For this problem, use the crime rate data set from KNNL Problem 1.28.

a. Find the least squares regression equation to predict the crime rate from the percent of individuals having at least a high school education. [Paste R or SAS output and then answer your question]

b. Give the ANOVA Table for this regression analysis.[Paste R or SAS output]

c. Find SSE and MSE for this model.

d. What is the estimate of from this analysis?

e. What percent of the variation in crime rates can be explained by the percent of high school graduates?

f. What is the correlation between crime rates and percent of high school graduates?

g. Based on your ANOVA table, is the linear relationship between X and Y statistically significant? Be sure to give an appropriate null and alternate hypothesis, test statistic, its associated degrees of freedom, and the p-value.

h. Give a scatter plot of crime rates vs. percent of high school graduates, with the regression line. Comment about linearity

i. Give the Residual Plot (residuals vs. fitted values). Test for Non-Linear and Non-constant variance.

j. Conduct Breusch-Pagan Test for the constancy of the error variance. Be sure to give an appropriate null and alternate hypothesis, test statistic, its associated degrees of freedom, and the p-value.

k. Brown-Forsyth test for constant variance [ skip this]

l. Index Plot to test for Independence of errors.

m. Conduct Durbin-Watson Test. Be sure to give an appropriate null and alternate hypothesis, test statistic, its associated degrees of freedom, and the p-value.

n. Outlier deduction test [Plot standardized Residuals versus fitted values]

o. Give a Histogram of the residuals.

p. Give a QQ-plot of the residuals to test for normality of error terms.

q. Conduct a Shapiro-Wilk Test on the residuals. Be sure to give an appropriate null and alternate hypothesis, test statistic and the p-value. Give the p-value for this test and explain what this means in terms of our model assumptions.