mattgbell
Help the elderly and unwell!
I've answered this many times, but happy to go again:
Before the results of a DBT can be taken as valid, you have to show that the test methodology is sufficiently sensitive to discern the small differences you are looking for. So some sort of control, where two items known to be subjectively (and measurably) different should first be tested under the chosen test methodology (which, by the way, is for the testers to determine) and only if the test results bear out the known outcome should the test proper go ahead. If the results of the control don't reliably identify the differences (ie to the required degree of statistical significance) then adjust the methodology and repeat until satisfied.
Not meaning to be captious, but how could you know they were subjectively different without testing them blind?
Another issue I have with this is that it allows a rather easy get out for critics of DBTs. Let's say our control involves a magnitude of audible difference between the two components of x. We find that we are able reliably to distinguish x in a DBT. All good so far: we have established that DBTs are able to distinguish differences of magnitude x.
Then say I propose another DBT of two different components. This time we fail to distinguish the components in a DBT. The critic of DBTs can now say: 'this test involves a difference of magnitude x-1, but we haven't established that DBTs can differentiate between differences of x-1, therefore I cannot accept the results of the test'.
You're left with the problem that any control of this type could be dismissed as being too easy.