LKeZ5arBDc5 - Sebastian Thrun from +Udacity
sez: "We learned we can drastically boost learning outcomes by adding a service layer around MOOCs. It has a huge impact on completion rates and learning outcomes."
It's weird that a statistician confuses correlation (people who enjoy our new 'service layer' have a higher completion rate) with causation (they have a higher completion rate because this 'service layer' helps them learning and mastering the certificate) and is completely oblivious to the obvious confounding variable - money.