Unfortunately, most people read the headline and then decide whether or not they agree with the article. My girlfriend works at a major online media publication and a lot of their Facebook shares and clicks happen before the user has viewed the story.
That makes sense, your likes/shares aren't for you they are for other people to see. So it's more important to like and share the things you want other people to see you like/share than to have them reflect your actual preferences.
Which is logical if you consider likes and reshares as the goal of sharing the story: sheer volume makes it impossible for me to evaluate each post in my Facebook feed in depth, so even if we assumed that people would actually read the article before liking, they are much more likely to read one with a good headline. A good hook is more important than good content (from the viewpoint of getting lots of reactions).
And when I can see how many people liked what I shared, that seems like a reasonable metric to optimize what to share
Which is an issue that most likely any with experience in systems like this knows - and instead of letting the user know that unless they read the article and are able to proceed to randomly prove they read it, there vote will not be counted.
For the UK, I have found that the articles featured on the front pages of major newspapers (e.g. The Telegraph or The Guardian) are very different from the ones featured in my Facebook stream. These are usually highly polar political pieces for Facebook.
It makes sense, on the other hand - an average Facebook reader is more prone to clickbait-y content than a user who specifically targets the front page.
