楼主问题解决没有啊，最近写论文和楼主用的同一个模型Mlogit,也要检验IIA可是不太会用hausman检验啊。。。。求楼主把命令粘贴看看。。。。。万分感谢！

mlogit jobsearch ymwork ymwo medu birth yfwork yfwo fedu fwork mwork sex edu party health hunyin
estimates store m100
mlogit jobsearch ymwork ymwo medu birth yfwork yfwo fedu fwork mwork sex edu party health hunyin if jobsearch !="guanxi":jobsearchlabel
estimates store m101
hausman m101 m100,alleqs constant

suest m100 m101
estat vce
test [m100_jihua=m101_jihua],cons


mlogit jobsearch ymwork ymwo medu birth yfwork yfwo fedu fwork mwork sex edu party health hunyin
estimates store m102
mlogit jobsearch ymwork ymwo medu birth yfwork yfwo fedu fwork mwork sex edu party health hunyin if jobsearch !="jihua":jobsearchlabel
estimates store m103
hausman m103 m102,alleqs constant

suest m102 m103
estat vce
test [m102_guanxi=m103_guanxi],cons

mlogit jobsearch ymwork ymwo medu birth yfwork yfwo fedu fwork mwork sex edu party health hunyin,base(1)
estimates store m106
mlogit jobsearch ymwork ymwo medu birth yfwork yfwo fedu fwork mwork sex edu party health hunyin if jobsearch !="shichang":jobsearchlabel,base(1)
estimates store m107
hausman m107 m106,alleqs constant

suest m106 m107
estat vce
test [m106_jihua=m107_jihua],cons

多谢，求问出现这个问题：Note: the rank of the differenced variance matrix (6) does not equal the number of coefficients
        being tested (78); be sure this is what you expect, or there may be problems computing the
        test.  Examine the output of your estimators for anything unexpected and possibly consider
        scaling your variables so that the coefficients are on a similar scale.
您可知该如何处理。。。。我大概想到是因为两个要比较的模型的变量个数不对等，因为里面有产业虚拟变量，两个模型中产业虚拟变量的个数不对等。。。。但是产业不是我关注的，其他主要变量之间的差别其实比较小，主要产业虚拟变量导致结果不太好， 出现了：chi2(5) = (b-B)'[(V_b-V_B)^(-1)](b-B)
                          =    -0.00   。。。。

mlogtest, hausman base

mlogtest, hausman base
这个命令经常报错？

如果HAUSMAN检验的卡方值为正值，但是P值为0，该如何解释呢？

您好，mlogit iia检验是如何安装的，我找不到安装文件，能不能教教我怎么安装这个命令

findit sg155