On Wednesday, October 7, 2015 at 9:24:00 PM UTC-4, rbowman wrote:
On 10/07/2015 01:36 PM, DerbyDad03 wrote:
1 - Yes, the warm up may have altered the results to some extent, but
not "utterly". Even if some of the drivers stopped, I'm confident in
saying that it wouldn't be enough to wipe out the entire 20% of
improved efficiency. There is no also guarantee that every of those
drivers have encountered - or properly deal with - 4 way stops.

If you design an experiment where you train the Group B participants
before hand and ignore Group A, you're not proving much other than
people perform better in an environment that they're familiar with.


What group was trained/ignored in the Mythbuster's experiment? They gave
time to all drivers to familiarize themselves with both the 4-way stop
intersection and the roundabout. Who was Group A and who was Group B?

In addition, I see that you snipped my #2, which I stated was even more
important than my #1: They were testing the *efficiency* of a roundabout
over a 4-way stop. The only way to do a valid comparison is to have all
participants trained in order to eliminate the "What the heck do I do now?"
variable.

If you want to complain about the testing method, complain about the
randomness of the left and right turn variable. In both tests they
used operators to randomly tell the drivers which way to turn (or not).
For a valid one-to-one comparison, they should have used the exact same
drivers, in the exact same positions, executing the exact same turns.
That method would have eliminated many variables, most importantly how
long each vehicle was in the intersection, thereby forcing other drivers
to wait.