While I don't actually disagree with your conclusion (and am more than confident that you know more than me about social media), this data only supports the hypothesis that "people share content using social media buttons", right? It doesn't necessarily refute the idea that "people will share content just as much if it doesn't have social media buttons" -- that is, there's no proof that the subset of Twitter links you found wouldn't be shared manually if there was no tweet button.
I can confirm this. All web projects I worked on which were minimalistic, clean and well designed had less stickiness (page impression count to visits) compared to sites stuffed with content, pics, links, extra widgets here and there.
No citation is needed for an unactionable assertion. Do you really believe in the entire history of every web page ever, this is untrue? The statement was more or less tantamount to use metrics to make decisions.
I consider myself a "power user", relatively speaking, of the web...in that I know how to copy-paste URLs and titles into Twitter using all manner of computing devices. And I can even look up the author's Twitter account to ping them in the tweet. You laugh at that, but just yesterday I had a media professor tweet asking for my email address, even though it's listed several times on my blog and I can't direct-message her with that info until she follows me.
But I'm starting to find it inconvenient when social buttons aren't used. Maybe it's because I'm getting older...but partially because web developers are either moving too fast for me, or have no concept of graceful degradation. For example it is now impossible, from what I can tell, to provide a "right-click" tap on the New York Times website when you're on an iPad. That is, you can't hold to highlight text (which is helpful when tweeting the title of Tweet).
So call me a surrender monkey. Social buttons are the web-widget we deserve in the fractured development landscape.
Edit: I also have to dispute the OP's contention that the buttons are never used...blog posts of mine that get shared a lot almost always involve the Tweet button...It's easy to tell because I've configured the button to post the title and my name in a way that you wouldn't if you were manually creating the Tweet itself.
But I don't want to tweet the title usually I want to find a nice quote and edit to fit. If I can't and the site tries to editorialise me pasting to twitter then I will not tweet.
Note that the first link is written by the guy who started Upworthy. Those social media buttons are probably pretty awesome if all you're doing is repackaging other people's content with linkbaity/one-weird-trick titles.
I work for a very large e-commerce site and we did a similar test, we found that when we put Facebook like/share buttons on our product pages, social activity increased but it had a severe impact on conversion. My theory at the time was people were simply distracted - the second they saw Facebook, they got lost in all their friends status updates and forgot to come back and complete their purchase.
Edit: I should clarify - the extra traffic we received as a result of the extra social activity was tiny and in no way made up for the damage to conversion.
http://luigimontanez.com/2012/actually-social-media-buttons-...
The Nieman Journalism Lab followed up a few months ago:
http://www.niemanlab.org/2013/11/tweet-buttons-are-less-of-a...