15.13. Consider Table 8.16 on page 245, treating happiness as the response variable. Table 15.24 shows results of fitting the cumulative logit model logit[P(y ≤ j)] =

αj + βx, using scores (1, 2, 3) for income, and the chisquared test of independence.

(a) Why does the table report two intercept estimates?

(b) Report and interpret the income effect.

(c) Using the model, test the hypothesis of no income effect.

Report the test statistic and P-value, and interpret.

(d) Suppose we instead used the Pearson chi-squared test of independence.Report the test statistic and P-value, and compare results to those in (c).Why are they so different?