Another criticism of the measuring system itself is that it fails the most important criterion of a sample: it is not random in the statistical sense of the word. A small fraction of the population is selected and only those that actually accept are used as the sample size. In many local areas of the 1990s, the difference between a rating that kept a show on the air and one that would cancel it was so small as to be statistically insignificant, and yet the show that just happened to get the higher rating would survive.[11] And yet in 2009 of the 114,500,000 U.S. television households[12] only 25,000 total American households (0.02183% of the total) participated in the Nielsen daily metered system.[13] In addition, the Nielsen ratings one TV per household three perhaps four network model encouraged a strong push for demographic measurements. This caused problems with multiple TV households or households where viewers would enter the simpler codes (usually their child's) raising serious questions to the quality of the demographic data.[14] The situation further deteriorated as the popularity of cable TV expanded the number of viewable networks to the point that the margin of error has increased due to the sampling sizes being too small.[15][16] Compounding matters is the fact that of the sample data that is collected, advertisers will not pay for time shifted (recorded for replay at a different time) programs,[17] rendering the 'raw' numbers useless.
A related criticism of the Nielsen ratings system is its lack of a system for measuring television audiences in environments outside the home, such as college dormitories, transport terminals, bars, and other public places where television is frequently viewed, often by large numbers of people in a common setting. In 2005, Nielsen announced plans to incorporate viewing by away-from-home college students into its sample. Internet TV viewing is another rapidly growing market for which Nielsen Ratings fail to account for viewer impact. Apple iTunes, atomfilms, Hulu, YouTube, and some of the networks' own websites (e.g., ABC.com, CBS.com) provide full-length web-based programming, either subscription-based or ad-supported. Though web sites can already track popularity of a site and the referring page, they can't track viewer demographics. To both track this and expand their market research offerings, Nielsen purchased NetRatings in 2007.[18]
