Topics Addressed

  1. Sample Surveys
  2. Business

Companies in the U.S spend hundreds of millions of dollars each year advertising their products on television, radio, and in the newspapers, and there is naturally a lot of interest in whether the message is getting through to the right people. Two surveys were conducted some years ago to measure the effect of an advertising campaign for Tide detergent. In one survey, interviewers came to the front door and asked housewives whether they used Tide. In the other, the interviewers asked to actually go into the laundry room and see what detergent was being used.

Would you expect the two surveys to reach similar conclusions?
What kinds of biases might each survey be susceptible to?

Explain briefly.

The Nielsen organization used to rate TV shows by the following method. They chose a panel of homes with TV sets, and attached a meter to each set. The meter recorded the times at which the set was on, and which channel it was tuned to. At the end of each month, the meters were read, and the Nielsen people computed the total number of hours spent by panel members watching each TV show. The Nielsen ratings were based on this total. The panel was seldom changed. The distribution of advertising revenue to the networks was (and still is) based in large part on ratings of this type, with a lot of money hanging in the balance. Using the language of sampling, briefly discuss whether this is a good way to estimate national television-watching habits, and suggest a better way if one occurs to you.


George Michailides