Abstract:

Because of the large sample sizes, the variety of market conditions, and the variety of methods used by Arbitron in syndicated audience measurement in the United States, the authors have had a rich opportunity to conduct rigorous experimentation in reducing the biases of both nonresponse error and response error in audience surveys. This paper presents a two-stage process model for the execution of audience search surveys. We describe survey strategies drawn from our experiments and experience in conducting diary surveys, set meter, people meter and personal portable meter panels over more than three decades.