Reconsider the wood specific gravity data referred to in Exercise 56.
a. Minitab's Best Regression option was used resulting in the accompanying output. Which model(s) would you recommend for investigating in more detail?

b. The accompanying Minitab output resulted from applying both the backward elimination method and the forward selection method. For each method, explain what occurred at every iteration of the algorithm. Response is spgrav on 5 predictors, with N = 20

Response is spgrav ps S3u r uP P m f f o a a i i o b b sb b ds s R-Sq Vara R-Sq (adj) C-p 1 56.4 53.9 10.6 0.021832 1 10.6 5.7 38.5 0.031245 1 5.3 0.1 41.7 0.032155 2 65.5 61.4 7.0 0.019975 2 62.1 57.6 9.1 0.020950 2 60.3 55.6 10.2 0.021439 X 72.3 67.1 4.9 0.018461 X 71.1 65.75.6 0.018846 X X 74.8 68.1 5.4 0.018179 X X X 3 71.2 65.8 5.6 0.018807 4 77.0 70.9 4.0 0.017353 X X X 72.7 65.4 6.7 0.018919 X X X 5 77.0 68.9 6.0 0.017953X X X X X Response is spgrav on 5 predictors, with N = 20 Step Constant 0.4421 0.4384 0.4381 0.5179 sprngfib 0.00011 0.00011 .00012 T-Value 1.17 1.95 1.98 sumrf1b T-Value 0.00001 0.12 sprwood T-Value 0.00531-0.00526-0.00498-0.004 38 5.20 -5.70 -6.56 5.96 spltabs T-Value -0.001B-0.0019 sumltabs T-Value 0.0044 3.01 0.0044 3.31 0.0031 2.63 0.0027 2.12 0.0180 77.05 0.0174 77.03 0.0185 72.27 0.0200 65.50 R-Sq Step Constant 0.7585 0.5179 sprwood 0.00444 0.00438 T-Value -4.92 -5.20 0.0027 2.12 sumltabs T-Value 0.0218 56.36 0.0200 65.50 R-Sq