Correlated versus uncorrelated applies to signals coming from multiple speakers at the same time while measuring. If correlated, then the same measurement signal is coming from the speakers in phase. This is not good for determining EQ settings, you will end up with phase cancellation at some frequencies, phase reinforcements at other frequencies, and will get a totally crazy frequency response curve. Move the measurement mic a few inches, and you get a totally different crazy frequency response curve.
By using uncorrelated pink noise, the measurement signal coming out of each speaker is random and completely unrelated to all the others, so there will be no phase cancellations or reinforcements, and you get a sane looking frequency response curve.