The really big problem in open source intelligence has been for some time that data to support just about anything can be found. OSINT investigations start with a premise, look for data that supports the premise and rarely look for data that contradicts it.

Sometimes this is just sloppy methodology. Other times it is intentional.

I think OSINT makes it sound like a serious military operation, but I think political opposition research is a much more accurate term for this sort of thing.