Eliminate the least important variable from the model in Exercise 5.3 and recompute the regression. Are all partial sums of squares for the remaining variables significant (α = .05)? If not, continue...

Eliminate the least important variable from the model in Exercise 5.3 and recompute the regression. Are all partial sums of squares for the remaining variables significant (α = .05)? If not, continue to eliminate the least important independent variable at each stage and recomputed

the regression. Stop when all independent variables in the model are significant (use α = .05). What do the results indicate about the need for the intercept? Does it make sense to have β0 = 0 in this exercise? Summarize the results of your final model in an analysis of variance table. Discuss in words your conclusions about what factors are important in peak flow rates.

Exercise 5.3

Use LQ = ln(Q) as the dependent variable and the logarithm of all nine independent variables plus an intercept as the “full” model. Compute the least squares regression equation and test the composite null hypothesis that all partial regression coefficients for the independent variables are zero. Compare the estimated partial regression coefficients to their standard errors. Which partial regression coefficients are significantly different from zero? Which independent variable would you eliminate first to simplify the model?