Figure 2-4
*There is no comprehensive list of e-mail addresses.
*True, so we would build a list asking people to apply.
*But half the households don’t have access to the internet.
*The proportion is growing rapidly, and we can compensate by oversampling.
*People who agree to receive questionnaires may not be representative of the general public.
*Yes, but by asking people for their characteristics ( age, race, gender, and so forth ), we can weight their responses accordingly. And just think about it: after the initial computer expenses, this system is fast and inexpensive.
*Yes, but can we say the results are reliable?
*We don’t know much now, so we do need to run some parallel telephone surveys to test the reliability of these initial Internet –based surveys.