To do this right is very difficult. Read the book Blink to get an idea of how humans form judgements. From psychology we know that people will preceive coffee from a red can as richer and more full bodied than the same coffee from a green can. Every single thing that is said to the subject, every thing in the enviroment, the order things are presented in, and even the lead experiementers beliefs, even when double blind designs are used influences the results.
do blind tests have any practical significance
do the results of a blind or double tests have any usefulness ?
a blind ab test does not prove anything. it only states facts, namely that a certain percentage of listeners were able to distinguish the sound of one component from another.
there are many issues as to the validity of these tests.
the question, "can a listener detect the sonic difference(s) between components", is never answered definitely by one test.
suppose a blind test is administered 10 or more times to the same group of listeners. assuming the first test is replicated 9 more times, there is a chance that there may be an inconsistency between the outcomes of the test.
suppose a statistically significant percentage, say less than 60 percent, for a large sample of listeners, identifies differences between 2 components, what does that mean ?
what is the application of such a result for a prospective buyer of a component ?
a blind ab test does not prove anything. it only states facts, namely that a certain percentage of listeners were able to distinguish the sound of one component from another.
there are many issues as to the validity of these tests.
the question, "can a listener detect the sonic difference(s) between components", is never answered definitely by one test.
suppose a blind test is administered 10 or more times to the same group of listeners. assuming the first test is replicated 9 more times, there is a chance that there may be an inconsistency between the outcomes of the test.
suppose a statistically significant percentage, say less than 60 percent, for a large sample of listeners, identifies differences between 2 components, what does that mean ?
what is the application of such a result for a prospective buyer of a component ?
- ...
- 44 posts total
- 44 posts total