13.47 The following regression model was fitted to 28 observations:

yn = 5 - 0.5x1t 10.11452

- 0.04x2t 10.01202

+ 2.725x3t 10.76612 R2 = 0.7329 where yn = operation efficiency (scale 0–10)

x1 = wait times of callers (in minutes)

x2 = number of complaints x3 = number of operators The numbers below the coefficients are the coefficient standard errors.

a. When the independent variable of cloud cover had been taken out from the model, the R2 = 0.4295.

Justify this statement.

b. Test at the 1% level the null hypothesis, all else being equal, for the significance of the coefficient on x1.

c. Determine whether the model has uniform variance over the predicted values at the 5% significance level.

d. Explain what cause heteroskedasticity.