How can you be confident that the full range of warning times for approaching vehicles in a given situation is longer than the crossing time?
As explained on the previous page, before you can conclude that a crossing situation is a Situation of Confidence,
you have to determine whether the full range of warning times includes any that are shorter than the student's crossing time.
This is the only way you can be confident it is clear to cross whenever it is quiet / (you see no vehicles approaching).
So, how many samples of approaching vehicles are enough to be confident that the warning times of all the approaching vehicles in that situation will be longer than your crossing time?
Many of us have struggled to figure that out (click here if you want to read about it).
Researchers often are faced with this same question -- how much data must be collected before concluding that there is no difference between the control and the experimental groups?
They have to balance the ideal data collection with what they can manage with the resources and time that they have, and decide how many samples they need in order to be confident of their results.
The same is true for analyzing crossing situations -- ultimately you have to decide how many samples you need for confirmation.
The next page has some suggestions that may be helpful.
Remember, the vehicles with the shortest warning times are NOT necessarily the fastest vehicles, or even the quietest ones!
- Research indicated the detection-to-arrival time was not affected by how quiet the car was, and it had little relation to the speed of the cars (Wall Emerson and Sauerburger, 2008).
- At the crossing where Dick and Lorraine were killed, one of the two cars I couldn't hear until they were 3 seconds away was going very slowly and the other was going very fast.