第六章
方差分析
方差分析是R.A.Fister发明的,用于两个及两个以上样本均数差别的显著性检验。由于各种因素的影响,研究所得的数据呈现波动状,造成波动的原因可分成两类,一是不可控的随机因素,另一是研究中施加的对结果形成影响的可控因素。方差分析的基本思想是:通过分析研究中不同来源的变异对总变异的贡献大小,从而确定可控因素对研究结果影响力的大小。
方差分析主要用于:1、均数差别的显著性检验,2、分离各有关因素并估计其对总变异的作用,3、分析因素间的交互作用,4、方差齐性检验。
第一节
Simple Factorial过程
6.1.1
主要功能
调用此过程可对资料进行方差分析或协方差分析。在方差分析中可按用户需要作单因素方差分析(其结果将与第五章第四节相同)或多因素方差分析(包括医学中常用的配伍组方差分析);当观察因素中存在有很难或无法人为控制的因素时,则可对之加以指定以便进行协方差分析。
6.1.2
实例操作
[例6-1]下表为运动员与大学生的身高(cm)与肺活量(cm3)的数据,考虑到身高与肺活量有关,而一般运动员的身高高于大学生,为进一步分析肺活量的差异是否由于体育锻炼所致,试作控制身高变量的协方差分析。
运 动 员 |
大 学 生 | ||
身高 |
肺活量 |
身高 |
肺活量 |
184.9 167.9 171.0 171.0 188.0 179.0 177.0 179.5 187.0 187.0 169.0 188.0 176.7 179.0 183.0 180.5 179.0 178.0 164.0 174.0 |
4300 3850 4100 4300 4800 4000 5400 4000 4800 4800 4500 4780 3700 5250 4250 4800 5000 3700 3600 4050 |
168.7 170.8 165.0 169.7 171.5 166.5 165.0 165.0 173.0 169.0 173.8 174.0 170.5 176.0 169.5 176.3 163.0 172.5 177.0 173.0 |
3450 4100 3800 3300 3450 3250 3600 3200 3950 4000 4150 3450 3250 4100 3650 3950 3500 3900 3450 3850 |
6.1.2.1 数据准备
激活数据管理窗口,定义变量名:组变量为group(运动员=1,大学生=2),身高为x,肺活量为y,按顺序输入相应数值,建立数据库,结果见图6.1。
图6.1 原始数据的输入 |
6.1.2.2 统计分析
激活 Statistics 菜单选ANOVA
Models中的Simple Factorial...项,弹出Simple Factorial
ANOVA对话框(图6.2)。在变量列表中选变量y,点击Ø钮使之进入Dependent框;选分组变量group,点击Ø钮使之进入Factor(s)框中,
并点击Define Range...钮在弹出的Simple Factorial ANOVA:Define
Range框中确定分组变量group的起止值(1,2);选协变量x,点击Ø钮使之进入Covariate(s)框中。
图6.2 协方差分析对话框 |
点击Options...框,弹出Simple
Factorial ANOVA:Options对话框。系统在协方差分析的方法(Method)上有三种选项:
1、Unique:同时评价所有的效应;
2、Hierarchical:除主效应外,逐一评价各因素的效应;
3、Experimental:评价因素干预之前的主效应。
本例选Unique方法,之后点击Continue钮返回Simple Factorial
ANOVA对话框,再点击OK钮即可。
6.1.2.3 结果解释
在结果输出窗口中可见如下统计数据:
先输出肺活量总均数和两组的肺活量均数,总均数为4033.25,运用员组均数为4399.00,大学生组为3667.50。
接着协方差分析表明,混杂因素X(身高)两组间是有差异的(F=10.679,P=0.002),控制其影响后,两组间肺活量的差别依然存在(F=9.220,P=0.004),故可以认为两组间肺活量的均数在消除了身高因素的影响之后仍有差别,运动员的肺活量大于大学生,即体育锻炼会提高肺活量。
最后系统输出公共回归系数,=
36.002,该值可用于求修正均数:
= -
(-)
本例为=
4399.00 - 36.002×(178.175 - 174.3325)=
4260.6623
= 3667.50 - 36.002×(170.49 - 174.3325)=
3805.8377
Y by GROUP Total
Population 4033.25 (
40)
GROUP 1
2
4399.00
3667.50 ( 20) (
20) Y by GROUP with
X
UNIQUE sums of squares
All effects entered simultaneously
Sum of
Mean
Sig Source of Variation
Squares
DF
Square
F of
F Covariates
1630763 1 1630762.635 10.679 .002 X
1630763 1 1630762.635 10.679 .002 Main Effects
1407847 1 1407847.095 9.220 .004 GROUP
1407847 1 1407847.095 9.220 .004 Explained
6981685 2 3490842.568 22.860 .000 Residual 5649992 37
152702.496
Total
12631678
39
323889.167
40 cases were
processed. 0 cases (.0 pct) were
missing. Covariate Raw Regression
Coefficient X
36.002
|
第二节
General Factorial过程
6.2.1
主要功能
调用此过程可对完全随机设计资料、配伍设计资料、析因设计资料、正交设计资料等等进行多因素方差分析或协方差分析。
6.2.2
实例操作
[例6-2]下表为三因素析因实验的资料,请用方差分析说明不同基础液与不同血清种类对钩端螺旋体的培养计数的影响。
基础液 (A) |
血清种类(B) | |||
兔血清浓度(C) |
胎盘血清浓度(C) | |||
5% |
8% |
5% |
8% | |
缓冲液 |
648 1246 1398 909 |
1144 1877 1671 1845 |
830 853 441 1030 |
578 669 643 1002 |
蒸馏水 |
1763 1241 1381 2421 |
1447 1883 1896 1926 |
920 709 848 574 |
933 1024 1092 742 |
自来水 |
580
1026 1026 830 |
1789 1215 1434 1651 |
1126 1176 1280 1212 |
685 546 595 566 |
6.2.2.1 数据准备
激活数据管理窗口,定义变量名:基础液为base,血清种类为sero,血清浓度为pct,钩端螺旋体的培养计数为X,按顺序输入相应数值,建立数据库。
6.2.2.2 统计分析
激活Statistics菜单选ANOVA
Models中的General Factorial...项,弹出General Factorial
ANOVA对话框(图6.3)。在对话框左侧的变量列表中选变量x,点击Ø钮使之进入Dependent
Variable框;选要控制的分组变量base、sero和pct,点Ø钮使之进入Factor(s)框中,并分别点击Define Range钮,在弹出的General Factorial ANOVA:Define
Range对话框中确定各变量的起止值,本例变量base的起止值为1、3,变量sero的起止值为1、2,变量pct的起止值为1、2。之后点击OK钮即可。
图6.3 析因方差分析对话框 |
6.2.2.3 结果解释
在结果输出窗口中,系统显示48个观察值进入统计,三个因素按其各自水平共产生12种组合。
分析表明,模型总效应的F值为10.55,P值 <
0.001,说明三因素间存在有交互作用。单因素效应和交互效应导致的组间差别比较结果是:
单因素组间比较:
A:基础液(BASE)
F =
4.98,P = 0.012,说明三种培养基培养钩体的计数有差别;
B:血清种类(SERO)
F =
61.265,P < 0.001,说明两种血清培养钩体的计数有差别;
C:血清浓度(PCT)
F =
3.49,P = 0.070,说明两种血清浓度培养钩体的计数无差别。
两因素构成的一级交互作用:
A×B:基础液(BASE)×血清种类(SERO)
F =
5.16,P = 0.011,交互作用明显;
B×C:血清种类(SERO)×血清浓度(PCT)
F =
15.96,P < 0.001,交互作用明显;
A×C:基础液(BASE)×血清浓度(PCT)
F =
0.78,P = 0.465,交互作用不明显。
三因素构成的二级交互作用:
A×B×C:基础液(BASE)×血清种类(SERO)×血清浓度(PCT)
F = 6.75,P
= 0.003,交互作用明显。
48 cases
accepted. 0 cases rejected because of
out-of-range factor values. 0 cases rejected because of missing
data. 12 non-empty
cells. 1 design will be
processed. - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Univariate Homogeneity of Variance
Tests Variable .. X
Cochrans C(3,12) =
.34004, P
= .036
(approx.)
Bartlett-Box F(11,897) = 1.69822, P
= .069 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - * * * * * * A n a l y s i
s o f V a r i a n c e --
design 1 * * * * *
* Tests of Significance for X using
UNIQUE sums of squares Source of Variation
SS
DF
MS
F
Sig of F WITHIN+RESIDUAL
2459233.75 36
68312.05 BASE
679967.38
2 339983.69 4.98
.012 PCT
238713.02
1 238713.02 3.49
.070 SERO
4184873.52
1 4184873.5 61.26
.000 BASE BY PCT
107005.54
2
53502.77
.78
.465 BASE BY SERO
705473.04
2 352736.52 5.16
.011 PCT BY SERO
1089922.69
1 1089922.7 15.96
.000 BASE BY PCT BY SERO 922307.37 2 461153.69 6.75
.003 (Model)
7928262.56 11 720751.14 10.55
.000 (Total)
10387496.31 47
221010.56
R-Squared =
.763 Adjusted R-Squared = .691
|
第三节
Multivarite过程
6.3.1
主要功能
调用此过程可进行多元方差分析。此外,对于一元设计,如涉及混合模型的设计、分割设计(又称列区设计)、重复测量设计、嵌套设计、因子与协变量交互效应设计等,此过程均能适用。
6.3.2
实例操作
[例6-3]甲地区为大城市,乙地区为县城,丙地区为农村。某地分别调查了上述三类地区8岁男生三项身体生长发育指标:身高、体重和胸围,数据见下表,问:三类地区之间男生三项身体生长发育指标的差异有无显著性?
学生编号 |
甲地区 |
乙地区 |
丙地区 | ||||||
身高 |
体重 |
胸围 |
身高 |
体重 |
胸围 |
身高 |
体重 |
胸围 | |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 |
119.80 121.70 121.40 124.40 120.00 117.00 118.10 118.80 124.20 124.90 124.70 123.00 125.30 124.20 127.40 128.20 126.10 128.70 129.50 126.90 126.50 128.20 131.40 130.80 133.90 130.40 131.30 130.20 136.00 141.00 |
22.60 21.50 19.10 21.80 21.40 20.10 18.80 22.00 21.30 24.00 23.30 22.50 22.90 19.50 22.90 22.30 22.70 23.50 24.50 25.50 25.00 26.10 27.90 26.80 27.20 24.40 24.40 23.00 26.30 31.90 |
60.50 55.50 56.50 60.50 57.70 57.00 57.10 61.70 58.40 60.80 60.00 60.00 65.20 53.80 59.50 60.00 57.40 60.40 51.00 61.50 63.90 63.00 63.10 61.50 65.80 62.60 59.50 62.60 60.00 63.70 |
125.10 127.00 125.70 114.90 124.90 117.60 124.20 117.90 120.40 115.00 126.20 125.10 114.90 121.50 114.00 118.70 120.60 122.90 119.60 112.30 121.30 121.20 120.20 120.30 120.00 123.30 122.10 123.30 109.90 125.60 |
23.00 21.50 23.40 17.50 23.50 18.90 20.80 20.30 20.00 19.70 21.20 22.10 19.70 22.00 19.00 19.10 20.00 18.50 19.50 20.00 20.00 21.20 23.10 21.00 22.20 20.10 21.00 21.50 17.80 23.30 |
62.00 59.00 61.50 52.50 58.50 57.00 58.50 61.00 56.00 56.50 56.50 58.50 56.00 57.00 54.50 54.50 55.50 56.00 59.50 58.00 58.00 59.00 59.50 59.50 59.50 56.50 57.50 61.00 56.50 60.50 |
118.30 121.30 121.80 124.20 123.50 123.00 134.90 123.70 105.20 112.20 118.60 112.00 121.50 124.50 119.50 122.50 115.50 122.50 124.50 125.00 117.50 127.30 122.30 121.30 120.50 116.00 120.50 114.50 131.00 122.50 |
20.40 20.00 26.60 22.10 23.20 22.90 32.30 22.70 20.20 20.80 21.00 23.20 24.00 21.50 20.50 23.00 19.00 22.50 25.00 25.50 23.00 22.50 22.00 21.00 22.00 19.00 20.00 19.00 25.50 24.50 |
54.40 54.30 61.10 58.60 60.20 58.20 64.80 59.90 54.50 57.50 57.60 58.20 60.30 55.60 55.50 56.70 54.20 57.60 57.90 60.30 59.00 58.90 58.20 55.60 55.10 53.50 54.40 53.40 58.30 58.70 |
6.3.2.1 数据准备
激活数据管理窗口,定义变量名:地区为G,身高为X1,体重为X2,胸围为X3,按顺序输入相应数值,变量G的数值是:甲地区为1,乙地区为2,丙地区为3。
6.3.2.2 统计分析
激活Statistics菜单选ANOVA
Models中的Multivarite...项,弹出Multivarite ANOVA
对话框(图6.8)。首先指定供分析用的变量x1、x2、x3,故在对话框左侧的变量列表中选变量x1、x2、x3,点击Ø钮使之进入Dependent
Variable框;然后选变量g(分组变量)点击Ø钮使之进入Factor(s)框中,并点击Define
Range钮,确定g的起始值和终止值。
图6.4 多元方差分析对话框 |
点击Options...钮,弹出Multivarite
ANOVA:Options对话框,选择需要计算的指标。在Factor(s)栏内选变量g,点击Ø钮使之进入Display Means
for框,要求计算平均值指标;在Matriced Within
Cell栏内选Correlation、Covariance、SSCP项,要求计算单元内的相关矩阵、方差协方差矩阵和离均差平方和交叉乘积矩阵;在Error
Matrices栏内也选上述三项,要求计算误差的相关矩阵、方差协方差矩阵和离均差平方和交叉乘积矩阵;在Diagnostics栏内选Homogeneity
test项,要求作变量的方差齐性检验。之后点击Continue钮返回Multivarite
ANOVA对话框,最后点击OK钮即可。
6.3.2.3 结果解释
在结果输出窗口中将看到如下分析结果:
系统首先显示共90个观察值进入统计分析,因分组变量g为三个地区,故分析的单元数为3。然后输出3个应变量(x1、x2、x3)的方差齐性检验结果,分别输出了Cochran
C检验值及其显著性水平P值、Bartlett-Box F检验值及其显著性水平P值。其中
身高:C = 0.39825,P = 0.540;F =
1.01272,P = 0.363;
体重:C = 0.43787,P = 0.227;F =
4.48624, P = 0.011;
胸围:C = 0.47239, P = 0.089;F =
2.06585, P = 0.127;
可见3项指标的方差基本整齐(P值均大于0.05)。
90 cases
accepted.
0 cases rejected because of out-of-range factor
values.
0 cases rejected because of missing
data.
3 non-empty 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.)
Bartlett-Box F(2,17030) = 1.01272, P = .363 Variable ..
X2
Cochrans C(29,3) =
.43787,
P = .227
(approx.)
Bartlett-Box F(2,17030) = 4.48624, P = .011 Variable ..
X3
Cochrans C(29,3) =
.47239,
P = .089
(approx.)
Bartlett-Box F(2,17030) = 2.06585, P = .127
|
Cochran C检验和Bartlett-Box
F检验对考查协方差矩阵的相等性比较方便,但还不够。于是系统接着分别输出了三类地区(即各个单元)各生长发育指标的离均差平方和交叉乘积矩阵和方差协方差矩阵。之后作Box
M检验,Box M检验提供矩阵一致性的多元测试,本例Boxs M = 36.93910,在基于方差分析的显著性检验中F =
2.92393;在基于χ2的显著性检验中χ2 = 35.09922, 两者P <
0.001,故认为矩阵一致性不佳。
Cell Number ..
1 Sum of Squares and Cross-Products
matrix
X1
X2
X3 X1
861.187 X2 380.137
230.519 X3
215.937
156.559
314.859
Variance-Covariance
matrix
X1
X2
X3 X1
29.696 X2
13.108
7.949 X3
7.446
5.399
10.857
Cell Number .. 1
(Cont.) Correlation matrix with Standard
Deviations on Diagonal
X1
X2
X3 X1
5.449 X2
.853
2.819 X3
.415
.581
3.295
Determinant of Covariance matrix of
dependent variables =
444.98354 LOG(Determinant) =
6.09804
Cell Number ..
2 Sum of Squares and Cross-Products
matrix
X1
X2
X3 X1
565.368 X2
147.222
78.910 X3
139.430
79.337
147.967
Variance-Covariance
matrix
X1
X2
X3 X1
19.495 X2
5.077
2.721 X3
4.808
2.736
5.102
Correlation matrix with Standard
Deviations on Diagonal
X1
X2
X3 X1
4.415 X2
.697
1.650 X3
.482
.734
2.259
Determinant of Covariance matrix of
dependent variables =
63.90640 LOG(Determinant) =
4.15742
Cell Number ..
3 Sum of Squares and Cross-Products
matrix
X1
X2
X3 X1
944.128 X2
307.722
217.030 X3
261.130
186.252
203.702
Variance-Covariance
matrix
X1
X2
X3 X1
32.556 X2
10.611
7.484 X3
9.004
6.422
7.024
Correlation matrix with Standard
Deviations on Diagonal
X1
X2
X3 X1
5.706 X2
.680
2.736 X3
.595
.886
2.650
Determinant of Covariance matrix of
dependent variables =
198.13507 LOG(Determinant) =
5.28895
Pooled within-cells
Variance-Covariance matrix
X1
X2
X3 X1
27.249 X2
9.599
6.051 X3
7.086 4.852
7.661
Determinant of pooled Covariance
matrix of dependent vars. =
272.06906 LOG(Determinant) =
5.60606
Multivariate test for Homogeneity
of Dispersion matrices
Boxs M =
36.93910 F WITH (12,36680) DF =
2.92393, P =
.000 (Approx.) Chi-Square with 12 DF =
35.09922, P =
.000 (Approx.)
|
下面系统输出将三类地区看成一个大样本时的离均差平方和交叉乘积矩阵。如X1、X2和X3的离均差平方和分别为662.884、121.562和114.902。在此基础上,进行多元差异的检验。通常有四种方法:
1、Pillai轨迹:V =
2、Wilks λ值:W =
3、Hotelling轨迹:T =
4、Roy最大根:R =
式中λmax为最大特征值,
λi为第i个特征值,s为非零特征值个数。根据这些值变换的F检验均有显著性(P<0.001),说明三类地区各生长发育指标之间的差别有高度显著性。
这一计算结果对上述三项生长发育指标进行了单因素的方差分析,可见:
X1: SS = 662.88356, F =
12.16335
X2: SS = 121.56200, F =
10.04439
X3: SS = 114.90200, F =
7.49893
差别均有显著性,说明三项生长发育指标各地区间的差别均有显著性。
Combined Observed Means
for G Variable ..
X1
G
1
WGT.
126.46667
UNWGT.
126.46667
2
WGT.
120.52000
UNWGT.
120.52000
3
WGT.
120.92000
UNWGT.
120.92000 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Variable ..
X2
G
1
WGT.
23.50667
UNWGT.
23.50667
2
WGT.
20.69667
UNWGT.
20.69667
3
WGT.
22.49667
UNWGT.
22.49667 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Variable ..
X3
G
1
WGT.
60.00667
UNWGT.
60.00667
2
WGT.
57.86667
UNWGT.
57.86667
3
WGT.
57.41667
UNWGT.
57.41667 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - WITHIN+RESIDUAL Correlations with
Std. Devs. on Diagonal
X1
X2
X3 X1
5.220 X2
.747
2.460 X3
.490
.713
2.768 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Statistics for WITHIN+RESIDUAL
correlations Log(Determinant) = .00000 Bartlett test of sphericity = . with 3 D.
F. Significance = . F(max) criterion = 4.50308 with (3,87) D.
F.
WITHIN+RESIDUAL Variances and
Covariances
X1
X2
X3 X1
27.249 X2
9.599
6.051 X3
7.086
4.852
7.661 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - WITHIN+RESIDUAL Sum-of-Squares and
Cross-Products
X1
X2
X3 X1
2370.683 X2
835.081
526.458 X3
616.497
422.147
666.527 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - EFFECT .. G Adjusted Hypothesis Sum-of-Squares
and Cross-Products
X1
X2
X3 X1
662.884 X2
230.323
121.562 X3
269.117
78.193
114.902 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Multivariate Tests of Significance
(S = 2, M = 0, N = 41 1/2) Test Name
Value
Approx.F
Hypoth. DF Error
DF Sig. of
F Pillais
.51227
9.87080
6.00
172.00
.000 Hotellings
.70427
9.85978
6.00 168.00
.000 Wilks
.55014
9.86643
6.00
170.00
.000 Roys
.31265 Note.. F statistic for WILKS'
Lambda is exact. - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - EFFECT .. G
(Cont.) Univariate F-tests with (2,87) D.
F. Variable Hypoth. SS Error SS Hypoth. MS Error MS F Sig.
of F X1
662.88356 2370.68267
331.44178
27.24923
12.16335
.000 X2
121.56200
526.45800
60.78100
6.05124
10.04439
.000 X3
114.90200
666.52700
57.45100
7.66123
7.49893
.001
|
之后按单元输出各项指标的观察值均数(Obs.Mean)、调整均数(Adj.Mean)、估计均数(Est.Mean)、粗误差(Raw
Resid)、标准化误差(Std.Resid)以及不分地区的总均数(Comined Adjusted Means for
G)。
Adjusted and Estimated
Means Variable ..
X1 CELL
Obs. Mean Adj.
Mean Est. Mean Raw Resid. Std.
Resid. 1
126.467
126.467
126.467
.000
.000 2
120.520
120.520
120.520
.000
.000 3
120.920
120.920
120.920
.000
.000 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Adjusted and Estimated
Means (Cont.) Variable ..
X2 CELL
Obs. Mean Adj.
Mean Est. Mean Raw Resid. Std.
Resid. 1
23.507
23.507
23.507
.000
.000 2
20.697
20.697
20.697
.000
.000 3
22.497
22.497
22.497
.000
.000 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Adjusted and Estimated Means
(Cont.) Variable ..
X3 CELL
Obs. Mean Adj.
Mean Est. Mean Raw Resid. Std.
Resid. 1
60.007
60.007
60.007
.000
.000 2
57.867
57.867
57.867
.000
.000 3
57.417
57.417
57.417
.000
.000 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Combined Adjusted Means
for G Variable ..
X1
G
1
UNWGT.
126.46667
2
UNWGT.
120.52000
3
UNWGT.
120.92000 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Variable ..
X2
G
1
UNWGT.
23.50667
2
UNWGT.
20.69667
3
UNWGT.
22.49667 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Variable ..
X3
G
1
UNWGT.
60.00667
2
UNWGT.
57.86667
3
UNWGT.
57.41667
|
最后,系统输出各变量的离差参数。用户可据此计算预测值,
预测值 Y = 总均数 + 该变量离差参数 +
变量间交互效应的离差参数
如本例因无变量间交互效应的离差参数,故甲地区8岁男生的身高预测值为 Y = 126.46667 + (-1.71555551)=
124.7511145。
上式中126.46667可从系统输出的Combined
Adjusted Means for G一栏中得到,离差参数-1.71555551 = 0 - 3.83111111 -
(-2.1155556),这是因为离差参数的合计总为0的缘故。余同,在此不作赘述。
Estimates for
X1 --- Individual univariate .9500
confidence intervals G Parameter Coeff. Std. Err. t-Value Sig. t Lower -95% CL- Upper
2
3.83111111
.77816
4.92327
.00000
2.28443
5.37780
3
-2.1155556
.77816
-2.71865
.00791
-3.66224
-.56887 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Estimates for
X2 --- Individual univariate .9500
confidence intervals G Parameter Coeff. Std. Err. t-Value Sig. t Lower -95% CL- Upper
2
1.27333333
.36670
3.47237
.00081
.54447
2.00220
3
-1.5366667
.36670
-4.19048
.00007
-2.26553
-.80780 - - - - - - - - - - - - - - - - - -
- - - - - - - - - - - - - - - - - - - Estimates for
X3 --- Individual univariate .9500
confidence intervals G Parameter Coeff. Std. Err. t-Value Sig. t Lower -95% CL- Upper
2
1.57666667
.41261
3.82117
.00025
.75655
2.39678
3
-.56333333
.41261
-1.36528
.17568
-1.38345
.25678
|