【文章內(nèi)容簡介】 G的數(shù)值是：甲地區(qū)為1，乙地區(qū)為2，丙地區(qū)為3。 統(tǒng)計分析 　激活Statistics菜單選ANOVA Models中的Multivarite...項，彈出Multivarite ANOVA 對話框（）。首先指定供分析用的變量xxx3，故在對話框左側(cè)的變量列表中選變量xxx3，點擊216。鈕使之進入Dependent Variable框；然后選變量g（分組變量）點擊216。鈕使之進入Factor(s)框中，并點擊Define Range鈕，確定g的起始值和終止值。 多元方差分析對話框 點擊Options...鈕，彈出Multivarite ANOVA:Options對話框，選擇需要計算的指標。在Factor(s)欄內(nèi)選變量g，點擊216。鈕使之進入Display Means for框，要求計算平均值指標；在Matriced Within Cell欄內(nèi)選Correlation、Covariance、SSCP項，要求計算單元內(nèi)的相關(guān)矩陣、方差協(xié)方差矩陣和離均差平方和交叉乘積矩陣；在Error Matrices欄內(nèi)也選上述三項，要求計算誤差的相關(guān)矩陣、方差協(xié)方差矩陣和離均差平方和交叉乘積矩陣；在Diagnostics欄內(nèi)選Homogeneity test項，要求作變量的方差齊性檢驗。之后點擊Continue鈕返回Multivarite ANOVA對話框，最后點擊OK鈕即可。 結(jié)果解釋 在結(jié)果輸出窗口中將看到如下分析結(jié)果： 系統(tǒng)首先顯示共90個觀察值進入統(tǒng)計分析，因分組變量g為三個地區(qū)，故分析的單元數(shù)為3。然后輸出3個應變量（xxx3）的方差齊性檢驗結(jié)果，分別輸出了Cochran C檢驗值及其顯著性水平P值、BartlettBox F檢驗值及其顯著性水平P值。其中 身高：C = ，P = ；F = ，P = ； 體重：C = ，P = ；F = , P = ； 胸圍：C = , P = ；F = , P = ； 可見3項指標的方差基本整齊（）。90 cases accepted. 0 cases rejected because of outofrange factor values. 0 cases rejected because of missing data. 3 nonempty cells. 1 design will be processed. CELL NUMBER 1 2 3 Variable G 1 2 3 Univariate Homogeneity of Variance Tests Variable .. X1 Cochrans C(29,3) = .39825, P = .540 (approx.) BartlettBox F(2,17030) = , P = .363 Variable .. X2 Cochrans C(29,3) = .43787, P = .227 (approx.) BartlettBox F(2,17030) = , P = .011 Variable .. X3 Cochrans C(29,3) = .47239, P = .089 (approx.) BartlettBox F(2,17030) = , P = .127 Cochran C檢驗和BartlettBox F檢驗對考查協(xié)方差矩陣的相等性比較方便，但還不夠。于是系統(tǒng)接著分別輸出了三類地區(qū)（即各個單元）各生長發(fā)育指標的離均差平方和交叉乘積矩陣和方差協(xié)方差矩陣。之后作Box M檢驗，Box M檢驗提供矩陣一致性的多元測試，本例Boxs M = ，在基于方差分析的顯著性檢驗中F = ；在基于χ2的顯著性檢驗中χ2 = , 兩者P ，故認為矩陣一致性不佳。Cell Number .. 1 Sum of Squares and CrossProducts matrix X1 X2 X3 X1 X2 X3 VarianceCovariance matrix X1 X2 X3 X1 X2 X3 Cell Number .. 1 (Cont.) Correlation matrix with Standard Deviations on Diagonal X1 X2 X3 X1 X2 .853 X3 .415 .581 Determinant of Covariance matrix of dependent variables = LOG(Determinant) = Cell Number .. 2 Sum of Squares and CrossProducts matrix X1 X2 X3 X1 X2 X3 VarianceCovariance matrix X1 X2 X3 X1 X2 X3 Correlation matrix with Standard Deviations on Diagonal X1 X2 X3 X1 X2 .697 X3 .482 .734 Determinant of Covariance matrix of dependent variables = LOG(Determinant) = Cell Number .. 3 Sum of Squares and CrossProducts matrix X1 X2 X3 X1 X2 X3 VarianceCovariance matrix X1 X2 X3 X1 X2 X3 Correlation matrix with Standard Deviations on Diagonal X1 X2 X3 X1 X2 .680 X3 .595 .886 Determinant of Covariance matrix of dependent variables = LO