Question

Part I

A firm that sells a dry skin cream exclusively through drug stores currently operates in 15 marketing districts in the Midwest. As part of an expansion feasibility study, the company wants to model district sales (Y) as a function of target population (X1), per capita income (X2), and number of drug stores (X3) in the region. Data collected for each of the 15 districts were used to fit a first-order linear model. A summary of the regression results is presented below (standard errors of estimated coefficients are in parenthesis). At the 0.05 level, can you conclude that the model is useful for predicting district sales?

Y-hat=3,000+3.2X1+0.4X2+1.1X3

Standard error for B1 = 2.4; B2 = 0.6; B3 = 0.8

R-square = 0.93; MSR= 4,412; MSE= 418.20

a.) No, the F statistic is not significant

b.) No, none of the t values is significant

c.) No, the F value is NOT significant

d.) Yes, the F value is significant

e.) Yes, R-square is very high

Part II

Based on these results, the company concludes that none of the three independent variables is a useful predictor of regional sales, Y. Do you agree with this statement? Choose the best answer.

a.) Yes. The F statistic is not significant

b.) Yes. None of the t values is significant

c.) No. The F statistic is significant

d.) No. The model may have empirical problem(s) that may confound the independent variables' ability to predict regional sales

e.) No. The disturbance term in the model is non-spherical, thus inhibiting the independent variables' ability to predict regional sales

Part III

Without performing any more rigorous evaluation of the results, is/are there any empirical a.problem(s) that might possibly be present in the model?

a.) Yes. The independent variables are likely collinear

b.) Yes. The error term is most likely autocorrelated

c.) Yes. Heteroscedasticity is most likely present in the model

d.) No. The model appears adequate

Answer 1