Section 9.4 (p.545)

1. Based on Equation (6), Q = 7.57. When the null hypothesis H₀ (the distributions of scores in the two cities are the same) is true, Q has approximately a c² distribution with (2-1)(3-1) = 2 df.. Therefore, 0.01 < Pr(Q ³ 7.57) < 0.025. This implies that we reject H₀ at a₀=0.025, but not at a₀=0.01.

4. Here we want to test the homogeneity; the table should be like the following:

	Supplier
	1	2	3
# of defectives	1	7	7
# of non-defective	14	8	8

Q = (1-5)²/5 + (7-5)²/5 + (7-5)²/5 + (14-10)²/10 + (8-10)²/10 + (8-10)²/10 = 7.2

Therefore, we should reject H₀ at a₀=0.05

5. Based on the discussion on “correlated 2x2 tables” (p.544), what we can get are simply the marginal totals as the shown in the following table:

		After demonstration
		Hit	Miss
Before demonstration	Hit			27
	Miss			73
		35	65

To fully construct the table, at least one of the cell has to be known.