Not surprising, but I wonder about the validity of the data. They don’t monitor everyone watching TV, only a relatively small sample, and I feel this sample is inherently biased towards people who do watch TV. So the actual drop could be even greater.