Honesty, I expect, would be more important than not thinking. While the individual questions are often immediately obvious as to their gender stereotypes, I meant if you know yourself then the outcome/"answer" of the test should be vaguely accurate, as opposed to constructing an answer to suit an outcome you'd prefer. Being confused about your gender may cause the outcome to not be what you are expecting. Hope that clears it up a little.
Using this as a self-evaluation is tricky. As I pointed out I took the test 3 years apart and I don't feel that my personal values had changed greatly nor my hormone levels or types and the outcome was a little different. Do you think it is that accurate that I should judge this as a change in myself or could it be that it's not perfect? It seems that you have gone through a definite change in your overall personality which was perhaps vaguely represented in the results, although imperfectly. Is this possible?