Three years of reported accident data was used. You can't generate more past data out of thin air. And, you can't analyze the frequency of things that never happened. So the case can be made to hold-off on striping for the time being, and rigorously collect new data going forward for a given reasonable period of time. If the frequency of accidents increases over the test period, then proceed with striping. If the frequency stays the same, continue collecting data. If the frequency diminishes, then cancel striping.
|