No less than 4 different people alerted me on my Facebook feed to the study on sentiment analysis and Facebook feeds Experimental evidence of massive-scale emotional contagion through social networks. There is quite a number of media reports on it researchers like Tal Yarkoni and Brian Keegan have commented on it. I have put some pointers on the Brede Wiki.
Lars Kai Hansen states that he “doubt[s] emotional filtering is routine at [Faacebook].”
I think that Facebook is actually doing emotional filtering. Here is a comment I made on a thread in Facebook: Doesn’t the brouhaha show how naïve Facebook users are? These A/B-testings occur all the time. Twitter and Facebook are continuously optimizing the news feed ordering (at least I should think so). Having sentiment analysis as a component is just one feature. If they didn’t do that the data scientists would not be worth their job. If you as a user care about privacy and not being manipulated you should stop using these social media. I suppose that for the scientist involved and possibly PNAS it is a bit of an issue if the IRB is not there. Susan Fiske seems not to be quite clear on that point.