1. Despite the positive correlation coefficient between A and B, the regression of A on B has a negative slope. Could this situation occur? Explain why or why not.
2. In an analysis of the SAT data, 1,000 is mistakenly added to every math score. For example, a Math score of 650 becomes 1650. What impact will this have on the slope coefficient in a bivariate regression of Math SAT on Verbal SAT as compared with the slope coefficient from the same biavariate regression using the original correct data? What impact will this mistake have on the slope coefficient in the regression of Verbal SAT on Math SAT? Explain your reasoning.