TV-related social media activity does
have a relationship to TV ratings. But does it cause TV ratings to climb? Nielsen says it has no proof -- yet.
In looking at an recent study Nielsen compiled with SocialGuide, its social TV
measurement unit, it says there is a "statistically significant relationship" between Twitter volume and TV ratings.
For example, in evaluating the fall 2012 premiere TV program ratings of
over 140 broadcast and cable programs, an 8.5% increase in Twitter activity was associated with a 1% rise in TV program ratings among 18- to-34-year-old viewers.
A higher volume of Twitter
activity was needed for older viewers to be associated with the same TV rating movement. Nielsen says a 14% increase in Twitter volume was associated with a 1% rise in in TV program ratings for 35-
to-49-year-olds.
Looking at the regular fall 2012 season episodes, the Twitter association is stronger in connection to TV program ratings. A 4.2% rise in Twitter volume is associated with
a 1% rise in ratings among 18- to-34-year-olds. There was an 8.4% rise in this social media Twitter activity is connected with a 1% climb in ratings among 35- to-49-year-olds.
Mike Hess,
executive vice president of media analytics for Nielsen, stated: “While our study doesn’t prove causality, the correlation we uncovered is significant, and we will continue our research to
deepen the industry’s understanding of this relationship.”
Nielsen says its analysis included a list of other variables that could impact ratings, such as ad spend, prior-year
rating, promotional spend, and differences in age and gender.
While most of these factors have a logical relationship with TV ratings, its analysis looked at "statistically significant
alignment" with TV ratings. Three factors showed this: prior-year ratings, ad spend and Twitter volume.
advertisement
advertisement
Statistical significance only speaks to minimal accuracy, not the magnitude of the relationship. The sample size affects significance, so it could be a tiny or large effect, depending on the absolute size of the correlation. I am hopeful that Nielsen will freely provide more than a bare bones news release.
Why, Mr. Friedman, do you persist in seeking a causal relationship between Twitter and TV? It is one thing to pursue your apparent pet hypothesis in a blog. It is another thing to flog it (sounds like 'blog it') in a news article. As far as I can tell, all Nielsen is really indicating is that there is a statistical relationship between two data sets it has compiled. Nielsen researchers also know there is a "statistically significant relationship" between the weight and height of TV Viewers. But that is not news -- and neither is this article.
No scientist would say weight causes height or vice versa. So let's get one thing straight from here on: "An empirically observable correlation between two interdependent variables is a necessary, but not sufficient, condition for causality." OK? Perhaps if you had tried to Tweet this story, you would have discovered that there was no story to Tweet. Onwards and upwards!
Here I am watching a high intense drama with a distracting on-screen logo and a hashtag crying for attention. I stop to tweet then ask the person next to me, 'what just happened'...