a lot of scientific studies are done that way. They take the
  long way around though. I saw one recently. They measured
  creativity and came up with metrics. They made choices based on
  the videos they saw. Then they chose more videos they considered
  creative and it matched the predictions of the model they came
  up with. Well, duh. Same people making the same choices come up
  with the same conclusions. It didn't predict anything. The only
  difference is: They took notes.