Having vast amounts of information freely available is a wonderful thing, but please, for the love of anything you consider holy or important, don't use abstracts you found in PubMed searches to prove your point (in comments, or on social media) unless you've actually read the full text of the article you're citing, and simultaneously you feel that you understood the article's methodology and limitations enough to think it worth sharing.
Thank you.