@charles1dad
Thanks for your interest in this important topic.
I do not conduct blind tests. When I am considering an upgrade I listen to the changed component in my system for a long period, say around a month. If I like it enough, I buy it. Obviously I am hearing differences, but I am not vainglorious enough to claim that I am hearing an improvement. I am conservative about upgrades and keep components for a long time. Even if I decide to change something, I usually keep it for use in my second system, as obviously I liked it when I bought it and that opinion has not changed.
Blind tests would be a great improvement; I think many would agree. But the idea is often criticised as being too difficult to arrange. I have given this a lot of thought.
1. Dealers are trying to sell goods and they hold much more equipment than hobbyists. Many of them have large listening rooms. They are in the best position to organise blind tests. It is to be hoped they don't feel blind tests will reduce their business, but if they hold my beliefs, they may do.
2. The dealer assembles two identical complete systems in the same space, save for the speakers (unless the component to be evaluated is a speaker). One system contains the component to be evaluated, the other has a different component, possibly the customer's present piece. A high grade switch box is constructed to quickly change the connection to the single pair of speakers. REASON 1 Both programmes must emanate from the same positional source. REASON 2: This avoids the delay and hassle of swapping components in and out which requires the subject to employ longer aural memory. Even a minute or two is detrimental - it is commonly believed that aural memory is poor. Pace those professional reviewers who compare a component today with one they last heard two years ago and claim their opinion is useful. Does ANYONE believe them?
3. The customer is blindfolded. REASON: he cannot see what is going on nor infer cues from the activity of the dealer. This avoids the oft-cited difficult problem of the operator needing to be 'blind', the 'double-blind' test, i.e. he does not know which component is playing.
4. The dealer plays the same material on system A, then system on B, then on one of the systems at random, X. The customer has to identify X and say if he prefers its sound. The cycle is repeated at least 10 times.
5. The results can be analysed statistically. if the customer is wrong half the time then it is probable he cannot discern a difference. If he succeeds nearly every time it is probable he can tell the difference and his opinion of which he prefers is validated.
