阅读使人充实，会谈使人敏捷，写作使人精确。培根
计量经济学上机模型分析方法总结
一、随机误差项的异方差问题的检验与修正
模型一：Depe
de
tVariableLOGYMethodLeastSquaresDate072912Time0903Sample131I
cludedobservatio
s31
Variable
Coefficie
tStdErrortStatisticProb
CLOGX1LOGX2
160252803254160507078
086097801037690048599
186128831359551043385
007320004000000
RsquaredAdjustedRsquaredSEofregressio
SumsquaredresidLoglikelihoodDurbi
Watso
stat
079650607819710170267081174712472491964720
Mea
depe
de
tvarSDdepe
de
tvarAkaikei
focriterio
Schwarzcriterio
FstatisticProbFstatistic
744870403646480611128047235554798060000000
模型二：Depe
de
tVariableLOGYMethodLeastSquaresDate072912Time0919Sample112I
cludedobservatio
s12
（一）异方差的检验1、GQ检验法
Variable
Coefficie
tStdErrortStatisticProb
CLOGX1LOGX2
374462603443690168904
119111300829990118844
314380441490771421228
001190002501890
RsquaredAdjustedRsquaredSEofregressio
SumsquaredresidLoglikelihoodDurbi
Watso
stat
066906505955240084955006495714286381810822
Mea
depe
de
tvarSDdepe
de
tvarAkaikei
focriterio
Schwarzcriterio
FstatisticProbFstatistic
723916101335811881064175983790978340006900
学问是异常珍贵的东西，从任何源泉吸收都不可耻。阿卜日法拉兹
f阅读使人充实，会谈使人敏捷，写作使人精确。培根
模型三：Depe
de
tVariableLOGYMethodLeastSquaresDate072912Time0920Sample2031I
cludedobservatio
s12
Variable
Coefficie
tStdErrortStatisticProb
CLOGX1LOGX2
035338102108980856522
160746101582200108601
021983813329427886856
083090215300000
RsquaredAdjustedRsquaredSEofregressio
SumsquaredresidLoglikelihoodDurbi
Watso
stat
087840208513810150490020382474251632123203
Mea
depe
de
tvarSDdepe
de
tvarAkaikei
focriterio
Schwarzcriterio
FstatisticProbFstatistic
776985103903630737527061630132507320000076
进行模型二和模型三两次回归，目的仅是得到出去中间7个样本点以后前后各12个样本点的残差平方和RSS1和RSS2，然后用较大的RSS除以较小的RSS即可求出
F统计量值进行显著性检验。2、怀特检验法（White）
模型一的怀特残差检验结果：
WhiteHeteroskedasticityTest
FstatisticObsRsquared
4920995Probability1335705Probability
00043390009657
TestEquatio
Depe
de
tVariableRESID2MethodLeastSquaresDate052913Time0904Sample131I
cludedobservatio
s31
Variable
Coefficie
tStdErrortStatisticProb
CLOGX1LOGX12
398213705792890041839
288285109160690066866
138r