Two points. One, as I pointed out elsewhere, the study was not on searching the internet, but on searching websites.

Two - and I'm sorry to pick on you because others are doing it - you're in real danger of hindsight bias here: http://lesswrong.com/lw/im/hindsight_devalues_science/