Dear all,
I can't find out how to store the estimated correlation after xtmixed.
I run:
xtmixed y cons1 cons4 if (Phase==1 & (cons1==1 |cons4==1 )) , nocons || Pat_Nr: cons1 cons4 , nocons reml cov(unstructured) ///
|| Woche: ,nocons res(unstructured,t(cons1) )
Obtaining starting values by EM:
Performing gradient-based optimization:
Iteration 0: log restricted-likelihood = 58.969166
Iteration 1: log restricted-likelihood = 62.175238
Iteration 2: log restricted-likelihood = 62.332964
Iteration 3: log restricted-likelihood = 62.339482
Iteration 4: log restricted-likelihood = 62.339706
Iteration 5: log restricted-likelihood = 62.339707
Computing standard errors:
Mixed-effects REML regression Number of obs = 66
-------------------------------------------------------------
| No. of Observations per Group
Group Variable | Groups Minimum Average Maximum
----------------+--------------------------------------------
Pat_Nr | 11 6 6.0 6
Woche | 33 2 2.0 2
-------------------------------------------------------------
Wald chi2(2) = 31.62
Log restricted-likelihood = 62.339707 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
y | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
cons1 | .0954103 .0193423 4.93 0.000 .0575001 .1333204
cons4 | .0414372 .0120296 3.44 0.001 .0178596 .0650148
------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects Parameters | Estimate Std. Err. [95% Conf. Interval]
-----------------------------+------------------------------------------------
Pat_Nr: Unstructured |
sd(cons1) | .0124668 .0832817 2.57e-08 6053.555
sd(cons4) | .0017214 .0020688 .0001632 .0181509
corr(cons1,cons4) | .999999 .0063395 -1 1
-----------------------------+------------------------------------------------
Woche: (empty) |
-----------------------------+------------------------------------------------
Residual: Unstructured |
sd(e0) | .0690406 .0086346 .0540317 .0882187
sd(e1) | .1089946 .0161692 .0814949 .1457739
corr(e0,e1) | .1526103 .1737057 -.1923375 .4639924
------------------------------------------------------------------------------
LR test vs. linear model: chi2(5) = 7.39 Prob > chi2 = 0.1935
Note: LR test is conservative and provided only for reference.
.
It's clear to me how to store corr(cons1,con4)=0.99999 (local corr=tanh(_b[atr1_1_1_2:_cons])).
But I can't find a solution to store corr(e0,e1), namely 0.1526. Where do I find this?
Regards
Kirstin
Related Posts with Saving residuals after xtmixed
Checking cointegration among all possible pairs in a datasetDear All., I have the following dataset referring to daily observations for a bunch of US stocks: …
Marginal structural models and Multivariate regressionIf we get similar estimates of multivariate regression compared with marginal structural models (MSM…
Number of ticks and format of labels on bar chartI am making a bar chart and am using the "over" option so that I can plot multiple bars for a series…
Determining whether to include a squared term and maximizationSay I have coffee = beta0 + beta1educ + beta2imptax + beta3ln(income) + beta4age + beta5agesq + u Wh…
r(mean) , panel subgroupDear Statalisters, I saw this old forum discussion , my question is about understanding the code : …
Subscribe to:
Post Comments (Atom)
0 Response to Saving residuals after xtmixed
Post a Comment