The only person it is capable of 'testing' is yeurself, and even then, all it tests is yeur PREFERENCE of certain values, not whether yeu actually think in that manner or not. As the preference of values themselves are inherently flawed, being a quantified version of an abstract concept, the point is not to assume it's 100% accurate and that yeu should treat it as pure perfection... that is the same mistake as those who abuse it make.
The point is to realize that it is a system designed to try to describe something in generalities which require specifics; MBTI can not and never will 100% accurately describe anyone, nor will it ever be able to accurately predict, even by its' own standards, whot someone's type "really" is; it can only measure yeur preferance, and yeu may not happen to like yeurself very much.
If yeu want further proof, keep in mind that it does not even TOUCH upon the idea of yeur base desires. That's Enneagram; I'm a 3 entp which's quite different from a 7 entp, though those are more common. There's not much yeu can do except try to further refine yeur studies through additional tests, and realizing that they're not going to be perfect as there's no possible way to break down the entirety of the human race into 16 boxes. Or we'd just have 16 clones that all look and act 100% identical.
As such, I don't see how stating "there's never been a double blind test of something which can't be accurately tested by two people simultaneously" has any relevance at all. There is *NO* test, and never will be any test capable of doing so for the psychological profile of anyone.