One aspect he ignores is the inflammatory nature of the particular variable being manipulated in the study : emotional state. Researchers should not underestimate the different reaction people will have to, on the one hand: experiments on usability, retention, click-through rate, engagement and on the other hand: experiments manipulating their emotional state.
The first type of experiments and analysis will annoy people. But it brings to mind businesses trying to maximize their profit. It's fits in with the idea of marketing research and is expected. The second type of experiment brings up ideas of the Matrix and Orwellian thought control. Yes, it's a huge leap. In many cases it will be an unfair one as well. But it is a fact this will be the reaction by the public who have valid concerns about how their data is being used and the media looking for inflammatory stories. The solution is not simple and as Keegan points out will require a more mature discussion than we are having right now. But researchers need to be prepared for these reactions and not act surprised when it happens.