Do the following for each of the four data sets:
a. Create a scatterplot of y versus x with a regression line ( = b0 + b1x), a plot of residuals versus the explanatory variable, a plot of residuals versus predicted (or “fitted”) values (y), and either a normal probability plot or a histogram of the residuals.
b. By hand, sketch on the scatterplot a curve that would fit the data better than the regression line.
c. Try various transformations of the explanatory and response variables to create a better linear regression model (as validated by graphical analysis of the residuals).