This was all the information I received from the question.

Q2) (10 Points) You suspect that Earning per Share (EPS) and the natural log, Ln of Sales per Share (LnSPS) are related. After you generate a regression using 30 years of quarterly data you want to test for the presence of Heteroscedasticity. You sort the data by LnSPS and break it into three equal samples from Low to High and discard the middle third. a) Based on the following data does the GQ test indicate a problem Lower Revenue Sample: Higher Revenue Sample: Adjusted R = 0.863945 Std Error of Slope = 0.349713 Standard Deviation EPS Standard Deviation Revenue 0.217308 0.070951 b) Regardless of your conclusion above you decide to transform your data and run a WLS regression. Some of the key output is: WLS Regression Coefficent t Values Intercept 0.507052 3.7689 Slope 0.157212 2.1568 If you expect the market to report $2,525 Sales per Share next quarter what is your best estimate of EPS? Q2) (10 Points) You suspect that Earning per Share (EPS) and the natural log, Ln of Sales per Share (LnSPS) are related. After you generate a regression using 30 years of quarterly data you want to test for the presence of Heteroscedasticity. You sort the data by LnSPS and break it into three equal samples from Low to High and discard the middle third. a) Based on the following data does the GQ test indicate a problem Lower Revenue Sample: Higher Revenue Sample: Adjusted R = 0.863945 Std Error of Slope = 0.349713 Standard Deviation EPS Standard Deviation Revenue 0.217308 0.070951 b) Regardless of your conclusion above you decide to transform your data and run a WLS regression. Some of the key output is: WLS Regression Coefficent t Values Intercept 0.507052 3.7689 Slope 0.157212 2.1568 If you expect the market to report $2,525 Sales per Share next quarter what is your best estimate of EPS