Question

3. In their paper titled "Do Colleges Get What They Pay For? Evidence on Football Coach Pay and Team Performance", Colbert and Eckard find that higher coach pays leads to improved team performance. And the performance is measured as the average number of points in the Sagarin computer rankings from 2006 to 2011. (Better performance means more points.) A regression table from this paper is presented below. Use the information in the last column to answer the following questions: Table 4. Regression Results-Dependent Variable: Rating (N = I10). Coefficient (t statistic) Independent Variable Coefficient (t statistic) Pay Pay 17.4(9.00)** -2.43(-4.55)** 12.1 (3.03)* -1.86(-2.27)* 1.41 (0.49) -0.021 (-0.29) 0.422(4.54)** 0.011 (1.62) 54.6(30.0)* 722 ea* AQ_BCS Capacity Net_Exp AP Pts Constant R2 54.8(48.5)** 681k Note. AP Associated Press; AQ = automatic qualifying; BCS = Bowl Championship Series; Exp = expenditure. t statistics based on clustered robust standard errors (conference clusters). *5% significance. **1% significance. What is the dependent variable? a. What is the coefficient on "Pay"? Explain what this means. b. Is this coefficient on "Pay" statistically significant at a 5% level? Why or why not? C. d. At the last row of the table, we can see the R2 of the regression is .722, what does this mean?

Answer 1

1. The dependent variable is Team performance which is measured as the average number of points in the sagarin computer rankings. So points is the dependent variable used for regression.

2. The coefficient on pay is 12.1 , this means for an unit increase in pay, the points increase by 12.1

3. Yes the coefficient is statistically significant as the t-value is greater than the critical t-score at 5% level of significance .

4. R²= .722 means 72.2% of the relationship between pay and performance can be explained by the regression equation.