1. 用图表检验Analyze -> regression -> Linear-> Plots
" ^- L% @/ H Y% W3 h0 LScatter plot of the standardised residuals on the standardised predicted values (ZRESID as the Y variable, and ZPRED as the X variable
; u* m Q7 B `0 w; ?# j% @/ H: j/ {7 L4 V* ~% j
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。
/ V& e) m( G q5 l3 {) l, s2. 用统计检验
( n8 p6 w8 {# `. g1 j% |& e2 E- o# a" B. eHeteroscedasticity——Testing and Correcting in SPSS.pdf
Gwilym Pryce March 2002.doc
(172.5 KB, 下载次数: 4)
- N' }8 D; p- D- o6 ^; [6 |" t* z: h- [! m3 R# E4 o
4 M, G5 h9 b) x7 t4 x
Levene’s Test" w* i8 A& `' m5 F7 i
Goldfeld-Quandt Test- ^# n" L6 t( L+ r2 b5 U
Breusch-Pagan Test
$ n5 H0 c5 H% S ` j' DWhite‘s Test (比较常用来检验异方差)9 a) @' z: D# \$ O7 R# H1 B* O
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)
: I6 }* p- g8 I! Q0 b2 h; Y& l. [
2 Z4 f; U( f3 N6 W! z) s7 s4 v% E
% e7 H3 g, v+ q" k5 t, G* T2 @0 f3 ~4 o1 f
White’s Test · Calculate n*R2 à R2 = 0.037, n=2016 à Thus, n*R2 = .037*2016 = 74.6.
+ N4 l; Y: Z- J· Compare this value with c2 (n), i.e.with c2 (2016) 7 P6 k( r0 T, V: \6 C
(c2 is the symbol for theChi-Square distribution) + `/ W3 w5 G/ B2 O+ o. y
c2 (2016) = 124obtained from c2 table. (For 955 confidence) As n*R2 < c2 ,heteroskedasticity can not be confirmed.
- k- n$ D8 i+ ^: ]
( R0 r) }" C7 ~0 @* @( `请参考:regression_explained_SPSS
regression_explained_SPSS.doc
(368 KB, 下载次数: 0)
; ~) L; q: g6 a7 S" y
|