EXAMPLE 18 Selection bias

Suppose Ashley would like to estimate the proportion of American voters who would favor abandoning the present system of Social Security in favor of a system where retirement funds would be invested in the stock market. Ashley goes to the mall with her clipboard and canvasses as many people as she can on Monday between 9 A.M. and 5 P.M. To each person, she asked the question “Do you favor or oppose abandoning the present Social Security system in favor of a system that invests retirement funds in the stock market?”

27

  1. Identify Ashley's target population.
  2. Identify Ashley's potential population.
  3. Discuss any possible problems.

Solution

  1. Ashley's target population is the population of all American voters.
  2. The collection of all the American voters who visited the mall on Monday between 9 A.M. and 5 P.M. represent her potential population.
  3. It appears that Ashley's survey may suffer from selection bias. The population of people who went to the mall on Monday between 9 A.M. and 5 P.M. is not representative of the target population of all American voters. Because many American voters work on Mondays between 9 A.M. and 5 P.M., they are not elements of the potential population. Further, the proportion of retirees at the mall during that time was likely larger than in the target population of all American voters. These retirees tend to oppose strongly any potential changes to the Social Security system and would probably tend to respond in the negative to the survey question.

NOW YOU CAN DO

Exercises 27–30.