中人网

标题: 请教有关探索性因子分析的问题 [打印本页]

作者: 卖桔者    时间: 2012-9-13 09:29
标题: 请教有关探索性因子分析的问题
请教kenny和各位:
+ ^* H& H  ?) j8 d; `  }      用SPSS软件做探索性因子分析(EFA),采用斜交旋转(Promax)提取因子,在输出的结果中有 pattern matrix 和 structure matrix,应该看哪个结果?谢谢!  U1 B9 B. U  n" O! |6 ]3 I
3 p3 g, w  t2 Q; a4 H- V9 D- P

作者: jkliang    时间: 2012-9-13 17:40
本帖最后由 jkliang 于 2012-9-13 17:42 编辑 + L; t5 K6 x- X& m1 ^

3 R3 }5 i# b! ?) B在某個SPSS問答網頁中尋得以下答案,請參考。
9 a0 @4 b  Q5 j' n' z1 z" F4 F' ]% ?& m3 @
The pattern matrix elements are analogous to the standardized regression coefficients from a, G3 O* ^& C8 u  b  u
multiple regression.  A given element of the pattern matrix indicates the importance of that
1 }- z; t+ D! h2 r6 \variable to the factor with the influence of the other variables partialled out.
4 x; S1 ?& w* O
, ?, x1 H+ T" n+ C3 o0 Z) T- |* V, Y, R( N. y9 L9 ?
The structure matrix elements are simple correlations of the variables with the factors.  They reflect all: q& j6 i& r+ G4 k9 b- C" h
paths from a factor to a variable (this includes factor to factor paths, since factors are correlated), while the
0 ?" n+ R# I9 cpattern coefficients reflect only the direct path from a factor to a variable.
& Y9 Z* ]: ^, J8 x
7 R$ h3 i: ?, P' a3 P3 _- W$ y! b4 E! k

作者: Kenneth    时间: 2012-9-14 17:34
谢谢 jkliang.  Simple correlations of the variables with the factors 就是载荷了。




欢迎光临 中人网 (http://bbs.chinahrd.net/) Powered by Discuz! X2.5