Roger Hill is an analyst covering equity stocks and he is interested in studying the effect that a company's reputation (REP), earnings quality (EarQual), and size (MktCap) has on its cost of capital (CostCap). REP is ranked on a scale of 1 to 10, with 10 being highest reputation. EarQual is ranked on a scale of 1 to 5, with 5 being highest earnings quality. MktCap is market capitalization in $ billions. Hill has collected data for 40 quarters from 2000-2010. Regression results and ANOVA results are summarized in Exhibits 1 and 2 (below). Hill has also collected t-distribution data and Durbin-Watson (DW) test statistic data which is shown in Exhibit 3 (below). Hill intends to use a 5% significance level in his analysis. Use the information here and in the Exhibits to answer the questions that follow:

Exhibit 1: Multiple Regression Results

VARIABLE	COEFFICIENT	STANDARD ERROR	t-STATISTIC
Intercept	+0.0558	0.0014	39.86
REP	-0.2065	0.0131	-15.76
EarQual	-0.3755	0.0239	-15.71
MktCap	-0.1758	0.0020	-87.9

Exhibit 2: ANOVA Results

	Sum of Squares	Significant F
Regression	1,465
Residual	12,050
Total	13,515
	Multiple R-squared = 0.31
	Durbin-Watson (DW) = 1.32

Exhibit 3: Selected Values for the t-distribution and the Durbin-Watson (DW) Statistic (n=40)

Selected t-distribution Values	Area in Right Tail	t-value
	0.050	1.684
	0.025	2.021
Durbin-Watson Statistics	dl	du
K = 2	1.39	1.60
K = 3	1.34	1.66
K = 4	1.29	1.72

1. Set up the regression equation for this regression.
2. What critical t-value from Exhibit 3 should Hill use when testing the significance of the regression coefficients?
3. Using the results in Exhibits 1 and 3, what will Hill most likely conclude regarding the effect of reputation on the cost of capital? What about the effect of earnings quality on the cost of capital? What about the effect of market capitalization on the cost of capital?
4. Using the information in Exhibit 2, what is the F-statistic to test the significance of the regression?
5. Explain the test situation where you would use a pooled vs an un-pooled variance test.
6. Explain the test situation where you would use a mean difference vs. a difference in means test.