE OF ACCIDENTS FOR EACH SEX WITHIN EACH CATEGORY OF SUREACB CONDITION Single Vehicle 1966 1971 Female Male Female Hale Dry 52.3 68.3 58.5 67.5 Net 19.0 18.9 19.5 17.8 Snow 28.7 16.8 22.0 1“. Multiple Vehicle 1966 1971 Female Male Female Male Driver 1 Dry 55.2 61.5 60.8 57.8 wet 23.0 23.5 23.3 26.9 Snow 21.8 15.0 15.9 15.3 Driver 2 Dr! 59.6 60.6 55.8 59.9 wet 22.1 23.6 23.1 25.9 Snow 18.3 15.8 21.1 1h.2 between sex and road condition does exist, with female driv- ers having relatively more accidents on snowy road surfaces and males having relatively more accidents under dry con- ditions. More confidence can be held in this conclusion for single vehicle accidents. The analyses done within locations and within light conditions showed snowy roads more strongly related to te- 99 males in rural than urban settings in single vehicle acci- dents. This was not consistently true {or multiple vehicle accidents. Snow was likewise more pronounced for the mul- tiple vehicle accidents of females after dark, while the results of the single vehicle accidents were mostly unaf- fected. These results can be round in tables #5 and b6. 100 HoHN moHH moon mobH NoGH ModH NoHN NomH tocm #own NomN mom NoQN noNN doom nonH Hgmm 903 “.00 comm n.mn n.3n “.mm w.~w o.mm h.Hn had N Hubdhfl HoHN Coma howm N.NH momd mooH womm v.0H Iona m.wfl moom NomH 3.3m noflm 3.6m momd oowm a0) Homo Nowm H.3m :.mw Noow Home How“ moan hhfi H Hflhdhfi Hausa 83.5 1928 can»: densm cups: Hanan cops: can: . 3.23m ous: cacao.“ Hfima wwafl onoaso> canduasz mend N.NH mobN 0.MH OohH HomH monm N.ON IOGQ Nowd homH Hofid O.HN hood oomH HoNN h.HH $03 Noam Home Hon“ Nonw b.Nm o.mo Hos: H.mw bud Hdhfifl £6989 Hakim GUARD Hakim CGDHD Hdhfim flunk: can: ode-.3 0H3.— cam-BL Humd 00$H oauano> «Hanan mZOHeduoq ZHZBHz ZOHBHQZOU mo<flm3m mo Hmdumedu m0auo «.aa n.2H n.o~ ~.n~ n.3a n: m.~n m.a~ scam m.o~ o.:~ ~.w~ e.- o.:~ n. - n.0a ~.n~ no: m.nn H.~m n.3n H.no a.~o n.~w o.om m.om use H nopdua snug unmdq saga panda sumo unwaq mama unwaq can: ode—com can: cannon Head coma oao«:o> candpasx a.w~ a. HH m.- «.ma «.BH H.mH o.an o.a~ roam s.mH a. ma m.H~ n.5H ~.o~ 3.5H m.nH n.ma an: m.ao o. as m.nn H.ne a.~w n.5o ~.Hm m.am ago some ocean anaa unmdq mama pnwdq anus ocean ode: nausea can: cacao» Hum” coma «Hodge» named» moneHono Hmqu ZHIBHz ZOHBHQZGO modmmam ho “moomadu mud” szBHz xmm mudm mam WBZMdHood mo m0¢Hszmum ”we w: Manda CHAPTER V DISCUSSION Time variables A very strong relationship between sex and both light condition and time of day for both types of accidents was exhibited by the data. These two variables are, of course, highly related; for time of day in large part de— termines the light conditions, although the correspondence is less than perfect due to seasonal variations and weather changes. The direction of the relationship, males having more accidents in the night-time and females having more during the day, is fully predictable from consideration of exposure differences. These variables differentiated the sexes more in the case of single vehicle accidents. This can be due to the fact that males drive relatively more at night, and single vehicle accidents are associated with night-time driving conditions. Further analysis of the light condition variable within urban and rural accidents showed that the sex difference was not an artifact of loca- tion differences, but rather held up within both types of location. Since the exposure hypothesis predicted real day - night driving differences, location was not expected to alter the relationship. 103 Exposure predicted that males would have more day- time accidents during the rush hours (7 - 9 AH and h - 6 PH) than female drivers, relative to all the daytime acci- dents. The data were consistently in line with this pre- diction. The relationship between sex and day of the week proved to be stable and strong. Likewise, the inference that females have relatively more accidents on weekdays and men on weekends is fully in line with what is expected from exposure considerations. aged Characteristics The results of the variables classified as road characteristics are neither strong nor unambiguous. None- theless, certain interesting relationships do appear. The classification of the highway upon which the accident took place was unrelated to sex in the single vehicle accident data, with no evidence of any consistent pattern emerging. The sane variable showed a significant relationship with sex on all multiple vehicle analyses. The pattern was fairly consistent for male drivers to have relatively more accidents on U. S. and state highways, while female drivers had relatively more accidents on county roads and city streets. The results of the multiple vehicle analyses are consistent with what would be expected by considering ex- posure alone. This researcher cannot see any reason why single vehicle data does not follow a similar pattern. 10“ However, U. S. and state highways which go through and often are the main streets of communities are coded as U. S. or state highways when, in fact, they are city streets. Furthermore, county roads in the suburbs are quite different in characteristics from county roads in rural areas. This makes the interpretation of the highway classification variable highly ambiguous. In the presence of this ambiguity and in the absence of any reasonable al- ternative hypothesis at the present, no real case can be made for rejecting exposure as inadequate in the case of single vehicle accidents. But neither is the absence of a relationship of the type exhibited by the multiple vehicle data totally understandable. There is a definite need for further research or further analysis to explain these re- sults. Analyses of road geometry were nonsignificant for multiple vehicle accidents, however, there was a strong significance for single vehicle accidents. The difference between the sexes was largest in the curve category, with males taking precedence. Males were also relatively higher in the grade-curve category, with female drivers being rel- atively higher in the grade and straight categories. These results could indicate that males leave the roadway on curves, thereby becoming victims of single vehicle mishaps relatively more frequently than females. The exposure hy- pothesis predicts this insofar as male drivers are expected to be found on highways where there are more likely to be 105 curves than female drivers. However, the data on the "highway classification“ variable do not seem to show this to be true in single vehicle accidents although again the variable is quite ambiguous. It could be argued that the difference in the curve category is explainable by the fact that males drive more in rural areas where there are more curves, but when accidents were analyzed only within rural locations the sex difference increased rather than de- creased. The same situation was true for accidents hap- pening at night. A reasonable explanation then becomes that males tend to drive at higher speeds and more often after drinking than females. Both high speeds and alcohol are more prevalent in rural and night-time single vehicle accidents, thus leading to the actual sex difference in the curve category. The sex differences are smaller in urban and daytime single vehicle accidents and absent altogether in multiple vehicle accidents because alcohol and high speeds are not as prevalent in these situations. The alternative explanation is that under the same circumstances, males are less able to successfully nego- tiate a curve than females. In the face of other research cited in this study, namely that in unusual circumstances female drivers display less skill, this explanation seems untenable. Under normal circunstances, drivers do not usually leave the road on a curve unless they are attempting to take the curve at an excessive speed or unless their ability is impaired such as with alcohol. Since males are 106 more often found at high speeds and with alcohol, it would appear reasonable to suggest that the sex difference on curves was caused by these other two variables. No conclusion can be drawn from the data on road surface. There could be a weak relationship there with the small number of accidents on non-paved roads serving to ob- scure it. Exposure does not seem to predict any relation- ship between sex and this variable. There were also a small number of accidents in the presence of a road defect. Despite this fact there is some evidence, though not strong, that female drivers do rela- tively more poorly in the presence of a road defect for both single and multiple vehicle accidents. A relationship of this sort is important because it is not predicted by exposure. There does not seem to be any reason to expect females to be driving in the presence of a road defect any more often than male drivers. This result suggests that female drivers are slightly less able to cope as adequately as male drivers in situations requiring greater than usual driving skill. Location That females have relatively more multiple vehicle accidents at intersections than males is predictable from exposure in terms of the locations in which the two sexes tend to drive. This implies that within locations, the sexes should have a very similar proportion of accidents at 107 intersections. Such was the case for urban accidents, but female drivers were involved more often in intersection accidents within rural areas also. The fact that in urban accidents the sexes were very similar seems to preclude any explanation in terms of driving ability at intersections. It is quite frankly difficult to understand the rural dif- ference. The most reasonable explanation seems to be dif- ferential driving habits of the two sexes. Females may have tended to drive in rural areas containing more inter- sections than male drivers. It does not seem reasonable to reject the exposure hypothesis. Intersections did not differentiate the sexes in single vehicle accidents. This is reasonable since when only one vehicle is involved an intersection is not much more likely to contain an accident than any other compara- bly length stretch of road. Single vehicle accidents are not as dependent upon high traffic volume. Men were shown to have relatively more rural acci- dents than female drivers at a significant level. This result held up for daylight and night-time accidents. The small but apparently reliable difference is in the direc- tion predicted by exposure. Collision Characteristics For the variable "direction analysis” only the four multiple vehicle analyses were significant. The categories "same direction“ and “entering fron an angle“ exhibited a 108 strong consistency, with males being relatively higher in the former, and females in the latter. Since females are more likely to be involved in an accident at an intersec- tion, they would be expected to be relatively higher in the “entering from an angle” category within rural accidents. However, this result was also exhibited in urban accidents. Since most same direction accidents must be rear end acci- dents, males must be more likely to have rear-end collisions than females. The reason for this is not clear. It could be a resultant of the tendency for males to drive at higher speeds; and, hence, be less able to stop in time. It could also be caused by poorer visibility at night, since the data show that same direction accidents are more likely to happen at night than entering from an angle accidents. Since male drivers are also associated with night, this would lead to the relationship which has been found. The male proportion being higher in this category naturally forces the female proportion to be higher elsewhere. This is perhaps why females were higher in entering from an angle for urban as well as rural accidents. Also the data have shown females to be proportionately more often guilty of a right-of-way violation in multiple vehicle accidents than males in urban as well as rural accidents. This could be the cause of the higher female proportion of entering from an angle acci- dents, cr the result of it, since a high correlation be- tween right-of-way violations and entering from an angle accidents is to be expected. 109 The validity of these explanations is less than obvious. However, the competing explanation is that males are less adroit at stopping their vehicles than are females under similar conditions and / or females are more likely to ignore a traffic signal at an intersection than males. This explanation seems perhaps less valid than the one put forth above, which is in line with the exposure hypothesis. The least which can be said is that the evidence is not great enough to reject the hypothesis at this point. There is very little which can be said concerning the ”fatal“ variable. It is predicted that males would be relatively more likely to be involved in fatal accidents than females. Hale drivers were involved in proportion— ately more fatal accidents than females, however, the num- ber of fatal accidents was too small for any analysis to reach significance. Hence, no conclusion can be made. Vehicle Characteristics There seems to be little which needs to be said concerning this class of variables. The high significance of “vehicle type" is, of course, easily explainable in terms of exposure. Females just do not typically drive anything but passenger cars. Both “vehicle defect“ and “vision obscured" were rare in occurrence, and further- more there is little theoretical reason to expect the sexes to differ on either of these two variables. 110 Driver Characteristics Sex and age were unrelated for multiple vehicle accidents, but inexperienced females and experienced male drivers had a disproportionate number of accidents. The same relationship on the eXperience variable held for sin- gle vehicle accidents, but age exhibited a consistent re- lationship with sex, with young male drivers (under 25) and middle-aged female drivers (25 to 6#) having relatively more accidents. The differences in the over 65 category were very slight. This inconsistency between the age and experience yardablea is easily explainable by the fact that females in the sample were much more likely to begin driving at a Imre advanced age than males. This can be seen very plainly in the analyses of sex and age within levels of experience. The tendency for more and more females to begin driving at all probably at least partially explains the sex difference and experience. There were probably a greater proportion of inexperienced females relative to all female drivers in the driving pepulation than the same ratio for males. It is also possible that part of the sex difference on the experience variable is because females drive less than males. If it is assumed that safe driving ability increases as a function of the amount of actual driving time, rather than the elapsed time since obtaining a li- 111 cense to drive, one year of experience for an average male means more actual driving experience than one year of ex- perience for an average female. This would result in the expectation that inexperienced females (in terms of years driving) would have a higher accident rate than inexperi- enced males. The analyses for the age variable leave two unan- swerable questions. First, why are females in the 20 to 24 year age group involved in proportionately more acci- dents than males in multiple vehicle accidents? Such was not the case for single vehicle accidents. Exposure pre- dicts the opposite since it was thought that young males drive disproportionately more than young females. This #2 also contradicts the research findings of Lauer and “3 although as was mentioned, the methods Swanson, et al., by which these studies controlled for mileage are highly questionable. Secondly, what causes the sex difference in single vehicle accidents and not in multiple vehicle acci- dents? The evidence points away from a third variable being responsible. Sex differences were manifested over both levels of location and both levels of light condition for single vehicle accidents. This would not be the case if either of these variables were responsible for the dif- uzlauer , op. cit . “aswanson, et al., Op. cit. 112 ference. Furthermore, since both the presence of alcohol and high speeds are related to rural locations and night- time, if either of these variables were responsible, larger sex differences would be expected at night and in rural locations. Thus, this variable is in need of further re- search. At this point all that can be said is that past research was only replicated in part and the exposure theory was only validated in part. No alternative expla- nation appears readily evident. The small and questionable relationship between sex and registration is easily explainable by exposure con- siderations. The reason for the weakness of the results probably lies in the fact that there were relatively few outstate drivers in the sample. It can only be concluded that these results did not contradict the exposure hypoth- esis. Driver Behavior The strong tendency for males to be more often guilty of a speeding violation than females is not pre- dictable from exposure differences. If the relationship were explainable either because males drive more at night or more in rural areas, the analyses done within locations and within light conditions would be expected to show little or no sex differences. Such was not the case. The sex differences were respectably large in each analysis. Thus, it appears that the tendency for males to be more 113 often guilty of a speeding violation is not explainable by exposure. 0n the other hand, it does not seem reasonable that this difference is explainable by differences in driving ability, i.e., that males drive more poorly when they are violating the legal limit than females when they are violating the legal limit. It seems much more likely that cultural roles, expectations, etc., lead men to do more speeding than women. In many segments of society speeding is considered acceptable and even sometimes desir- able masculine behavior, while accepted feminine behavior does not include driving an automobile at excessive speeds. The previous discussion of intersections and direc- tion analysis is quite relevant to the right-of-way viola- tion variable. If exposure were the explanation for the small relationship of females and right-of-way violations, analyses of these two variables within locations should show little or no sex differences. However, the sex dif- ference did hold up within each location. As with the intersection and direction analysis variables, no expla- nation for these results is readily available. As for the other violations, it is unfortunate, but apparently true that the frequency of such violations was in all cases too small to find any relationships, if indeed any exist. When all the non-alcohol and non-speeding viola- tions were grouped together into one category, with the absence of any such violation in the other, data from sin- 11h gle vehicle accidents were inconclusive, while the multiple vehicle data exhibited a small but reliable tendency for females to have more non-speeding or drinking violations. This is in line with what is predicted from exposure dif- ferences. The reason that the single vehicle results were ambiguous is probably because these violations are rarer in single vehicle accidents and less relevant. The sex differences on the alcohol, speed, and sleep variables were all large and consistent. The sex difference on sleep is easily explainable in terms of the tendency for males to drive more during the late-night hours. The exposure theory predicted males to have a dis- prcpcrtionate number of accidents with alcohol. That this difference holds up within locations and light conditions gives strength to the conclusion that the sexes are differ- entiated by alcohol and the results are not an artifact of another variable. The added strength of the sex difference on alcohol within night-time accidents is no doubt a result of the pronounced tendency for alcohol to be more often present in night-time driving. The tendency for male drivers to have accidents at higher speeds than female drivers also held up within loca- tions and light conditions. Part of the apparent sex dif- ference on speed was shown to be a result of the mutual relation each of these variables had with the light con- dition variable. However, a sex difference was still in evidence within light conditions. As with the case of the 115 related speed violation variable, this consistent differ- ence is not explainable by exposure, but rather males do apparently drive faster than females other things being equal. Again, cultural role differences between the sexes seems most adequate to explain this driving difference. Heather variables Both the “weather“ variable and the highly related “surface condition“ variable showed females to be involved in a higher than expected number of accidents in snow con- ditions for single vehicle accidents and to a lesser extent for multiple vehicle accidents. The results of the analy- sis within locations and light conditions for these two variables generally gave indication of a larger sex differ- ence on snowy roads and in snowy weather at night and in rural locations, with females having a higher proportion in each case. These results do not seem predictable by exposure, for by and large snow is evenly distributed over all times and areas. In fact exposure leads to the opposite predic- tion, since females may be less likely to drive during bad weather conditions. The relationship between female drivers and snow add credence to the previously stated hypothesis that female drivers tend to be more vulnerable to situations requiring more than a usual amount of driving skill. Al- though snow is not infrequent in Michigan, it seems reason- able to consider driving during falling snow or snowy road 116 conditions as presenting an especially stressful situation. This would seem to be especially true of night-time driving and the higher speeds and oftentimes poorer roads of rural areas. It is hard to see any reason for suspecting females drive more in snowy weather in these conditions than males. It is more reasonable to suggest safe driving ability dif- ference. The high demands for skill which a snow-covered road often makes on a driver, particularly at night and in rural locations, were apparently more frequently met by male drivers than female drivers. CHAPTER VI CONCLUSION Many statistically significant and replicated dif- ferences between male and female automobile drivers both for single and multiple vehicle accidents have been found and reported in this study. The overall model or hypoth- esis adOpted at the outset was that all of these differ- ences were potentially explainable in terms of driving ex- posure differences between the sexes. Many of the differ- ences were indeed explainable in this manner. There were some exceptions, however. Females were found to be positively related to road defects and snowy weather. Either of these results standing alone would not be sufficient to cast much doubt on the ex- posure hypothesis. However, in combination with results of previous studies they indicate that there may very well be something else going on. In Pennsylvania Turnpike accidents, Blotzer, et al., found females were relatively more often in the categories they named "deficiencies in routine driving skills” and #h "failure to cape with road conditions“. Penn found fe- males to be guilty of “faulty driving" more frequently than unBlotzer, et al., op. cit., p. 37. 118 male drivers.u5 In the presence of an unexpected motor scooter, Uhr found female drivers to be markedly more likely to make a dangerous and inappropriate response than “6 Baker found females less able to successfully ‘57 males. cope with flat tires while driving. There seems to be a definite common thread running through these results. / All indicate that female drivers tend to have more accidents in the presence of stressful conditions, in other words, all indicate that female driv- ers are more likely to make an inappropriate response when driving in a situation requiring added skill./ Blotzer, et. al., explained his results in terms of the way drivers learn to cope with stressful situations. They suggested that since this knowledge is not really taught and not required to obtain an operator's license, drivers must learn from experience. flSince females tend to drive less frequently under stressful conditions they learn more slowly:’8 There may be other reasons for this sex difference. Two areas, the psychological and the physiological, seem most promising. Physically, men are of course, generally stronger, but this probably makes little difference with usPenn, op, cit., p. 3. uéUhr, op. cit., p. 69. u7hker, OE. Gite, ppe 8-90 “BBlotzer, et al., 0p. cit. 119 the ease of operation of modern day vehicles. Because of athletics and physical labor, men are probably better co- ordinated than females which may make a difference in quick manipulation of a vehicle. From a psychological point of view, men are prob- ably more accustomed to needing to make a quick rational decision and then acting upon it than females. Females, furthermore, are expected to react to situations in a more emotional manner than are males in this culture. Under stress it is the male rather than the female who is usually looked to for calm rational decision making. This is not to say that females can not be trained to make appropriate responses. It does, in fact, appear that females do need more training than males. The sugges- tion can be made to devise some sort of training device for training all potential drivers to cope with snow and ice, heavy traffic, road defects, etc. Special emphasis should be placed then upon the training of female potential driv- ers. A device of this sort could have an appreciable effect in the prevention of accidents. [There was also some indication that male drivers tended to exceed the speed limit and generally be traveling faster more frequently than female drivers when involved in an accident] The interpretation given to this result was in terms of cultural roles and driving confidence rather than in terms of differences between the sexes in driving abilities when exceeding the speed limit. 120 The practical implications of an interpretation stressing ability would be in terms of increased training for males in driving at speeds exceeding the legal limit. This does not seem at all relevant to the problem. Rather, what appears to be needed is increased edu- cation fooused on the hazards of excessive speed, particu- larly aimed at male drivers. Also, attempts should be made to alter the association between excessive speed and mascu- linity which has developed in some segments of this culture. The fact does remain that males were found signifi- _ cantly more often having consumed alcohol before accidents and consistently were driving at higher rates of speed di- rectly preceding accidents. Both of these factors are con- ducive to serious accidents. The data of this study con- tained only two categories of severity, fatal and nonfatal, and only a few of the entire sample of accidents contained a fatality. Despite this lack, a reasonable assumption to make is that males tended to have more severe accidents than females. In this way females were safer drivers. Without trying to minimize the need for improved driving on the part of women, if a choice had to be made as to which of the two sexes was in greater need of being the target of a safety campaign, male drivers would undoubtedly get the nod. BIBLIOGRAPHY BIBLIOGRAPHY Allen, Terrence H. "A Factor Analysis of Accident Records.” Highway Research Record, No. 79. January, 1965, PD. 17 "‘ 250 Allgaier, Earl. ”Some Road-User Characteristics in the Traffic Problem.” Traffic Quarterly, Vol. h, 19509 PPO 59 ' 770 Baker, J. Standard. "Research and Traffic Records." Traf- fic Digest and Review, July, 1971, pp. 8 - 10. Baker, J. Standard. “What are the Causes of Traffic Acci- dents?" Traffic Di est and Review, October, 1961, pp. ll - 17. Beach, Dayle. ”Solving Knotty Accident Classification Problems." Traffic Safet , Vol. 71, August, 1971, pp. 15 - 38. Blotzer, Paul., Krumm, Richard L., Krus, Donald H., and Stark, Donald E. Accident Causation - Pennsylvania Turn ike Joint Safet Research Grou . rrisburg: Hestinghouse Air Brake Co., 1955. Brody, L. Personal Factors in Safe Operation of Hotor Vehi- cles. New York: New York University, 9 7. _g§pppell, B. J. "Driver Age and Sex Related to Accident ‘ Time and Type.” Traffic Safety guarterly Research 83'10" v01. 10, 19 . ppm 3 - 30 campbell, Hilliam G. Form and St la in Thesis Hritin . Boston: Houghton hifflin Co., 1537. Capelle, Donald 6., et al. ‘An Introduction to Highway Trans ortation En ineerin ., ed. Dona . pelle, Donald E. Cleveland, and Woodrow w. Rankin. Hashington: Institute of Traffic Engineers, 1968. Goldstein, L. G. Human variables in Traffic Accidents. Hashington: Nations Academy of ciences - ational Research Council, 1962. Haddon, William., Suchman, Edward A., and Klein, David. Accident Research Methods and A roaches. New York: Harper and Row, 1935. 123 “Highway Research News Briefs.“ Highway Research News Vol. 31, Spring, 1968, pp. - 9. ' Johnson, Charles F. “The Human Factor in Automobile Acci- dents.“ Traffic Safet , Vol. 72, March, 1972, pp. 13 - 33. Kreml, Franklin M. “The President's Task Force on Highway Safety.“ Traffic Di est and Review, Vol. 19, March, 197 , pp. - . Lauer, A. R. “Age and Sex in Relation to Accidents.“ Accident Research Methods and Approaches, eds. William Haddon, Edward A. Suchman, and David Klein, pp. 130 - 138. New York: Harper and Row, 196“. Lauer, A. R. “Characteristics of the Driving Population with Respect to Age, Sex, Driving Habits, and Accident Involvement.“ Iowa Academ of Science, l95h Proceedings, 361, pp. 39 - 95. Iauer, A. R., et al. Structure and.Characteristics of the Drivin Po ulation A Stud of A e and Sex in Rel;- tion to Accident Susce tibiIit . Ames, Iowa: Iowa State College, 1935. McFarland, Ross A., and Moore, Roland C. Youth and the Automobile. Washington: Golden Anniversary fifiite House Conference on Children and Youth, 1960. McFarland, Ross A., Moore, Roland C., and warren, A. Bertrand. Human variables in Motor Vehicle Acci- dents. Boston: Harvard School of Public Health, 1933. Michigan Department of State Police. Michigan Traffic Accident Facts: 1266. Lansing: Michigan Depart- ment of State Po ice, 1967. Michigan Department of State Police. Michigan Traffic Accident Facts: 1221. Lansing: Michigan part- ment 0 State 0 ice, 1972. Penn, Hugh 8. “Causes and Characteristics of Single Car Accidents.“ Hi hwa Research Record, No. 79, January, 1965, pp. - . Presidents Committee for Traffic Safety. Highway Safety Action Pro ram. Washington: U. S. Government Printing Office, 1960. Recht, J. L. “The 1971 Traffic Story.“ Traffic Safety, V01. 72’ fireh, 1972. pp. 8 " 330 12h Schwenk, Lilliam C. “Age and Sex of Motor Vehicle Fatality Victims.“ owa cade of Science, 1956 Pro- ceedings, 6%, pp. 313 - 319. Schwenk, Lilliam C. “Age and Sex in Relation to Fatal Traffic Accidents for 1957 - A Continuation Study.“ Iowa Acade of Science, 1958 Proceedings, 65, pp. 522 - fizz. . Siebrecht, E. “A Preliminary Report of Accident Character- istics of Iowa Drivers.“ Iowa Acade of Science, 1953 Proceedings, 60, pp. 332 - 337. Solomon, David. Accidents on Main Rural Hi hwa s Related to S ed Driver and Vehicle. 0. S. Bipartment of C 5 f l ommerce, ureau 0 Pub ic Roads, Washington: U. S. Government Printing Office, 196h. Swanson, Clifford 0., Schwenk, Lilliam C., and Lauer, A. R. “Age and Fatal Motor Vehicle Accidents.“ Hi hwa Research Abstracts, Synopsis Issue, Vol. 27, cem r, , p. 69. Tarrants, William E. “Myths and Misconceptions in Traffic Safety.“ Higpway Research News, Vol. 31, Spring, 19689 PP- ‘ o The 196# California Driver Record Study, Part II. Acci- dents, Traffic Citations, and Negligent gpprator cunt Sex. Sacramento: California partment of Motor Vehicles, March, 1965. The 196# California Driver Record Study, Part v. Driver Record c Sex and Marital Status. Sacramento: California EEpartment of Motor Vehicles, June, 1965. Uhr, Leonard. “Sex as a Determinant of Driving Skills: Vomen Drivers.“ Journal of Applied Psychology, Vol. “3, 1959. Do a vanosdall, Frederick E. “An Introductory Study to Show the Relationships Between Michigan Drivers by Age, Sex, and EXposure in Miles of Motor Vehicle Operation.“ Unpublished Master's thesis, Michigan State University, 1966. Viteles, M. 8., and Gardner, H. M. “Women Taxicab Drivers: Sex Differences in Proneness to Motor Vehicle Acci- dents.“ Personnel Journal, Vol. 7, 1929, pp. 349 - 355. Winer, B. J. Statistical Principgls in Expgrimental Desigp. New York: cGraw - Hi , l9 2. APPENDIX APPENDIX A CATEGORIES WITHIN EACH VARIABLE Time Variables Light Condition Light Dark Dusk-dawn Time of Day 12 PM 3.AM 7 AM 9.AM 11 AM 1 PM E PM 6 PM 9 PM Day of Week 3 AM 7 AM 9.AM 11 AM 1 PM a PM 6 PM 9 PM 12 PM Sunday Monday Tuesday Wednesday Thursday Friday Saturday Road Characteristics Highway Classification 0. S. State County City Road Geometry Straight Grade 127 Curve Grade-curve Road Surface Paved Not Paved Road Defects Road Defect No Road Defect Location Intersection Intersection Nonintersection Locality Urban Rural Collision Characteristics Directional Analysis Entering From Angle Same Direction Opposite Direction Stopped Parked Pedestrian Single Vehicle Fatal Fatal Non-fatal Vehicle Characteristics vehicle Type Car Pickup Truck Other 128 vehicle Defects One or More Vehicle Defects No vehicle Defects Vision Obscured Vision Obscured Vision Not Obscured Driver ghgracteristigs Sex Female Male A80 19 and under 20 to 2# 25 to bk #5 to 6b 65 and Over lxperience Lsss Than One rear One To Five Tears More Than Five Iears Registration Instate Outstate Driver Behavior Speed Violation Speed Violation No Speed Violation Right-of-way‘Violation Right-of-way Violation No Right-of-way Violation Following-too-close Violation No Violation 129 Passing Violation Passing Violation No Passing Violation Turning Violation Turning Violation No Turning Violation Traffic Control Violation Traffic Control Violation No Traffic Control Violation Violations Other Than Speeding or Drinking Any Violation Other Than Speeding or Drinking No Violations Other Than Speeding or Drinking Alcohol Not Drinking Drinking Drunk Sleep Sleep No Sleep Speed 0-10 11-20 21 31- 0 1-50 51 61 - 7O 71 and Over Weathsg Vagiables Weather Clear Rain Snow Poe Surface Condition Dr! Wet Snow 130 "I7'11@ifllflfi'flfifl'ifl'flfilflWI“