A philanthropic organization collects and buys data on its donor base. The full database contains about 4.5 million donors and over 400 variables collected on each, but the data set ch13_MCSP_Donor_Profiles is a sample of 916 donors and includes the following variables:

• Age (in years)

• Homeowner (H 5 Yes, U 5 Unknown)

• Gender (F 5 Female, M 5 Male, U 5 Unknown)

• Wealth (ordered categories of total household wealth from 1 5 Lowest to 9 5 Highest)

• Children (number of children)

• Donated Last (0 5 Did not donate to last campaign, 1 5 Did donate to last campaign)

• Amt Donated Last ($ amount of contribution to last campaign)

The analysts at the organization want to know how much people donate on average to campaigns, and what factors might inf luence that amount. Compare the confidence intervals for the mean Amt Donated Last by those known to own their homes with those whose homeowner status is unknown. Perform similar comparisons for Gender and two of the Wealth categories. Write up a short report using graphics and confidence intervals for what you’ve found. (Be careful not to make inferences directly about the differences between groups. We’ll discuss that in Chapter 14. Your inference should be about single groups.)

(The distribution of Amt Donated Last is highly skewed to the right, so the median might be thought to be the appropriate summary. But the median is $0.00, so the analysts must use the mean. From simulations, they’ve ascertained that the sampling distribution for the mean is unimodal and symmetric for samples larger than 250 or so. Note that small differences in the mean could result in millions of dollars of added revenue nationwide. The average cost of their solicitation is $0.67 per person to produce and mail.)