In Reply to: Re: If you did nothing and none raised their hands you would have some validation. posted by PhilNYC on August 20, 2006 at 08:57:39:
As noted by others, while you think you had a fairly straightforward test going on, there other issues. It is also important that the person administering the test to the subjects be unaware of what (if any) change took place in the tweak. You probably shouldn't know the results until the stats have been compiled afterwards.Of course, there are different levels of validity based on how rigorous the test.
Can a test like the one you describe be the basis on which to make a buy, keep or discard a piece of audio equipment? Certainly! It is your money and your system.
Can one reach a solid, defensible conclusion about the widespread applicablility of the issue in question? No. Too many variables, lack of adequate controls, too small a sample and so on. The best your test could do in that area is suggest that you might have an interesting issue to pursue further. Frankly, it is not easy to design and conduct good audio tests. That's why a lot of people don't do it, especially if they are not sure in advance that they'll like the results.
One of the common problems in this forum is people do tests such as yours and then want to make universal pronouncements. (Making clear that is NOT what you just did.)
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Re: Test administrator needs to be blind as well... - mls-stl 08/20/0609:31:49 08/20/06 (5)
- Re: Test administrator needs to be blind as well... - PhilNYC 09:49:36 08/20/06 (3)
- Actually few experiments are done double blind - Norm 10:40:17 08/20/06 (2)
- Re: same thing stated another way - mls-stl 11:29:55 08/20/06 (1)
- I was merely pointing out that double blind testing is relatively uncommon - Norm 11:45:58 08/21/06 (0)
- Good point about not generalizing to what others might hear. nt - Norm 09:40:25 08/20/06 (0)