1. 用图表检验Analyze -> regression -> Linear-> Plots
; A3 b8 H1 G9 H" m( ~; Q, g+ p4 o- dScatter plot of the standardised residuals on the standardised predicted values (ZRESID as the Y variable, and ZPRED as the X variable8 d( N) I2 g1 C
" j( L0 s8 A% P9 m- W8 `如果图表显示有可能存在异方差,需要用统计检验来进一步检测异方差是否确实存在。) R/ {% G( h1 P' W
2. 用统计检验
& X$ E# J) R7 ?4 P1 A K& dHeteroscedasticity——Testing and Correcting in SPSS.pdf
Gwilym Pryce March 2002.doc
(172.5 KB, 下载次数: 4)
" w0 x3 E# h% }+ y/ c f! g. v% v" v- _; H1 G+ {
0 y% N, ^4 P7 p0 T& b" [
Levene’s Test
+ W1 b- C* n- D+ @8 @Goldfeld-Quandt Test$ F; o b: O3 r V* v* r2 r3 o5 y
Breusch-Pagan Test- @+ ~' F* `9 @, `8 R }- v I7 d
White‘s Test (比较常用来检验异方差)0 N, x+ @3 _3 O2 L& [
Assume you want to run a regression of wage on age, work experience,education, gender, and a dummy for sectorofemployment (whether employed in the public sector). wage = function(age, workexperience, education, gender, sector) or, as your textbook will have it, wage = b1 + b2*age + b3*work experience+ b4*education + b5*gender + b6*sector The White’s test is usually used as a test for heteroskedasticity. In this test, a regression of the squares ofthe residuals is run on the variables suspected of causing theheteroskedasticity, their squares, and cross products. (residuals)2 = b0 + b1 educ + b2 work_ex+ b3 (educ)2 + b4 (work_ex)2 + b5(educ*work_ex)
3 [4 y* r C* n0 `! a' @ e7 ~% o9 T. q! t2 @2 y: |- R
5 A" J5 z3 }! Y. y7 V
3 d0 F/ o5 Z6 t2 F/ T B* c
White’s Test · Calculate n*R2 à R2 = 0.037, n=2016 à Thus, n*R2 = .037*2016 = 74.6.
2 Q a& N. k* M% g9 l; z- G+ R· Compare this value with c2 (n), i.e.with c2 (2016) 6 M( F' u0 u! Q1 S" i4 D/ y/ z: a
(c2 is the symbol for theChi-Square distribution) z% k( [0 m% v Y% E( N6 q# r
c2 (2016) = 124obtained from c2 table. (For 955 confidence) As n*R2 < c2 ,heteroskedasticity can not be confirmed.
8 Q* |. K) ~! ^3 t8 g8 V9 ~7 L! Z$ D' f* c( m3 Q/ p
请参考:regression_explained_SPSS
regression_explained_SPSS.doc
(368 KB, 下载次数: 0)
) N' E. [3 w. i! C( v |