Case 1: a bunch of self-proclaimed "social media experts" decide to make some semi-random changes to Facebook. These results are never made public, but people still complain about the fact that Facebook's layout keeps changing.
Case 2: Facebook consults with academic experts in their fields to perform a focused and limited experiment and then shares the results with the rest of the world.
Case 1 happens ALL THE TIME and we don't even know it. Case 2 suddenly pops up and people become unhinged.
Personally, I find Case 2 to be far more ethical, even though Case 1 is well within their legal rights. I think this is a great opportunity to encourage large companies to continue to engage in fairly public research (which is something Facebook has been famous for when it comes to Comp Sci).