老师同学好,我在做exploratory factor analysis的时候,用stata默认的factor命令,为什么旋转之后出现出现前四个因子解释了超过100%的方差?我运行的命令是:
factor Q21_1-Q24_4
rotate
这是旋转后的表:
| . rotate |
| Factor analysis/correlation Number of obs = 812 |
| Method: principal factors Retained factors = 10 |
| Rotation: orthogonal varimax (Kaiser off) Number of params = 175 |
| -------------------------------------------------------------------------- |
| Factor Variance Difference Proportion Cumulative |
| -------------+------------------------------------------------------------ |
| Factor1 5.99990 3.44700 0.4649 0.4649 |
| Factor2 2.55290 0.26643 0.1978 0.6627 |
| Factor3 2.28647 0.06143 0.1772 0.8399 |
| Factor4 2.22505 1.88952 0.1724 1.0123 |
| Factor5 0.33553 0.16772 0.0260 1.0383 |
| Factor6 0.16781 0.00215 0.0130 1.0513 |
| Factor7 0.16566 0.06032 0.0128 1.0642 |
| Factor8 0.10534 0.00337 0.0082 1.0723 |
| Factor9 0.10197 0.08302 0.0079 1.0802 |
| Factor10 0.01895 . 0.0015 1.0817 |
| -------------------------------------------------------------------------- |
| LR test: independent vs. saturated: chi2(231) = 1.3e+04 Prob>chi2 = 0.0000 |
为什么最后一列累计百分比超过100%。
顺便说,从因子分析检验和factor loading看,倒是挺符合假设的,非常感谢。