利用spss进行异方差检验的方法，望借鉴

馨冉回忆

1. 用图表检验Analyze -> regression -> Linear-> Plots
$ U2 r* Z7 D* b4 NScatter plot of the standardised residuals on the standardised predicted values (ZRESID as  the Y variable, and ZPRED as the X variable
7 r) s4 c; `' e; _3 B7 n
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。
1 h! I5 i4 d7 x$ }% i% ^; L2. 用统计检验
+ b- l; d% {2 D. X- ZHeteroscedasticity——Testing and Correcting in SPSS.pdf Gwilym Pryce March 2002.doc (172.5 KB, 下载次数: 4)

2013-10-13 16:09 上传

点击文件名下载附件
下载积分: 体力 -2 点

! v6 N0 |! d) |4 K3 m* @7 j
- d4 y" F" g' W2 p
Levene’s Test
8 I$ q, Z6 G1 q' M3 ^$ EGoldfeld-Quandt Test
Breusch-Pagan Test
White‘s Test （比较常用来检验异方差）
& h  z( b8 N8 e& u2 j3 R' Z/ K

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)

6 ^# C& Q1 j+ d6 t& o

- u( @' e" _" w1 e) {) b9 g

5 d( I  u3 Z: o  X

White’s Test

·        Calculate n*R2                            à              R2  = 0.037, n=2016           à    Thus, n*R2 = .037*2016 = 74.6.

( Z8 ~$ R  a, W$ _9 v; h/ q

·        Compare this value with c2 (n), i.e.with  c2 (2016)  
1 D7 s2 w0 C! a+ ^+ A! ~2 |(c2  is the symbol for theChi-Square distribution)

$ {' M4 c& X8 x9 |- z  l

c2 (2016) = 124obtained from c2 table.            (For 955 confidence)              As n*R2  < c2 ,heteroskedasticity can not be confirmed.

8 u; C  k6 ~; O! M6 z请参考：regression_explained_SPSS regression_explained_SPSS.doc (368 KB, 下载次数: 0)

2013-10-13 16:12 上传

点击文件名下载附件
下载积分: 体力 -2 点