The average rating user, when queried as to what he or she wants most from ratings, will probably respond "accuracy," meaning numbers that are free from errors or trustworthy for decision making. On the other hand, we know that all ratings by their very nature are estimates and are therefore subject to errors and variations of several types. The question then becomes not one of absolute accuracy but of relative accuracy. How much accuracy can be obtained at what cost? The expense must be related to the monetary exposure of the decision involved.
Broadcast ratings have employed a wide range of techniques over the years. As we saw in Chapter 1, radio's two first syndicated services the early 1930s-- the Cooperative Analysis of Broadcasting (CAB) using recall and the Hooper coincidentals--both employed telephone interviewing. In the mid 1940s, the Nielsen Audimeter won acceptance, and roster recall utilizing personal interviews was introduced by Pulse.
The success of direct mail in the conduct of station coverage surveys in the 1930s led CBS to experiment with mail diaries in the early 1940s. This prompted the adoption of the mail diary by American Research Bureau in 1948 and later by Nielsen for local market reports.
Today the emphasis is on meters (which register set usage and station tuning) as well as respondent diaries (either household or personal). Other techniques and variants of these basic methods have been used. However, the only successful new methodology currently in use is RADAR, which utilizes a carefully selected