the code is as follows:
Code:
melogit outcome exposure covariate1 covariate2 covariate3 ||household:||family:, vce(cluster household)
melogit c_use age urban || district: age what is the 2nd age accounting for and is that something I should include?
Also, integration points, when should you change the amount of integration points you use and how do you use them, I know the default is 7 but I am not sure if I need to change them or how to interpret the output if I do.
Also, when using melogit are the point estimates accounted for clustering using fixed effects then the SE using robust at the household level? is it possible to change this or to use random effects, if so how do I do that?
The very last two parts what do the var(_cons) mean in these circumstances
Code:
Fitting fixed-effects model: Iteration 0: log likelihood = -6079.9391 Iteration 1: log likelihood = -5550.3319 Iteration 2: log likelihood = -5549.0637 Iteration 3: log likelihood = -5549.063 Iteration 4: log likelihood = -5549.063 Refining starting values: Grid node 0: log likelihood = -5824.9555 Fitting full model: Iteration 0: log pseudolikelihood = -5824.9555 (not concave) Iteration 1: log pseudolikelihood = -5711.0415 (not concave) Iteration 2: log pseudolikelihood = -5555.8322 Iteration 3: log pseudolikelihood = -5522.1834 Iteration 4: log pseudolikelihood = -5521.3746 Iteration 5: log pseudolikelihood = -5521.3738 Iteration 6: log pseudolikelihood = -5521.3738 Mixed-effects logistic regression Number of obs = 30,000 ------------------------------------------------------------- | No. of Observations per Group Group Variable | Groups Minimum Average Maximum ----------------+-------------------------------------------- household | 7,000 1 3.8 25 family | 20,000 1 1.3 4 ------------------------------------------------------------- Integration method: mvaghermite Integration pts. = 7 Wald chi2(8) = 59.79 Log pseudolikelihood = -5521.3738 Prob > chi2 = 0.0000 (Std. Err. adjusted for 7,000clusters in household) ------------------------------------------------------------------------------------ | Robust outcome| Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------------+---------------------------------------------------------------- exposure| 2 | .4139325 .074874 4.19 0.000 .1671822 .4606829 3 | .956007 .4214943 2.22 0.027 .1078934 1.760121 4 | .70234631 .1889264 3.74 0.000 .3370741 1.077652 | covariate1| -.0023418 .0591824 -0.10 0.923 -.1216872 .1103036 covariate2 | .2345331 .0712209 3.69 0.000 .1233427 .4025234 covariate3| .013492 .0372542 0.40 0.686 -.0579577 .0880762 covariate4| .0163457 .0238631 0.70 0.484 -.0300612 .0634805 covariate5| .0343457 .0965659 0.36 0.722 -.154881 .2236504 _cons | -3.783458 .1728209 -21.92 0.000 -4.12689 -3.449445 -------------------+---------------------------------------------------------------- household | var(_cons)| .4873455 .1111736 .3113988 .7623431 -------------------+---------------------------------------------------------------- household>family | var(_cons)| .7645611 .2636417 .3902342 1.5042231 ------------------------------------------------------------------------------------
0 Response to Using and interpreting multi level mixed effects logictic regression with nested clustering
Post a Comment