(2)影响因素的显著性分析
Dependent Variable: y
Sum of
Source DF Squares Mean Square F Value Pr > F
Model 6 260.6666667 43.4444444 30.08 0.0325
Error 2 2.8888889 1.4444444
Corrected Total 8 263.5555556
Source DF Anova SS Mean Square F Value Pr > F
x1 2 54.8888889 27.4444444 19.00 0.0500
x2 2 37.5555556 18.7777778 13.00 0.0714
x3 2 168.2222222 84.1111111 58.23 0.0169
模型的F值为30.08,p值为0.0325,大于在0.05水平上的F值,由此可知,模型有效,此次的评定的结果是可靠的。F值为19.00,p值为0.0500,可认为在0.05的水平上是显著的,因此辅料A的三个水平对总分的影响有显著的差异;F值为13.00,p值为0.0714,小于在0.05的水平上的F统计量的值,因此辅料B的三个水平对总分的影响无显著的差异;F统计量的值为58.23,p值为0.0169,大于0.01而小于在0.05的水平上的F统计量的值,因此辅料C的三个水平对总分的影响有显著的差异。
由上表,还可以说明各因素对总分的影响大小。比较F值大小:FC > FA > FB。F值越大,影响作用越大,各个因素对总分的影响程度大小的次序为C>A>B。
(3)最佳配方的确定
SAS中“means”指令的输出结果为:
The ANOVA Procedure
Level of --------------y--------------
x1 N Mean/均值 Std Dev/标准误
1 3 79.0000000 7.21110255
2 3 83.6666667 4.04145188
3 3 78.0000000 6.00000000
Level of --------------y--------------
x2 N Mean Std Dev
1 3 77.6666667 8.96288644
2 3 82.6666667 4.50924975
3 3 80.3333333 3.51188458
Level of --------------y--------------
x3 N Mean Std Dev
1 3 77.0000000 3.60555128
2 3 86.3333333 2.08166600
3 3 77.3333333 5.50757055
sas是个工具, 你要学会怎么去用它.而不是学sas.作者: shiyiming 时间: 2005-10-26 19:37 标题: regression I don't think it is necessary to use ANOVA to detect the relationship between variety and the results. you can use regression and regard variety as variable, and then compare the slope. If the slope is not significant to 0, so no relationship. ANOVA just see the difference between variety, can't rank. If you use regression you can take LSD comparison.
I don't know whether I said the right thing!作者: shiyiming 时间: 2005-10-27 09:55 标题: thanks thank you , liudreams!作者: shiyiming 时间: 2005-10-27 19:58 标题: 你的方法是正确的 To tiantian: