When you send out a Tweet, please don't forget that the vast, vast, majority of your followers aren't going to click your link (and those that do click, the vast, vast, majority of them probably won't actually read the article they click to). Even if you consider every single Twitter user "engagement" as a link click, I'd bet that for most of your tweets, 5% or less will actually bother to click through.
What that means is that if you're a trusted authority, you owe it to your followers to ensure that your message is the message of the tweet itself. Meaning sharing a hyperbolic headline, or the title of a weak study, without your own qualification in those 140 characters, may lead your followers to believe the hyperbole or that the crappy study you linked to because it was "interesting", was in fact important.
Bottom line, if you think a study's hugely preliminary, say so. If it's an animal model, say so. If you haven't read the actual study, say so. And if it's just bollocks, please don't link to it at all.
Curious about your Twitter Analytics, click here and have a peek.