Dear all,
I am tryin to run a multilevel model (individuals nested in countries), while controlling for year-effects (2012-2018). See my code below. The result omits the constant, and I don't understand why nor what this means. I tried including explicit dummies for each year instead, but the results remain the same. Why is this the constant omitted? I do not have missing values for these variables. I also tried changing the base category, without success.
Best,
Harry
I am tryin to run a multilevel model (individuals nested in countries), while controlling for year-effects (2012-2018). See my code below. The result omits the constant, and I don't understand why nor what this means. I tried including explicit dummies for each year instead, but the results remain the same. Why is this the constant omitted? I do not have missing values for these variables. I also tried changing the base category, without success.
Code:
mixed demosatisfaction equalityvalue gini gindifnat gindifeu i.year || country:
Mixed-effects ML regression Number of obs = 171,631
Group variable: country Number of groups = 28
Obs per group:
min = 2,775
avg = 6,129.7
max = 10,099
Wald chi2(10) = 5146.44
Log likelihood = -192192.93 Prob > chi2 = 0.0000
----------------------------------------------------------------------------------
demosatisfaction | Coefficient Std. err. z P>|z| [95% conf. interval]
-----------------+----------------------------------------------------------------
equalityvalue | .0100899 .0045895 2.20 0.028 .0010946 .0190852
gini | .0806512 .0012274 65.71 0.000 .0782454 .0830569
gindifnat | -.0127411 .0101866 -1.25 0.211 -.0327064 .0072242
gindifeu | -.0779339 .0099304 -7.85 0.000 -.0973972 -.0584707
|
year |
2013 | -.0215762 .0067843 -3.18 0.001 -.0348731 -.0082792
2014 | -.0485003 .0085952 -5.64 0.000 -.0653466 -.031654
2015 | -.0392094 .0085768 -4.57 0.000 -.0560195 -.0223993
2016 | .0025607 .0074815 0.34 0.732 -.0121028 .0172243
2017 | .0538484 .0068795 7.83 0.000 .0403648 .067332
2018 | .0823129 .0074179 11.10 0.000 .0677741 .0968518
|
_cons | 0 (omitted)
----------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects parameters | Estimate Std. err. [95% conf. interval]
-----------------------------+------------------------------------------------
country: Identity |
var(_cons) | .0380295 .0101915 .022491 .0643032
-----------------------------+------------------------------------------------
var(Residual) | .5492214 .001875 .5455587 .5529086
------------------------------------------------------------------------------
LR test vs. linear model: chibar2(01) = 10590.29 Prob >= chibar2 = 0.0000
Harry
