| Mass Communication Research Ratings Research |
 |
|
Background
Systematic research came to broadcasting soon after the medium of radio was born in the 1920s. The driving force was the needs of advertisers who wanted to know about the size and composition of audiences. Research developed more rapidly in broadcast media than in print media, because newspapers and magazines could rely on their subscription lists to inform advertisers about their readers.
Ratings Research
Ratings research, which dominate broadcasting research, uses a variety of social science techniques to ascertain the size and characteristics of broadcast audiences. In fact, broadcast researchers are among the most sophisticated in the world when it comes to such things as sampling, statistical analysis and gathering data via questionnaires, diaries, and various electronic devices.
As has been noted, ratings research is driven primarily by the needs of advertisers who use the information to make decisions about how much advertising to buy in which "day parts" or times of day. Advertising representatives who work for broadcast companies need to understand ratings in detail to answer the demands of their clients.
Of course, ratings research is also used by broadcast programmers who decide which programs to buy and when to air them, and by producers who decide what programs to develop and produce.
Ratings research is dominated by a few companies:
- Arbitron does local market radio
- Statistical Research, Inc. does network radio ratings.
- A. C. Nielsen Company does national and local market television. Nielsen also is using its substantial resources and expertise to develop ratings of World Wide Web sites.
Sampling: Broadcast research provides an excellent example of the need to understand sampling principles and apply them well. With nearly 100 million households in America, it would be impossible to do a census to find out who is paying attention to what. It is essential to deal with samples and use them to generalize to the overall population.
The sampling problem is made even more difficult by the fact that ratings must be broken down into markets (geographic areas served by competing media outlets) and care must be taken to represent subgroups such as race, gender and ethnic groups.
For our purposes here it is sufficient to say that ratings researchers use a variety of techniques (multistage sampling) and statistical weighting to derive estimates of audience sizes and compositions.
Data Gathering Techniques:
- Audimeters: Devices that automatically record when a set is turned, to which channels, and for how long.
- Diaries: Used by selected households to record descriptions of who is listening or watching, what station is tuned in, and the day part or time.
- Telephone surveys that ask people about their normal listening or viewing behaviors. These are common in radio research and often involve purposive samples by demographic group, because radio stations are targeted to specific markets.
- Telephone Coincidental Surveys: Telephone surveys in which random samples are interviewed to ask what they are watching or listening to at the moment. These are considered to be more accurate than surveys that rely on recall.
- People Meters that ask all members of chosen households to press a button when they are watching television and automatically records station tuned in, and length of time spent watching. Nielsen says that people meters provide the most accurate television ratings.
Here's a highly simplified example that will help us understand ratings and shares. It is about television ratings that deal with household viewing (contrasted with radio ratings that deal in individual listening.)
We'll keep the arithmetic simple by assuming that we're concerned with a small town with a population of 1,000 households and only three TV stations. Data at a specific day part might be:
| Station | Households Viewing |
| WCAT | 50 |
| WDOG | 150 |
| WSEX | 300 |
| Not Watching | 500 |
| Total Households | 1,000 |
A television rating is the percentage of all households watching a specific station. Thus the rating is the number of households watching a station divided by the total number of household, with the result multiplied by 100 so it is expressed as a percentage rather than a decimal.
To calculate the rating for WCAT divide 50 by 1,000 and multiply the result by 100:
50/1,000 = .05, multiplied by 100, yields a rating of 5
For WDOG the rating is (150/1000)= 15
For WSEX the rating is (300/1000)*100 = 30
Understanding Shares
A television share is the percentage of all households watching television that are watching a specific station. The combined number of all households watching some station is call the HUT, which stands for households using television. In our example this is 500.
This gives a share for WCAT of 10, that is (50/500)*100.
For WDOG the share is 30, that is (150/500)*100.
For WSEX the share is 60, that is (300/500)*100
Radio ratings and shares are calculated in the same way with an important exception: Radio stations focus on individuals rather than households. The total number of persons tuned in to some radio station is called the PUR, which stands for Persons Using Radio.
Thus a radio rating for a given station is the number of individuals listening to that station divided by the number of persons in the market, with the result multiplied by 100.
A radio share for a given station is the number of individuals listening to that station divided by the PUR (number of persons using radio).
Some Things to Remember about Ratings and Shares
- Ratings and shares are estimates derived from samples. Thus various statistical concepts such as standard error and confidence interval apply to them. When we hear things like: "Ten million Americans watched the third presidential debate," we should remind ourselves that these are estimates and that the figure is accurate plus or minus some percentage.
- Ratings are reported in what broadcast professionals call "the books." While it is beyond the scope of this class to learn how to read the book, it's important to know the kinds of data they contain. In general, the books break down the ratings by gender and age group. Therefore it is meaningful to talk about ratings among "females 12 to 17 years old," or "males 18 to 35." Of course, the ratings and shares for demographic subgroups have larger sampling errors than those dealing with the whole population do.
- Ratings are of interest primarily to persons buying and selling advertising (and of course, to station managers who are interested in advertising revenues). This is because they want to know how many people and what kinds of people they are reaching with their advertising dollars.
- Shares are of interest primarily to programmers and program producers as they want to determine how to maximize audiences given the constraints of day parts. Obviously there are more people available to attend media during prime time than there are during the workday. Programmers want to schedule and produce programs that will appeal to the people who are available at specific times. That's why cartoons dominant Saturday morning, sports are on weekends, and soap operas on weekdays.
Other Concepts from Ratings Research
There are dozens of concepts that would be needed for a full understanding of ratings research and we won't try to cover all of them here. However, a few are important enough for all communication professionals to know. They include:
- Cost per Thousand, abbreviated CPM (remember "M" is the Roman numeral for 1,000). This is a convenient index for advertisers that indicates the efficiency of their advertising. There's nothing mysterious about Cost Per Thousand. It is the cost of an advertisement divided by the audience size in thousands.
- Cume (cumulative audience, a.k.a. reach) is an estimate of the number of persons who attended at least five minutes in a particular day part across a week. Sometimes referred to as "unduplicated audience," it lets advertisers know how many people are likely to see an advertisement broadcast repeatedly at a specific time of day. A person who watches the same program every day would count only once in the cume.
- Gross Rating Point (GRP) is the combined rating of two or more day parts, representing the size of the gross audience. In contrast to the cume, it is of interest to the advertiser who buys time for multiple commercials broadcast at different times of day.
Non-Ratings Research in Broadcasting
Ratings research accounts for the vast majority of research in broadcasting; however, broadcast researchers are sophisticated users of social science techniques and apply their knowledge to a variety of other problems. Broadcast researchers use many approaches to find out such things as:
- Which proposed television dramas are likely to draw an audience.
- What kinds of people are attracted to what kinds of radio station music formats, and,
- How much appeal celebrities have as commercial spokespersons.
In addition to the standard data gathering approaches of sample surveys, content analysis and experiments, broadcast researcher have developed approaches that are specific to the problems they face. In all cases, they are aware of such concepts as sampling error, questionnaire construction, experimental design and various kinds of validity.
If you don't
understand something in this Web note, please e-mail
Dr. Sitton.
Revised 083011 — http://www.uamont.edu/FacultyWeb/sitton/crz/mrea/rating.html