Using software, analyze the relationship between x = college education and y = percentage single-parent families, for the data in Table 3.6, which are in the U.S. Statewide Crime data file on the text CD.
a. Construct a scatterplot. Based on your plot, identify two observations that seem quite different from the others, having a y value relatively large in one case and somewhat small in the other case.
b. Find the regression equation (i) for the entire data set, (ii) deleting only the first of the two outlying observations, and (iii) deleting only the second of the two outlying observations.
c. Is either deleted observation influential on the slope? Summarize the influence.
d. Including D.C., ŷ = 21.2 + 0.089x, whereas deleting that observation, ŷ = 28.1 - 0.206x. Find ŷ for D.C. in the two cases. Does the predicted value for D.C. depend much on which regression equation is used?

  • CreatedSeptember 11, 2015
  • Files Included
Post your question