gzw6681467
续表:
题项 成 分
1 2 3 4 5
D13 0.266 0.743 0.166 -0.111 0.199
D10 0.191 0.713 0.354 -0.176 -2.577E-02
D5 0.477 0.641 0.247 -0.]43 0.517
D18 0.454 0.561 0.307 -0.499 0.347
D2 0.222 0.236 0.818 -3.200E-02 0.120
D1 0.269 0.292 0.750 -0.391 0.299
D14 0.147 0.578 0.663 -0.241 3.908E-02
D9 0.255 0.355 0.662 0.256 -0.105
D17 0.496 0.563 0.570 -0.399 0.194
D19 0.254 0.254 0.171 -0.875 0.110
D12 0.272 0.237 0.145 -7.292E-02 0.870
D4 0.508 0.180 0.]70 -0.313 0.789
萃取方法:主成分分析
旋转方法:含Kaiser正态化的Oblimin法。
注:此处的结构矩阵类似于使用最大变异法中的转轴后的成分矩阵,这五个共同因素中,第四个共同因素包含第19题、第五个因素包含第12题与第4题,这二个因素包含的题项太少,无法测出所要测量的构想,建议删除。
以上是<<SPSS统计应用实务----问卷分析与应用统计>>第84页的内容,这是在对题项分析后抽取5个主成分后所得的成分矩阵.她在注释中说"第五因素包含第12题与第4题......",在此不知道为什么说第五因素包含第12题与第4题,而不包括其它题,如第5题,判断的依据是什么?恳请各位大师们指点迷路!!
anita_jiu
I think your question is related to factor loading value which represents the correlation between variables and components in essense. The determination of such a value depends on a number of common elements, such as:
1. correlation coefficients between variables, and
2. sample size
According to the matrix you presented above, the 5th component - both D12 and D4 have the highest factor loading values (.7 may be considered as high). There is another variable here, D5, which also holds a relatively high correlation (.517).
The 4th component - It is clear that D19 (-0.875) is dominating this component although D18 (-.499) appears to have 'somewhat' high correlation with the component as well. However, one can argue that D19 is so high and different from the other variables so that only D18 shall be remained in this component. When a component is formed by a single variable, this implies that the component is unstable or invalid and therefore should be removed from the analysis. I suppose this is one possible explanation to your query.
mhwxl
你可以看一下 吴明隆 的那本书(第2版)第三章 讲的很清楚
hgsean
rotated component matrix is more useful than component matrix that you provided here!
bjnanhai
select the most important component in one. two.....and five .then continue to analyse
gzw6681467
非常感谢各位高手指导,我已在你们的指引下基本掌握了对这个报表的分析.再次深表谢意!!!!
wansuiwansui
zenmcainyouqiana