I see your point but don't fully agree. There's only 3 data points for each cohort, and the change seems to be within the margin of error, at that. Have you considered that the year-on-year decline might also be correlated with the modest increase in economic growth over the last few years? This occurs to me because I notice that the older two cohorts (which would have the highest number of retirees) actually show a slight increase in viewership for this year over 2012.

I do think TV is in long term decline, though as I explained in a nearby comment I think that's more to do with the model of TV distribution than production. The subheading on the graph actually acknowledges this (saying that streaming video services are disrupting the traditional pay TV audience model), but the headline is misguided - I'm not convinced people are spending that much less time sitting in front of the TV than before.

To be honest, I'm a bit surprised the decline isn't steeper, although I'm basing that partly on changes in my own viewing habits (n=1) as the selection on Netflix improves - Iwatch about 6 hours of news and 4 hours of entertainment on TV in a typical week, but since rescuing a dog has made it hard to go to the movies, over the last few months we've been systematically working our way through every episode of Star Trek instead.

I guess I didn't make myself clear. People may be watching less broadcast and cable TV at time-of-broadcast, but that doesn't necessarily mean less time spent watching entertainment produced specifically for television. Broadcasters (which I'll use here to include cable TV channels like HBO etc.) have services like Hulu through which they capture revenue, and they are also willing to finance the production costs of TV series if they are going to get some of the long-tail revenue from Netflix etc. The lower initial viewership aka TV ratings doesn't necessarily indicate a lower lifetime audience. That's why you don't see the stocks in freefall even though superficially the audience appears to be shrinking.

That's what I mean about bogus stats. I never power my phone off (why would I?) and its got 3G internet everywhere I go, therefore on the graph by the same logic I have 24 hours of mobile use per day.

Maybe I'm a contrarian or I see lot of this here on HN for what it is but I find these comments funny and ironic. How do you find time to comment HN with all ths living to do?

20 hours over the weekend plus 4 hours a day at night. It's possible.

Not everyone has a full time job. The labor force participation rate is only about 63% in the US.

Also, I think a lot of people need television as background noise. They don't necessarily watch it, but it's on while they do chores and what not.