"black hole 09" wrote:
i say thay should try an experment 16 people 1 roller coaster 1 desibel meter 2 trips around the track 1st time without sound and the 2nd time with sound and see what the readings are on the meter lets see i bet they stay the same only one way to find out...
you couldn't do it that way, too many confounders, not least that a rider may be more or les likely to scream the second time round anyway.
The only way to do it fairly would be to have a random 50/50 chance of having the speech on the headphones, measure the scream levels all day, then do a load of statistical hypothesis testing to see if the average 'speaker' scream is louder than the average 'non-speaker' scream.
My guess is that people won't scream louder because of the voice individually, but it would make every scream at the same time (just after he says 'don't look down') instead of waiting for the drop - when a lot of people are speechless anyway.