CAB was a marketing research organization. Its interest was not just in radio listening but in what advertisers in general got out of advertising. Radio program ratings were but one means of answering a number of questions and concerns that CAB addressed. Radio program ratings provided an answer to the question of how large an audience was. During this period, radio advertisers were also program producers, purchasing time periods from the radio networks and filling these periods with their own programs. Sponsors used CAB information about radio's listening audience to build programs that attracted the type of listeners most likely to buy their products. Because radio audiences were broadly based and were not already predetermined as magazines were by editorial policy, most advertisers concentrated merely on attracting large numbers of listeners.

Origins

     Because national advertisers were some of the first companies to explore the use of radio as a means of advertising, it was perhaps not surprising that the Association of National Advertisers (ANA), through its radio commission (or committee), made the first attempt to answer basic questions regarding radio's anonymous audience and the national advertisers' possible customers. Because advertisers created programs and purchased time from networks, the first question that served as the focus for the first generation of audience research was that of network programs' relative popularity: what network programs were the most popular; that is, which programs drew the largest audiences relative to other programs? This question of program ratings served as the focus for CAB.

     Crossley had started Crossley Inc. in 1918; it was one of three major organizations (the others being Roper and Gallup) that engaged in political polling. Political polling results were ranked by percentage points, much like ratings. When the ANA returned to Crossley to repeat an earlier audit, Crossley suggested that he should instead study radio's listening audience and listeners' program preferences. Although the radio committee of the ANA decided not to finance the study, the ANA did agree to endorse it if Crossley would underwrite it. By 1929 Crossley had gained endorsement from 30 sponsors, and by 1930 he began field work. Cross­ ley's service functioned primarily as a network program rating service. Crossley, in fact, coined the term rating, although initially he used only the "identified listening audience," or what is now called the share, as the base for his ratings. Because national advertisers and agencies had developed Crossley's service, it initially served the 33 cities where the radio networks had outlets.

CAB's Method: The Telephone Recall

     Although Crossley used a variety of techniques for different clients, including printed roster, mechanical recorder, personal interview, and coincidental, he selected a simple "next-day recall" telephone method to provide the first regular measure of network program audiences. The coincidental method employs an interview method in which the respondent is asked to state his or her listening to radio at the precise time ("coincidental" with) the interview. By comparison, next-day recall required respondents to recall behavior for the previous 24 hours. The recall method raised questions about respondent memory. Crossley's technique was to dial a telephone­ based list and interview respondents about their previous day's listening.

     Crossley chose the telephone recall method for four primary reasons. First, radio and telephone ownership had originally exhibited a high degree of congruence when he began his radio work in 1929-though radio homes would soon outstrip those with telephones. Second, telephones covered a wide area quickly. Third, recall meant that a great deal of information about listeners could be collected at little expense. Finally, sponsor identification was an important concern during the period. As Daniel Yankelovich noted in 1938, for many years advertisers held the belief that recall or registration was the most useful index of advertising effectiveness. Because most sponsors or agencies typically developed their own programs during this period or at least sponsored an entire program, advertisers sought to know the degree of registration between the programs and their product. Recall measured conscious impressions. The chief drawback, of course, was that not all homes had telephones. By the late 1930s, twice as many homes had radios as owned telephones, meaning a large proportion of the radio audience was not reached by CAB researchers. Still, over the years Crossley's sample size grew four-fold from its starting point of about 100,000 households.

Enter Hooperatings

     Crossley's service, under stiff competition after 1934 with the up-and-coming Hooperatings service, gradually lost ground. The foundation of Crossley's survey technique was originally quota sampling. This was a type of nonprobability sampling, meaning that the degree of sampling error could not be calculated (sampling error is the difference between the sample's results and the results that would be obtained if the whole population were surveyed). Quotas were set as to the number of respondents of varying demographics, such as geographical areas, age, gender, economic levels, and so forth that interviewers were to obtain. The goal in setting quotas was to ensure that the sample was distributed with respect to these characteristics in proportion to presumably known population totals. In response to attacks from C.E. Hooper of Hooperatings, Cross­ ley changed to random sampling, a type of probability sample in which sampling error could be calculated. A random sample ensured that each unit used in the sample had an equal chance or probability of being selected. He also shifted to the coincidental method, or having interviewers call while the program audience was still tuned to their radios. Hooper also questioned CAB's tabulation procedures, the phrasing of questions asked, and distribution methods.

     As the authority of CAB was whittled away by the compiler of a more convincing set of statistics, backers of CAB became alarmed. By January 1946 Hooper had substantial subscribers, and industry backers saw CAB's switch from telephone recall to Hooper's method of telephone coincidental as rendering CAB superfluous. The final straw for CAB was the financial withdrawal of ABC, CBS, and NBC, which left MBS as the only radio network member. The four networks had provided 40 percent of CAB revenues. The three advertising association backers, ANA, AAAA, and NAB, were left holding the bag. An attempt was made to cover this large operating cost gap through increasing dues and assessment. However, CAB had gradually lost more and more ground to Hooper, and in June 1946 CAB suspended its 17-year-old service.

     Run as a cooperative membership organization, CAB was operated by a board of governors consisting of advertisers, agencies, and broadcasters. This cooperative structure led to a crucial difficulty that impaired CAB. This structure had hindered its efficient operation, making it much less responsive to the marketplace than Hooper. The committee's divergent ideas and politics led to a bureaucratic situation that Crossley called "too many chiefs." Its committee structure meant that decisions were long in the making; decisions were typically promises and were tied to interest rather than economic considerations; furthermore, results were not measured in terms of profits or loss in the marketplace. Hooper's private enterprise, on the other hand, was conscious of cost, was aware of the degree of acceptance from its clients, and was more responsive in general to the marketplace.

     Both CAB's and Hooper's rating services were rating indices limited to a handful of urban cities. Neither provided national size estimates of the number of listeners to a given program, but rather were limited to comparative figures. Neither was designed to measure a true national program audience but was instead limited to urban telephone homes.