1. Studies have shown that people who marry and divorce have, on average, fewer children than do people who stay married, suggesting that having more children makes divorce less likely.
a. Explain why there might be sampling bias in this study.
b. Suppose that the chances of getting divorced are not affected by the number of children you have. Explain how it could still be the case that people who marry and divorce have, on average, fewer children than do people who stay married.
c. What would have to be done to have a controlled experiment?
2. There were 233,467 deaths in California in 2007; the names of these decedents were arranged in alphabetical order. A random number generator was used to select 100 numbers between 1 and 233,467. State whether each of the following sampling procedures is likely to produced biased estimates of the average age at which these people died.
a. The research assistant wrote down only the first 50 random numbers.
b. The research assistant put the random numbers in numerical order; for example, the numbers 7, 3, 9 were changed to 3, 7, 9.
c. The research assistant put the ages of the 100 randomly selected people in numerical order; for example, the people with ages 85, 72, 91 were put in this order: 72, 85, 91.