A pollster for the Public Policy Institute of Califor- nia explains how the Institute selects a sample of Califor- nia adults ("It's About Quality, Not Quantity," San Luis Obispo Tribune, January 21, 2000):

That is done by using computer-generated random residential telephone numbers with all California pre?xes, and when there are no answers, calling back repeatedly to the original numbers selected to avoid a bias against hard-to-reach people. Once a call is completed, a second random selection is made by asking for the adult in the household who had the most recent birthday. It is as important to randomize who you speak to in the household as it is to random- ize the household you select. If you didn't, you'd primarily get women and older people.

Comment on this approach to selecting a sample. How does the sampling procedure attempt to minimize certain types of bias? Are there sources of bias that may still be a concern?