Controlled experiment? Causation/correlation? (Score 0)

The experiment is not not really controlled, and therefore cannot prove the causation that is implied by the headline. They call random people and ask if they use IM and then ask other questions. For it to be controlled (and to prove causation, i.e. IM helps productivity), they would need to first randomly select the experimental group and control group, and THEN have one use IM and other not. Since they don't do this, they only show correlation, A 3rd factor could have caused those users to use IM and to be productive. Interestingly, in the TFA, when they list the hypotheses proven, they are correctly stated as correlations (although not using that word), as in: "IM users report lower levels of disruptive interruptions than non IM users". That is a statement of correlation, not causation, as opposed to the headline of the article (and Slashdot's). I also agree with the commenters who point out 1) the definition of productivity is questionable and 2) It is based on their own assesment of whether they waste time with interruptions, not by being observed by the experimenters. All in all, not much is proven IMO.