Trying to make a straight correlation between TV audience and Twitter activity is as nonsense as ignoring the potential Twitter provides to get insightful metrics for TV.
Some shows will trigger more Twitter activity than others. Granted. That does not make any attempt to map TV audience with Twitter invalid. It just shows that getting meaningful insights requires some more thoughtful analysis and modeling than the mere counting of tweets.