Understanding the Nielsen ratings
Understanding the Nielsen ratings involves comprehending a system that measures television viewership to inform programming and advertising decisions. Established by Arthur Nielsen in 1950, the ratings have evolved to reflect the complex media environment of today, where households often own multiple televisions and have access to numerous channels. Nielsen Media Research employs statistical sampling methods to gather data from a representative subset of households, typically around 9,000 nationally, to estimate the viewing habits of the broader population. Key metrics used in this analysis include ratings, which represent the percentage of potential viewers watching a show, and shares, which indicate the percentage of actual viewers watching television at the time.
Nielsen has adapted its methods over the years to account for changes in viewing behavior, such as the rise of digital video recorders (DVRs) and streaming services. Ratings not only affect decisions on renewing or canceling shows but also guide the allocation of billions of dollars in advertising. Despite its importance, the ratings system faces challenges, such as incomplete data from paper diary submissions and the need for more inclusive measures of diverse viewing environments. Overall, Nielsen ratings serve as a crucial indicator of television program performance and viewer engagement in a rapidly changing media landscape.
Understanding the Nielsen ratings
Summary: Television viewing data are estimated using metrics collection and statistical modeling.
The Nielsen Ratings are a measure of how many people are watching certain television programs. When Arthur Nielsen began measuring television viewing in 1950, there were three networks and about 9% of U.S. households had a television. In the twenty-first century, homes often have multiple televisions receiving scores of channels. Even when the number of television sets was small, it was not possible to gather complete viewing data for every single person who owns a television. Instead, Nielsen Media Research uses statistical sampling methods to take a representative subset of viewers and then extrapolates from the sample’s viewing activities to the whole population of viewers. The statistical methods Nielsen uses to collect its data have been refined several times in response to changes in viewer behavior. People who develop and analyze ratings typically have expertise in analytics, metrics, and statistical modeling.
Advanced statistical methodologies, like data mining and software such as Mathematica, are used to extract patterns from Nielsen data that help explain which segments of the population view particular shows. Networks make decisions about whether to cancel or renew programs based on Nielsen ratings. Companies also use Nielsen audience estimates to allocate tens of billions of television advertising dollars each year.
Statistical Sampling and Data Collection
No one knows exactly how many households have televisions, but 2010 estimates suggest that the average U.S. household has just fewer than three televisions. Using statistical sampling, Nielsen can obtain representative data using small a proportion of households: approximately 9000 in its national sample, another 1000 in its Hispanic sample, and various smaller amounts in selected local markets. For the national television sample and major local markets, “people meters” record what television shows household members watch using an electronic set meter, along with a remote control that distinguishes each individual member of the household. Set meters are also used to collect data in mid-sized local markets, but with paper diaries for individual demographics. Meters transmit data to Nielsen every night, where it is checked mathematically for transmission or recording errors before analysis. In the smallest markets, viewers record programs in paper diaries and mail them to Nielsen. Historically, Nielsen tracked only television programs that were viewed live at the time they aired. However, people are increasingly using digital video recorders (DVRs), streaming video, and other delayed viewing technologies, which biases live ratings and affects both programming and advertising decisions. Nielsen began adding DVR households to its sample in 2006 and now regularly reports same-day and seven-day DVR playback ratings as well as its traditional live viewer ratings. People’s failure to return paper diaries is also a growing source of bias, and research methodologists are working on revising this method to make completing the diaries easier to encourage greater response.
Television Metrics
Nielsen’s primary metrics for television viewing are rating, share, and projected audience. A program’s rating is a percentage that represents the number of households that watched the program out of the total number of households that could have watched the program. In this case, the denominator of the fraction is fixed according to the Nielsen sample size. The 1983 finale of the television show M*A*S*H holds the record for highest Nielsen rating, 60.2, which means slightly more than 60% of possible sample households tuned in to watch. At the time there were about 83 million television households, so one sample rating point represented 1% or 830,000 households in the population. However, it would be very unusual for every household to be watching television at the same time.
Share adjusts for this fact by computing the percentage of households that watched a specific program out of the number of households that were actually watching television during that time. This is a more complicated calculation, since the number of televisions being used at any given moment changes constantly. Shares are often used to measure how competitive a program is in its particular time slot. The M*A*S*H finale had a 77 share, which means 77% of households watching television at all were tuned to that program. Ratings and shares are also computed for several age, race, and other subgroups, as these are very important to advertisers. Since data are recorded at the household level—and many people may watch the same program in one house, or outside the home in places like sports bars or dorms—the number of individual viewers in a subgroup or population can only be estimated from the demographic data recorded by people meters and diaries. In 2007, Nielsen also began to measure college students’ viewing habits by treating them as if they were watching an additional television set at home.
Projected audience is the estimated number of people reached in the overall population, which is calculated using statistical modeling. The M*A*S*H finale had a projected audience of 106 million viewers. The number of television households grows every year, as does the number of channel choices, so it can be difficult to compare ratings from across years, especially over large stretches of time. For example, although Super Bowl XLIV surpassed the M*A*S*H finale in terms of estimated viewers (106.5 million), it had a lower rating (46.4). Another reason that the numbers may be difficult to compare is that Nielsen produces rapid overnight ratings for many media outlets and these values are later adjusted. Further, only selected numbers are made public, such as the daily or weekly top 20 shows, which vary from week to week. Comprehensive data is generally available only to Nielsen’s clients, and networks may advertise only the statistics that are most favorable.
Bibliography
Balnaves, Mark, Tom O’Regan, and Ben Goldsmith. Rating the Audience: The Business of Media. New York: Bloomsbury, 2011.
Nielsen Company. “How DVRs are Changing the Television Landscape.” http://blog.nielsen.com/nielsenwire/wp-content/uploads/2009/04/dvr‗tvlandscape‗043009.pdf.
Webster, James, Patricia Phalen, and Lawrence Lichty. Ratings Analysis: The Theory And Practice Of Audience Research. 3rd ed. Mahway, NJ: Lawrence Erlbaum Associates, 2008.