Method = 1 (Strong Hierarchy) was chosen. The program 1. Standardizes all predictors to have mean 0 and variance 1 and renames them as x1-xp (see table below), numbered as they occur in the data set. 2. Fits main effects, interactions, and squared terms according to a modified forward selection procedure that on average has False Selection Rate (FSR) = 0.05. Developed by Hugh Crews, August 2008. Macro call parameters: Dataset = diabetes Model = age--s6 @q Y = y Gamma = 0.05 Method = 1 Terms = 20 Include = 0 Cbound = none The number of effects forced into the model is 0. Note: The terms=k option was used. Suboptimal results may occur if k is so small that gammahat never reaches its maximum. Variable names Variable Original Obs name name 1 x1 age 2 x2 sex 3 x3 bmi 4 x4 bp 5 x5 s1 6 x6 s2 7 x7 s3 8 x8 s4 9 x9 s5 10 x10 s6 11 y y Forward Sequence Selection Effect p-to- monotonized Obs Step 1+S Entered enter p-to-enter 1 0 1 Intercept <.0001 0.00000 2 1 2 x3 <.0001 0.00000 3 2 3 x9 <.0001 0.00000 4 3 4 x4 <.0001 0.00004 5 4 5 x5 0.0015 0.00145 6 5 6 x3*x4 0.0074 0.00739 7 6 7 x6 0.0049 0.00739 8 7 8 x2 0.0001 0.00739 9 8 9 x2*x4 0.0437 0.04373 10 9 10 x10 0.2254 0.22535 11 10 11 x10*x10 0.0238 0.22535 12 11 12 x4*x10 0.1451 0.22535 13 12 13 x4*x5 0.2825 0.28252 14 13 14 x4*x6 0.2767 0.28252 15 14 15 x6*x9 0.3571 0.35713 16 15 16 x3*x5 0.3515 0.35713 17 16 17 x3*x6 0.3244 0.35713 18 17 18 x3*x9 0.3593 0.35929 19 18 19 x6*x6 0.4200 0.42003 20 19 20 x2*x6 0.5225 0.52246 Sequence of Possible Models Obs alpha gammahat 1+S 1 0.00000 0.00000 1 2 0.00000 0.00000 2 3 0.00000 0.00000 3 4 0.00004 0.00009 4 5 0.00145 0.00349 5 6 0.00739 0.01663 8 7 0.04373 0.10689 9 8 0.22535 0.48826 12 9 0.28252 0.48432 14 10 0.35713 0.44116 17 11 0.35929 0.39921 18 12 0.42003 0.42003 19 13 0.52246 0.47021 20 Fast FSR estimates alpha Obs estimate 1+S 1 0.017391 8 The GLM Procedure Number of Observations Read 442 Number of Observations Used 442 The GLM Procedure Dependent Variable: y y Sum of Source DF Squares Mean Square F Value Pr > F Model 7 1378030.050 196861.436 68.74 <.0001 Error 434 1242979.074 2864.007 Corrected Total 441 2621009.124 R-Square Coeff Var Root MSE y Mean 0.525763 35.17728 53.51642 152.1335 Source DF Type I SS Mean Square F Value Pr > F x3 1 901427.3137 901427.3137 314.74 <.0001 x9 1 302887.7968 302887.7968 105.76 <.0001 x4 1 53985.3203 53985.3203 18.85 <.0001 x5 1 31277.2901 31277.2901 10.92 0.0010 x3*x4 1 21756.4843 21756.4843 7.60 0.0061 x6 1 23599.4572 23599.4572 8.24 0.0043 x2 1 43096.3878 43096.3878 15.05 0.0001 Source DF Type III SS Mean Square F Value Pr > F x3 1 179342.2649 179342.2649 62.62 <.0001 x9 1 301375.5202 301375.5202 105.23 <.0001 x4 1 67960.5993 67960.5993 23.73 <.0001 x5 1 69123.4994 69123.4994 24.14 <.0001 x3*x4 1 28514.9230 28514.9230 9.96 0.0017 x6 1 47235.7450 47235.7450 16.49 <.0001 x2 1 43096.3878 43096.3878 15.05 0.0001 Standard Parameter Estimate Error t Value Pr > |t| Intercept 149.0532781 2.72627676 54.67 <.0001 x3 24.5382015 3.10090490 7.91 <.0001 x9 38.8008819 3.78246270 10.26 <.0001 x4 14.4929663 2.97519700 4.87 <.0001 x5 -37.1911915 7.57032070 -4.91 <.0001 x3*x4 7.8075508 2.47437638 3.16 0.0017 x6 28.2957719 6.96743947 4.06 <.0001 x2 -10.9463978 2.82187542 -3.88 0.0001