1. How does the chi-square test for the differences in three or more proportions differ from the chi-square test for independence?
2. In a chi-square test for the difference in three proportions, why can't the alternate hypothesis be stated as p(population 1) ? p(population 2) ? p(population 3)?
3. In some of the chi-square tests we studied, there was a requirement that the expected frequencies meet a minimum value. Explain why that is necessary.