I just read the abstract of a paper that critiques an older paper because it reported a strong effect from its research, but "there are serious methodological concerns, which make the overall findings questionable." The original study should never have been published. Reporters just quote these results without understanding that methodology is everything. You can find any relationship if you bias your methods accordingly, and then get that relationship into the public domain.
Drug companies do this with new drugs. The anti-climate change lobby does this with warming studies. Food companies do this with junk food studies. Health food companies do it with health food or supplements. And again, and again.
The study in question here linked childhood consumption of candy with adult violence.