It was an announcement we knew was in the works, but Nielsen has made public that it will be tripling the size of its ratings sample by that mythic year, 2011, in which the media industry is hanging all its hopes. (I say this because every projection I come across extends a forecast out to 2011.)
The announcement, made earlier this week, has seen Nielsen proclaim that their numbers will be much more precise now, since they will be based on 37,000 homes and 100,000 people, rather than the current 12,000 homes and 35,000 people that Nielsen says it uses today.
For those of you who read the C3 blog regularly, you know I rail fairly often on the Nielsen's for a variety of reasons, one of which is representativeness. (I'm not alone in doing this; C3 Consulting Researcher Jason Mittell has been known to do this from time-to-time as well.) I feel the same way about this as I do about second-by-second measurement:
I still think it's important to realize that these measurement numbers, taken on their own, are somewhat arbitrary and only provide a limited amount of insight, but it is generalized and quantifiable insight that makes shows and commercials comparable, which has definite value. In a perfect world, there would be a way to avoid the generalizations that come along with quantification, but we don't live in a perfect world, and there has to be some way to make comparisons.So, if numerical ratings or measurements are going to be at the center of the business model for television advertising, we should seek to make them as in-depth and detailed as possible.
I wrote about some of the issues of such a small sample size back when the Nielsens started measuring college viewers, writing:
I was not a math major and in fact haven't had a math class since my senior year of high school, but 30 percent of the 450 families who do have college students means something like 135 college students, correct? I know that the choices are somewhat limited on college campuses, but it seems pretty idiosyncratic that the number of American television viewers who are on college campuses will be represented by about 135 students.
What does this tripling of the sample size mean? It means that, for what the Nielsen ratings are worth, they are much more accurate, just as second-by-second ratings are more accurate than minute-by-minute or even larger measures. It's a major move in the right direction, especially for those poor shows that get 0.0 ratings in which one would suspect that the choices made by even one or two Nielsen families could make significant ratings differences in.
These do not solve many of the larger issues about the ratings system, such as the simplicity of having one major metric used to measure success, and the exclusion of any degree of quality in the quantitative measurement, valuing only impressions rather than the "engagement" of those viewers, to use the current favored buzzword.
But I think Nielsen deserves credit for making a strong commitment to try and make its sample more representative and more useful, since those ratings are likely to remain at the center of the television industry for the foreseeable future.
I too met this news with some conflict -- it would be like being told that now 100,000 people in the country could vote in elections, not just 35,000. Better, but still kind of missing the point. The mere fact that they had to increase *threefold* seems a tacit acknowledgment of how crappy their science has been to date, too, no?
In some sense, though, the equally interesting announcement was about the increase in local meters, since this could start to make sweeps weeks a thing of the past, and hence could start to kill off sweep stunts. Interesting shifts
Jon, good point about the increase in local meters, which I didn't discuss. I think moving beyond the sweeps concept could be of great benefit for the television industry. Enjoyed your analogy as well. With all the new technologies we have in place, it seems that there just have to be much better ways to track impressions, the usefulness of impressions aside. I know that we have a lot of privacy issues to balance against that, but I am hoping that the industry's reputation for slow change doesn't get in the way of the need to find new ways to measure and value television viewing in an increasingly multichannel environment.