Viewers of American Idol could vote for their favorite singer by using land-line telephones, cell phones, or text-messaging; the winner is presumably the favorite of all Americans.
a. Which of these best describes the potential source of bias: using a self-selected sample, sampling frame not matching population, using a convenience sample, or taking a haphazard sample?
b. In which case would we have a better idea of which singer is preferred by the American public in general-if 10,000 people or if 100,000 people call in their vote; or doesn't it make a difference?