REGRESSION MODELS AND CAUSAL INFERENCE IN STUDIES OF THE RELATIONSHIPS BETWEEN ACCIDENTS AND ROAD AND TRAFFIC CHARACTERISTICS
REGRESSIONSMODELLER OCH KAUSALSAMBAND VID STUDIER AV TRAFIKOLYCKORS VAEG-OCH TRAFIKBEROENDE
The best correlation between the number of accidents and the road and traffic variables studied was obtained for the vehicle mileage of the stretches of road. The regression analysis also showed that it is the vehicle mileage which explains the greatest part of the variance in the number of accidents. In the regression analysis it is chiefly the average annual daily traffic of the stretches of road together with the sight variable "the number of minimum sight distance spots per kilometer for which the sight distance is less than 150 meters" which explains the greatest part of the variance in the accident density. The variable "the number of minimum sight distance spots per kilometer for which the sight distance is less than 150 meters" explains the greatest part of the variance in the accident rate. In the attempt which was made to throw some light on the casuality between different accident measures and some of the most important road variables it was shown that the sight variable "the number of minimum sight distance spots" has a large influence on the different accident measures. When this variable is included in the analysis together with the road width, the influence of the road width on the accident rate and the accident density is eliminated. As regard the accident measure "the number of accidents" it was found that in addition to the trivial influence of the length and average annual daily traffic of the stretches of road there was an influence of both the above mentioned sight variable and the road shoulder width. In conclusion it can be said that some of the sight variables studied show a stronger relationship to the different accident measures than what is the case for the road width.
- Corporate Authors:
- Nilsson, G
- Publication Date: 1974
- Features: Figures; References; Tables;
- Pagination: 52 p.
- TRT Terms: Crash causes; Crash rates; Regression analysis; Sight distance; Traffic volume
- Subject Areas: Data and Information Technology; Highways; Safety and Human Factors;
- Accession Number: 00099559
- Record Type: Publication
- Source Agency: Highway Safety Research Institute
- Report/Paper Numbers: No. 43
- Files: TRIS
- Created Date: Nov 5 1975 12:00AM