If one of the tests was inaccurate, couldn't it also be true that both tests were inaccurate? And is it also possible that the test/concept itself is flawed given such variance in test results of the same people over time?