请教xttest0命令输出结果解读

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

Test:   Var(u) = 0
                             chibar2(01) =  1242.97
                          Prob > chibar2 =   0.0000

已知P显著应采用随机效应模型，我想知道输出结果中的chibar2(01) 是代表？ 是chibar2是卡方？(01)是自由度吗？

急 在线等 谢谢各位了！

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏1 回帖

关键词：xttest ttest 输出结果 test Est 自由度 在线 模型

这么多人看了 木有一个人知道吗

etetete 发表于 2013-6-29 15:36
这么多人看了 木有一个人知道吗

好像是的

qifei1988 发表于 2013-6-29 20:31
好像是的

可是像hausman检验里的 卡方是chi2啊 英文不是chibar2.。。。。

What is chibar2?

    The likelihood-ratio (LR) test that is displayed is testing on the boundary of the parameter
    space.  You are probably testing whether an estimated variance component (something that is
    always greater than zero) is different from zero by using an LR test.

    Suppose for now that the two models being compared differ only with respect to the variance
    component in question, in which case the test statistic will be displayed as "chibar(01)".  In
    such cases, the limiting distribution of the maximum-likelihood estimate of the parameter in
    question is a normal distribution that is halved, or chopped off at the boundary -- zero here.
    The distribution of the LR test statistic is therefore not the usual chi-squared with 1 degree
    of freedom but is instead a 50:50 mixture of a chi-squared with no degrees of freedom (that is,
    a point mass at zero) and a chi-squared with 1 degree of freedom.

    The p-value of the LR test takes this into account and will be set to 1 if it is determined
    that your estimate is close enough to zero to be, in effect, zero for purposes of significance.
    Otherwise, the p-value displayed is set to one-half of the probability that a chi-squared with
    1 degree of freedom is greater than the calculated LR test statistic.

    Sometimes you are testing whether a variance component is zero in addition to testing whether k
    other parameters (not affected by boundary conditions) are zero.  Such situations often arise
    when comparing mixed-effects models, such as those fit by xtmixed.  For such tests, the
    distribution of the likelihood-ratio test statistic is a 50:50 mixture of chi-squared
    distributions with k and k+1 degrees of freedom, shown on the output as "chibar(4_5)", for
    example.  As for chibar(01), significance levels are adjusted accordingly.

    Finally, if you are testing more than one boundary-affected parameter, the theory is much more
    complex and usually intractable.  When this occurs, Stata will either display significance
    levels that are conservative and marked as such or will not display an LR test at all.

至于 检验的涵义，不是看stata，
你去看计量经济学的书
greene的书上专门有介绍re中检验的内容，明白理论就知道检验值的涵义了

你需要先懂理论，在谈软件的检验。

有高手路过么