Hi to everybody, how to report the results of a mixed effects model?Can I explain better how to write if there is a diagnosis effect, a time effect and an interaction effect? and therefore how to write the statistics? Thank you very much in advance everyone.
Below my output and my post-hoc
xtmixed RITIRO_sociale Diagnosi_numbis##timepoint || progressivo_soggetto:, var reml
Performing EM optimization:
Performing gradient-based optimization:
Iteration 0: Log restricted-likelihood = -1015.2903
Iteration 1: Log restricted-likelihood = -1015.2802
Iteration 2: Log restricted-likelihood = -1015.2802
Computing standard errors:
Mixed-effects REML regression Number of obs = 342
Group variable: progressivo_~o Number of groups = 190
Obs per group:
min = 1
avg = 1.8
max = 2
Wald chi2(7) = 49.03
Log restricted-likelihood = -1015.2802 Prob > chi2 = 0.00000
-------------------------------------------------------------------------------------------
RITIRO_sociale | Coefficient Std. err. z P>|z| [95% conf. interval]
--------------------------+----------------------------------------------------------------
Diagnosi_numbis |
2 | 1.146703 1.105662 1.04 0.300 -1.020354 3.313759
3 | 3.41232 1.229787 2.77 0.006 1.001982 5.822659
4 | 4.544643 1.593348 2.85 0.004 1.421738 7.667548
|
24.timepoint | .7807385 1.029916 0.76 0.448 -1.237861 2.799338
|
Diagnosi_numbis#timepoint |
2 24 | -.1016103 1.188776 -0.09 0.932 -2.431569 2.228349
3 24 | -.2780768 1.340183 -0.21 0.836 -2.904788 2.348634
4 24 | 4.972854 1.780246 2.79 0.005 1.483636 8.462072
|
_cons | 50.14286 .960825 52.19 0.000 48.25967 52.02604
-------------------------------------------------------------------------------------------
------------------------------------------------------------------------------
Random-effects parameters | Estimate Std. err. [95% conf. interval]
-----------------------------+------------------------------------------------
progressiv~o: Identity |
var(_cons) | 11.38482 2.378743 7.559202 17.14652
-----------------------------+------------------------------------------------
var(Residual) | 14.46436 1.736417 11.43179 18.3014
------------------------------------------------------------------------------
LR test vs. linear model: chibar2(01) = 26.97 Prob >= chibar2 = 0.0000
estat ic
Akaike's information criterion and Bayesian information criterion
-----------------------------------------------------------------------------
Model | N ll(null) ll(model) df AIC BIC
-------------+---------------------------------------------------------------
. | 342 . -1015.28 10 2050.56 2088.908
-----------------------------------------------------------------------------
Note: BIC uses N = number of observations. See [R] IC note.
contrast timepoint@Diagnosi, effect
Contrasts of marginal linear predictions
Margins: asbalanced
-------------------------------------------------------------
| df chi2 P>chi2
--------------------------+----------------------------------
RITIRO_sociale |
timepoint@Diagnosi_numbis |
1 | 1 0.57 0.4484
2 | 1 1.31 0.2527
3 | 1 0.34 0.5578
4 | 1 15.70 0.0001
Joint | 4 17.93 0.0013
-------------------------------------------------------------
-------------------------------------------------------------------------------------------
| Contrast Std. err. z P>|z| [95% conf. interval]
--------------------------+----------------------------------------------------------------
RITIRO_sociale |
timepoint@Diagnosi_numbis |
(24 vs base) 1 | .7807385 1.029916 0.76 0.448 -1.237861 2.799338
(24 vs base) 2 | .6791281 .5936846 1.14 0.253 -.4844723 1.842729
(24 vs base) 3 | .5026616 .8575333 0.59 0.558 -1.178073 2.183396
(24 vs base) 4 | 5.753593 1.452084 3.96 0.000 2.90756 8.599625
-------------------------------------------------------------------------------------------
. margins timepoint, at(Diagnosi_numbis=1) pwcompare(effects) mcompare(bonferroni)
Pairwise comparisons of adjusted predictions Number of obs = 342
Expression: Linear prediction, fixed portion, predict()
At: Diagnosi_numbis = 1
note: option bonferroni ignored since there is only one comparison
------------------------------------------------------------------------------
| Delta-method Unadjusted Unadjusted
| Contrast std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
timepoint |
24 vs 18 | .7807385 1.029916 0.76 0.448 -1.237861 2.799338
------------------------------------------------------------------------------
. margins timepoint, at(Diagnosi_numbis=2) pwcompare(effects) mcompare(bonferroni)
Pairwise comparisons of adjusted predictions Number of obs = 342
Expression: Linear prediction, fixed portion, predict()
At: Diagnosi_numbis = 2
note: option bonferroni ignored since there is only one comparison
------------------------------------------------------------------------------
| Delta-method Unadjusted Unadjusted
| Contrast std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
timepoint |
24 vs 18 | .6791281 .5936846 1.14 0.253 -.4844723 1.842729
------------------------------------------------------------------------------
. margins timepoint, at(Diagnosi_numbis=3) pwcompare(effects) mcompare(bonferroni)
Pairwise comparisons of adjusted predictions Number of obs = 342
Expression: Linear prediction, fixed portion, predict()
At: Diagnosi_numbis = 3
note: option bonferroni ignored since there is only one comparison
------------------------------------------------------------------------------
| Delta-method Unadjusted Unadjusted
| Contrast std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
timepoint |
24 vs 18 | .5026616 .8575333 0.59 0.558 -1.178073 2.183396
------------------------------------------------------------------------------
. margins timepoint, at(Diagnosi_numbis=4) pwcompare(effects) mcompare(bonferroni)
Pairwise comparisons of adjusted predictions Number of obs = 342
Expression: Linear prediction, fixed portion, predict()
At: Diagnosi_numbis = 4
note: option bonferroni ignored since there is only one comparison
------------------------------------------------------------------------------
| Delta-method Unadjusted Unadjusted
| Contrast std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
timepoint |
24 vs 18 | 5.753593 1.452084 3.96 0.000 2.90756 8.599625
------------------------------------------------------------------------------
.
contrast Diagnosi_numbis@timepoint, effect mcompare(bonferroni)
Contrasts of marginal linear predictions
Margins: asbalanced
-------------------------------------------------------------
| df chi2 P>chi2
--------------------------+----------------------------------
RITIRO_sociale |
Diagnosi_numbis@timepoint |
18 | 3 14.00 0.0029
24 | 3 39.01 0.0000
Joint | 6 43.26 0.0000
-------------------------------------------------------------
Note: Bonferroni-adjusted p-values are reported for tests on
individual contrasts only.
----------------------------------------
| Number of
| comparisons
--------------------------+-------------
RITIRO_sociale |
Diagnosi_numbis@timepoint | 6
----------------------------------------
-------------------------------------------------------------------------------------------
| Bonferroni Bonferroni
| Contrast Std. err. z P>|z| [95% conf. interval]
--------------------------+----------------------------------------------------------------
RITIRO_sociale |
Diagnosi_numbis@timepoint |
(2 vs base) 18 | 1.146703 1.105662 1.04 1.000 -1.770317 4.063722
(2 vs base) 24 | 1.045092 1.110614 0.94 1.000 -1.884993 3.975178
(3 vs base) 18 | 3.41232 1.229787 2.77 0.033 .1678253 6.656815
(3 vs base) 24 | 3.134243 1.250465 2.51 0.073 -.1648054 6.433292
(4 vs base) 18 | 4.544643 1.593348 2.85 0.026 .3409807 8.748305
(4 vs base) 24 | 9.517497 1.693169 5.62 0.000 5.050481 13.98451
-------------------------------------------------------------------------------------------
.
.margins Diagnosi_numbis, at(timepoint=18) pwcompare(effects) mcompare(bonferroni)
Pairwise comparisons of adjusted predictions Number of obs = 342
Expression: Linear prediction, fixed portion, predict()
At: timepoint = 18
------------------------------
| Number of
| comparisons
----------------+-------------
Diagnosi_numbis | 6
------------------------------
---------------------------------------------------------------------------------
| Delta-method Bonferroni Bonferroni
| Contrast std. err. z P>|z| [95% conf. interval]
----------------+----------------------------------------------------------------
Diagnosi_numbis |
2 vs 1 | 1.146703 1.105662 1.04 1.000 -1.770317 4.063722
3 vs 1 | 3.41232 1.229787 2.77 0.033 .1678253 6.656815
4 vs 1 | 4.544643 1.593348 2.85 0.026 .3409807 8.748305
3 vs 2 | 2.265618 .9425999 2.40 0.097 -.2212033 4.752439
4 vs 2 | 3.39794 1.38379 2.46 0.084 -.252855 7.048736
4 vs 3 | 1.132323 1.484845 0.76 1.000 -2.785081 5.049726
---------------------------------------------------------------------------------
. margins Diagnosi_numbis, at(timepoint=24) pwcompare(effects) mcompare(bonferroni)
Pairwise comparisons of adjusted predictions Number of obs = 342
Expression: Linear prediction, fixed portion, predict()
At: timepoint = 24
------------------------------
| Number of
| comparisons
----------------+-------------
Diagnosi_numbis | 6
------------------------------
---------------------------------------------------------------------------------
| Delta-method Bonferroni Bonferroni
| Contrast std. err. z P>|z| [95% conf. interval]
----------------+----------------------------------------------------------------
Diagnosi_numbis |
2 vs 1 | 1.045092 1.110614 0.94 1.000 -1.884993 3.975178
3 vs 1 | 3.134243 1.250465 2.51 0.073 -.1648054 6.433292
4 vs 1 | 9.517497 1.693169 5.62 0.000 5.050481 13.98451
3 vs 2 | 2.089151 .9463817 2.21 0.164 -.4076473 4.58595
4 vs 2 | 8.472405 1.482834 5.71 0.000 4.560307 12.3845
4 vs 3 | 6.383253 1.590282 4.01 0.000 2.187681 10.57883
---------------------------------------------------------------------------------
Comment