Plagiarizing yourself is sticky. Technically when you publish an article, you sign away the copyright on the text, but you own the know-how/IP. So if you've not published in a peer review journal where the copyright is no longer yours, then you're free to recycle text, so long as you're not purporting something to be new that isn't. However, if you've signed away a copyright, then you cannot use the same text, even if you're writing a paper about some data you already published another paper about, any thus the data collection method wasn't similar, but literally the same. Still need to figure out a new way to write up the methods. At least, that's the way I understand the rules.