I've skimmed through the report linked, and I'm pretty skeptical of their methodology. For example, on their "Television Methodology" section:

> Television data are derived from Nielsen’s National TV Panel that is based on a sample of over 40,000 homes that are selected based on area probability sampling.

The report does not go into any details on how these 40,000 homes were selected, or any demographic breakdowns beyond saying that the data is "inclusive of multicultural audiences." It's difficult to say how representative these 40,000 homes are of the "average" American.

It's also not entirely clear how they derive their measurements. There's little that discusses how they handle potential over or under-reporting, and there's nothing that says how they differentiate between different people in the same household.

While this report certainly isn't useless, there's enough grey area here that I would be hesitant to draw any conclusions from the data presented. We'd need a much more comprehensive breakdown of their methodology before we can be confident that their data is accurate and representative.