ABSTRACT
The ideas behind this paper started to germinate some 15 years before. I recently started at Carnegie Mellon when I was asked to do some work for Mr. Rogers’ Neighborhood, a very popular children’s TV program. (I did not know much about Mr. Rogers; the TV program of my youth was the Howdy Doody Show.)
Mr. Rogers’ Neighborhood sold magazine subscriptions to a publication they produced. They did a survey of their subscribers asking them to rate the various features in the magazine. I was hired to address the question of which features would appeal most to viewers of the show who had not yet subscribed. Thus I had no direct information about the population of interest. I did have one additional piece of information from the subscribers who answered the questionnaire: when they answered. I reasoned that the subscribers who answered quickly are likely to be the most enthusiastic Mr. Rogers supporters. Hence those who delayed their responses would be those most like the target population, the viewers less enthusiastic but still possible subscribers. So I looked at how supportive viewers were of various magazine features as a function of how long it took them to answer the questionnaire, and made my recommendations accordingly. I never did hear from Mr. Rogers’ Neighborhood about how well it worked.
