A randomized controlled experiment: Find a collection of 200 people who you will be able to test 5 years from now. Send copies of this report to a random selection of 100 of them. Send a similar article not about television to the other 100. Test after 5 years.