Crash Modification Factors in the Highway Safety Manual: Resources for Evaluation (2023) / Chapter Skim
Currently Skimming:
Appendix G Development of Crash Modification Functions for Adding a TWLTL to a Two-Lane Road
Appendix G Development of Crash Modification Functions for Adding a TWLTL to a Two-Lane Road
Pages 148-159
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 148... ...
Appendix G Development of Crash Modification Functions for Adding a TWLTL to a Two-Lane Road G-1
|
From page 149... ...
G-12 Table 1. Summary Statistics .....................................................................................................................
|
From page 150... ...
The original analysis found statistically significant reductions at the 95-percent confidence level in total, injury, and rear-end crashes using data from four States where installations were evaluated. The positive effects for rear-end crashes comprised the largest crash type reduction.
|
From page 151... ...
The observed crash frequency in the after period is modeled using a negative binomial model with the EB estimate of expected crashes in the after period had no treatment been applied used as an offset. Variables that may affect the CMF value are then also added to the model.
|
From page 152... ...
4.1 GLM Results The GLM approach was determined to be unsuccessful. Although some models could be fit to the data, the high variability in observed crash counts between sites results in predicted CMFs that are mostly greater, in some cases much greater, than 1.0 even though the treatment is found to be effective overall.
|
From page 153... ...
The model indicates that the CMF for total crashes decreases as the segment AADT increases and is always less than 1.0. The parameter for the AADT variable is however of very low statistical significance and the R2 value is also very low.
|
From page 154... ...
totrateb >=0 and totrateb <1 totrateb >=1 and totrateb <2 totrateb >=2 and totrateb <3 totrateb >=3 and totrateb <4 totrateb >=4 and totrateb <6 totrateb >=6 and totrateb <7 totrateb >=7 and totrateb <9 totrateb >=9 Within each bin the average EB total crash frequency per mile was determined and used as the variable in the model. Table 2 shows the parameter estimates, standard error, p-value and the R2 value.
|
From page 155... ...
CMFi = a+b(TOTRATEi) where CMFi = the predicted CMF value TOTRATEi = the average EB expected total crash frequency per mile Table 3.
|
From page 156... ...
The model indicates that the CMF for injury crashes decreases as the segment AADT increases. The parameter for the AADT variable is however of very low statistical significance and the R2 value is also low.
|
From page 157... ...
For the calibration dataset, CMFs between approximately 1.47 and 0.00 are predicted for individual sites. Figure 3.
|
From page 158... ...
Model Results for Rear-End Crashes Parameter Estimate Standard Error p-value a 1.5702 0.5406 0.1009 b -9.0957 6.2461 0.2826 R2 0.5146 The following graph plots the data points being modeled and the fitted line. For the calibration dataset, CMFs between approximately 1.52 and 0.00 are predicted for individual sites.
|
From page 159... ...
Although some models could be fit to the data, the high variability in crash counts between sites results in predicted CMFs that are mostly greater, in some cases much greater, than 1.0 even though the treatment is found to be effective overall. For the meta-regression approach some success was found for linear models using the AADT before treatment as the predictor variable.
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.