In: Statistics and Probability
A sophisticated economic model of unemployment predicts that if welfare is reduced by 20%, the proportion of unemployed people should decrease by 6 percentage points. We want to test this prediction based on data collected prior and after a welfare reform that reduced welfare by 20%: We have a sample of n1 = 2524 randomly selected people, of which 425 are unemployed, collected before welfare reform, and we have a sample of n2 = 1278 randomly selected people, of which 161 people are unemployed, collected after welfare reform. Denote by p1 the unemployment rate in the population prior to welfare reform, and let p2 the unemployment rate (in the population) after welfare reform.
(a) State the implication of the economic theory as a hypothesis about the relationship between p1 and p2.
(b) Find a test statistic whose distribution under the null hypothesis is (approximately) standard normal.
(c) Can you reject the validity of the economic theory on 5% level against a two-sided alternative?