고령자와 비고령자의 교통사고 심각도 모형 비교분석 = (The) severity of traffic accidents model's comparative analysis of elderly and non-elderly

다국어 초록 (Multilingual Abstract)

The severity of traffic accidents model's
comparative analysis of elderly and non-elderly


Woo-Dong Jeung

Department of Urban Information & Geodetic and Cadastral Engineering, Graduate School, Kyungil University

Supervised by professor Jae-Seok Kim

In this study, characteristics of traffic accident were reclassified according to the age groups such as elderly and non-elderly. In order to estimate traffic accident severity model using ordered logit model, accident severity which was given the weight of the accident fatalities was put as a dependent variable and personal characteristics, transportation facilities and road environment were set up as independent variables. Thus traffic accident reduction methods were suggested with selecting influential explanatory variables from estimated models. the detail results from this study are as follows.

1. The accident data between elderly population and non-elderly was classified through discriminant analysis. In the elderly group, 43 cases out of 400 cases were classified as a closer category to the non-elderly group. Also, in the non-elderly group, 55 cases out of 400 cases were classified as closer category to the elderly group. So cases were reclassified into 412 cases as elderly and 388 cases as non-elderly group respectively. Therefore, the criteria of traffic accident could be established using not only the age but also traffic facilities and road environments.

2. In the characteristics of the item at analysis, 60s age were the most common by 38.3% and over 70 years old age were the least by 4.4% among the traffic accident offenders. Also, for the traffic accident victims, 60s age were the most by 43.9% and 20s age were the fewest by 5.1%. For the non-elderly group, 40s age were the most by 27.6% in the offenders and 70s and 80s ages were the least by 0.3%. For the victims age, 40s age appeared as the most by 22.9% and 60s were the least by 7.0%.

3. In order to estimate ordered logit model, the accident severity were set up as the dependent variable and 17 variables named X1~X19 were set up as independent variables excluding covariates.

4. According to results of p-value<0.05, ‘accident type(X1)’, ‘property demage(X2)’, ‘the offender wearing protective gear or not(X7)’, ‘the victim wearing protective gear or not(X8)’, ‘offender’s speed just before accidents(X9)’, ‘victim’s speed just before accidents(X10)’, ‘Victim’s DUI or not(X12)’, ‘weather conditions(X13)’, ‘existing traffic signal or not(X15)’, ‘lane width(X16)’ and ‘existing median barrier or not(X18)’ were satisfied into statistical significance as dependent variables in the elderly group and ‘property demage(X2)’, ‘type of victim’s vehicle(X4)’, ‘the offender wearing protective gear or not(X7)’, ‘the victim wearing protective gear or not(X8)’, ‘offender’s speed just before accidents(X9)’, ‘victim’s speed just before accidents(X10)’, ‘lane width(X16)’ and ‘existing median barrier or not(X18)’ were satisfied into statistical significance as dependent variables in the non-elderly group. As the results of analysis, the elderly group has overall more significant variables than non-elderly group. In the significant variables of elderly group, 2 variables were significant in the characteristics of demage type, 5 variables were significant in the characteristics of drivers related, 1 variable was significant in the weather characteristics and 3 variables were significant in the characteristics of road structures. This results reflect that traffic accidents refer to personal characteristics in the elderly group. Furthermore in the significant variables of non-elderly group, 2 variables were significant in the characteristics of demage type, 4 variables were significant in the characteristics of drivers related, 2 variables were significant in the characteristics of road structures. Non-elderly group also had the personal characteristics as the common, however the variable which showed the most significance was in typical characteristics.

5. Based on the results of the analysis, safety improvement methods were proposed for each characteristic.
First, for mitigation of pedestrian crashes, facilities for segregation of pedestrian and vehicle, pedestrian crossing installation, traffic signal installation, an appropriate space for signal waiting, the acquisition of visibility and size for transportation safety facilities and training and promotion of safety are needed. Also, for mitigation of vehicle crashes, the shortening driving aptitude test period and education of driving are required.
Second, as for the drivers related characteristics, providing periodicals for transportation safety and regular education for drivers and testing regarding educational programs. Also training and management system should be established for supporting educational activities.
Third, for weather characteristics, the more installation of portable enforcement cameras should be used and vehicle inspection cycle should be shortened. Also vehicle emission inspection carried out freely by government should be extended.
Fourth, the installation of traffic signal should be reviewed in the area where the traffic signal have not been installed yet. For road structure characteristics, human resources for the enforcement should be increased, and criteria of safety facilities should be refined as consolidated regulations. Moreover, investment should be increased for developing new materials for safety facilities.

In addition, various factor analysis were limited in this study because of lack of geometric information and human factors. In a future, the traffic accident researches should be done first in order to conduct research for the variety of variables.

목차 (Table of Contents)

• 제 1 장 서론 1
• 1-1. 연구의 배경 1
• 1-2. 연구의 목적 3
• 1-3. 연구의 내용 및 방법 4
• 1-4. 선행연구의 고찰 9
• 제 2 장 교통안전정책과 분석이론 14
• 2-1. 우리나라의 교통안전정책 실태 14
• 2-2. 외국의 교통안전정책 실태 19
• 2-3. 판별분석 25
• 2-4. 교차분석 32
• 2-5. 순서형 로짓모형 35
• 제 3 장 자료의 적정성과 특성 40
• 3-1. 자료의 수집과 내용 40
• 3-2. 자료의 적정성 43
• 3-3. 분석자료의 특성 60
• 3-4. 교차분석 72
• 제 4 장 사고심각도 모형 비교 86
• 4-1. 변수의 정리 및 입력방법 86
• 4-2. 사고심각도 모형 89
• 4-3. 분석결과 94
• 4-4. 사고개선을 위한 적용방법 99
• 제 5 장 결론 106
• [ 참 고 문 헌 ]
• [ 부 록 ]