1中正大學運休所應用統計學20051229 (獨立樣本)的單因子變異數分析 data onea; input a heart; label a=a因子 heart=心肺指數; cards; 151.4 145.6 152 140.2 142.3 165.2 140 152.9 168.7 1 55.6 2 67.1 2 62.9 2 55.2 252.9 267.1 257.3 252.9 261.6 256.8 262.1 3 54.2 3 75 3 62.9 3 69.8 3 46.4 3 66.2 3 54.2 3 70.9 3 75 3 55 ; proc anova data=onea; class a; model heart=a; means a / lsd tukey scheffe cldiff; proc glm data=onea; class a; model heart=a; means a / lsd tukey scheffe cldiff; run;重複量數(相依樣本)的單因子變異數分析 data aa; input s ; do a=1 to 3; input heart ; output; end; label s=受試者 a=a因子 heart=心肺指數; cards; 151.467.154.2 245.662.975 35255.262.9 440.252.969.8 542.367.146.4 665.257.366.2 74052.954.2 852.961.670.9 968.756.875 10 55.662.155 ; proc anova data=aa; class s a; model heart = a s a*s; test h=a e=a*s; means a / tukey; run;受試者組別心肺指數 1 2 3 4 5 6 7 8 9 101 1 1 1 1 1 1 1 1 151.4 45.6 52 40.2 42.3 65.2 40 52.9 68.7 55.6 11 12 13 14 15 16 17 18 19 202 2 2 2 2 2 2 2 2 267.1 62.9 55.2 52.9 67.1 57.3 52.9 61.6 56.8 62.1 21 22 23 24 25 26 27 28 29 303 3 3 3 3 3 3 3 3 354.2 75 62.9 69.8 46.4 66.2 54.2 70.9 75 55受試者心肺指數 1心肺指數 2心肺指數 3 1 2 3 4 5 6 7 8 9 1051.4 45.6 52 40.2 42.3 65.2 40 52.9 68.7 55.667.1 62.9 55.2 52.9 67.1 57.3 52.9 61.6 56.8 62.154.2 75 62.9 69.8 46.4 66.2 54.2 70.9 75 55人數不相等時採用 glm人數相等時採用 anova重複量數基本上是人數相等的 如果有遺漏值則採用 glm 進行分析2中正大學運休所應用統計學20051229The ANOVA ProcedureClass Level InformationClassLevelsValuesa31 2 3Number of Observations Read30 Number of Observations Used30The ANOVA ProcedureDependent Variable: heart心肺指數Sum of SourceDFSquaresMean SquareF ValuePr FModel2708.206000354.1030004.690.0179Error272039.06200075.520815Corrected Total292747.268000R-SquareCoeff VarRoot MSEheart Mean0.25778614.988398.69027157.98000SourceDFAnova SSMean SquareF ValuePr Fa2708.2060000354.10300004.690.0179The ANOVA Proceduret Tests (LSD) for heartNOTE: This test controls the Type I comparisonwise error rate, not the experimentwise error rate.Alpha0.05 Error Degrees of Freedom27 Error Mean Square75.52081 Critical Value of t2.05183 Least Significant Difference7.9742Comparisons significant at the 0.05 level are indicated by *.Difference aBetween95% Confidence ComparisonMeansLimits3 - 23.370-4.60411.344 3 - 111.5703.59619.544* 2 - 3-3.370-11.3444.604 2 - 18.2000.22616.174* 1 - 3-11.570-19.544-3.596* 1 - 2-8.200-16.174-0.226*The ANOVA ProcedureTukeys Studentized Range (HSD) Test for heart3中正大學運休所應用統計學20051229 NOTE: This test controls the Type I experimentwise error rate.Alpha0.05 Error Degrees of Freedom27 Error Mean Square75.52081 Critical Value of Studentized Range3.50643 Minimum Significant Difference9.636Comparisons significant at the 0.05 level are indicated by *.Difference aBetweenSimultaneous 95% ComparisonMeansConfidence Limits3 - 23.370-6.26613.006 3 - 111.5701.93421.206* 2 - 3-3.370-13.0066.266 2 - 18.200-1.43617.836 1 - 3-11.570-21.206-1.934* 1 - 2-8.200-17.8361.436The ANOVA Procedure 人數不等時採用scheffe事後比較法 Scheffes Test for heartNOTE: This test controls the Type I experimentwise error rate, but it generally has a higher Type II error rate than Tukeys for all pairwise comparisons.Alpha0.05 Error Degrees of Freedom27 Error Mean Square75.52081 Critical Value of F3.35413 Minimum Significant Difference10.066Comparisons significant at the 0.05 level are indicated by *.Difference aBetweenSimultaneous 95% ComparisonMeansConfidence Limits3 - 23.370-6.69613.436 3 - 111.5701.50421.636* 2 - 3-3.370-13.4366.696 2 - 18.200-1.86618.266 1 - 3-11.570-21.636-1.504* 1 - 2-8.200-18.2661.866The GLM ProcedureClass Level InformationClassLevelsValuesa31 2 3Number of Observations Read30 Number of Observations Used30The GLM ProcedureDependent Variable: heart心肺指數4中正大學運休所應用統計學20051229Sum of SourceDFSquaresMean SquareF ValuePr FModel2708.206000354.1030004.690.0179Error272039.06200075.520815Corrected Total292747.268000R-SquareCoeff VarRoot MSEheart Mean0.25778614.988398.69027157.98000SourceDFType I SSMean SquareF ValuePr Fa2708.2060000354.10300004.690.0179SourceDFType III SSMean SquareF ValuePr Fa2708.2060000354.10300004.690.0179The GLM Proceduret Tests (LSD) for heartNOTE: This test controls the Type I comparisonwise error rate, not the experimentwise error rate.Alpha0.05 Error Degrees of Freedom27 Error Mean Square75.52081 Critical Value of t2.05183 Least Significant Difference7.9742Comparisons significant at the 0.05 level are indicated by *.Difference aBetween95% Confidence ComparisonMeansLimits3 - 23.370-4.60411.344 3 - 111.5703.59619.544* 2 - 3-3.370-11.3444.604 2 - 18.2000.22616.174* 1 - 3-11.570-19.544-3.596* 1 - 2-8.200-16.174-0.226*The GLM ProcedureTukeys Studentized Range (HSD) Test for heartNOTE: This test controls the Type I experimentwise error rate.Alpha0.05 Error Degrees of Freedom27 Error Mean Square75.52081 Critical Value of Studentized Range3.50643 Minimum Significant Difference9.636Comparisons significant at the 0.05 level are indicated by *.5中正大學運休所應用統計學20051229Difference aBetweenSimultaneous 95% ComparisonMeansC