So how are they lying with numbers? By not making clear that people who decided to watch are strongly biased in favor of the Prez and enjoy his speeches. This year 91% liked the speech, last year only 83% liked that speech. However, as my previous post notes, fewer people chose to watch this year, making the sample of speech-watchers even more pro-Obama.
This article makes as much sense as going to a rock concert and asking people on their way out if they like the music. Chances are the only people who chose to go were people who already liked that music.
To get better data CBS should have asked random sample of likely voters whether they watched the SOTU, ask viewers if they liked the speech, and ask non-viewers why they chose not to watch, when SOTU was very widely available. CBS didn't do this because they knew they'd get results they wouldn't like.
