Also, some of the participants had a good idea: if one of your personality dimension letters changes when taking the test multiple times, you can fill it out with an X. Can we add an instruction for them to do this on the next survey?
I don’t take this test all too often (in fact, didn’t take the one in the survey IIRC), but if we can do this, here’s my personality type: IXXX. Oh wait.
(Yes, seriously, if I take an online MBTI test several times at evenly spaced time intervals within the same month, the first varies between .6 and .95 towards I, and the others just jump around in a manner I can’t predict (yet, anyway, probably could eventually if I did more timewasting internet-test-taking))
I predict similar (perhaps less pronounced?) variation would be present in around 30% of LWers (not too confident in this number), and that we could reduce the variation dramatically by eliminating confused questions and tabooing ambiguous or vague words / phrases, replacing them with multiple questions containing various common meanings, and an even greater (bitwise) reduction by giving more contextual information from which the respondent can infer or judge values and weight variables on “It depends, but I suppose most of the time I would...” -type answers. (much more confident in these last two predictions than the first)
I don’t take this test all too often (in fact, didn’t take the one in the survey IIRC), but if we can do this, here’s my personality type: IXXX. Oh wait.
(Yes, seriously, if I take an online MBTI test several times at evenly spaced time intervals within the same month, the first varies between .6 and .95 towards I, and the others just jump around in a manner I can’t predict (yet, anyway, probably could eventually if I did more timewasting internet-test-taking))
I predict similar (perhaps less pronounced?) variation would be present in around 30% of LWers (not too confident in this number), and that we could reduce the variation dramatically by eliminating confused questions and tabooing ambiguous or vague words / phrases, replacing them with multiple questions containing various common meanings, and an even greater (bitwise) reduction by giving more contextual information from which the respondent can infer or judge values and weight variables on “It depends, but I suppose most of the time I would...” -type answers. (much more confident in these last two predictions than the first)