bootstrap likelihood ratio OR Lo-Mendell-Rubin adjusted likelihood ratio test

[Radhouene DOGGUI]

Dear Statalisters,

Please I am looking for your help to obtain the syntax in order to calculate bootstrap likelihood ratio test (BLRT) and Lo-Mendell-Rubin adjusted likelihood ratio tests for the number of classes in latent class models .

Thanks

[Carlo Lazzaro]

Radhouenne:
as far as your first question is concerned (I'm not familiar with Lo-Mendell-Rubin adjusted likelihood ratio test), the following toy-example may be helpful:

. use "C:\Program Files\Stata16\ado\base\a\auto.dta"

. logit foreign mpg

Iteration 0:   log likelihood =  -45.03321  
Iteration 1:   log likelihood = -39.380959  
Iteration 2:   log likelihood = -39.288802  
Iteration 3:   log likelihood =  -39.28864  
Iteration 4:   log likelihood =  -39.28864  

Logistic regression                             Number of obs     =         74
                                                LR chi2(1)        =      11.49
                                                Prob > chi2       =     0.0007
Log likelihood =  -39.28864                     Pseudo R2         =     0.1276

------------------------------------------------------------------------------
     foreign |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         mpg |   .1597621   .0525876     3.04   0.002     .0566922     .262832
       _cons |  -4.378866   1.211295    -3.62   0.000    -6.752961   -2.004771
------------------------------------------------------------------------------

. bootstrap (e(ll)), reps(200) : logit foreign mpg
(running logit on estimation sample)

Bootstrap replications (200)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5 
..................................................    50
..................................................   100
..................................................   150
..................................................   200

Logistic regression                             Number of obs     =         74
                                                Replications      =        200

      command:  logit foreign mpg
        _bs_1:  e(ll)

------------------------------------------------------------------------------
             |   Observed   Bootstrap                         Normal-based
             |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |  -39.28864   4.084001    -9.62   0.000    -47.29314   -31.28414
------------------------------------------------------------------------------

. estat bootstrap, all

Logistic regression                             Number of obs     =         74
                                                Replications      =        200

      command:  logit foreign mpg
        _bs_1:  e(ll)

------------------------------------------------------------------------------
             |    Observed               Bootstrap
             |       Coef.       Bias    Std. Err.  [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _bs_1 |   -39.28864   .9831143   4.0840012   -47.29314  -31.28414   (N)
             |                                      -45.72757  -29.04522   (P)
             |                                      -46.91825  -31.80264  (BC)
------------------------------------------------------------------------------
(N)    normal confidence interval
(P)    percentile confidence interval
(BC)   bias-corrected confidence interval

.

[Radhouene DOGGUI]

Radhouenne:
as far as your first question is concerned (I'm not familiar with Lo-Mendell-Rubin adjusted likelihood ratio test), the following toy-example may be helpful:

. use "C:\Program Files\Stata16\ado\base\a\auto.dta"

. logit foreign mpg

Iteration 0: log likelihood = -45.03321
Iteration 1: log likelihood = -39.380959
Iteration 2: log likelihood = -39.288802
Iteration 3: log likelihood = -39.28864
Iteration 4: log likelihood = -39.28864

Logistic regression Number of obs = 74
LR chi2(1) = 11.49
Prob > chi2 = 0.0007
Log likelihood = -39.28864 Pseudo R2 = 0.1276

------------------------------------------------------------------------------
foreign | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
mpg | .1597621 .0525876 3.04 0.002 .0566922 .262832
_cons | -4.378866 1.211295 -3.62 0.000 -6.752961 -2.004771
------------------------------------------------------------------------------

. bootstrap (e(ll)), reps(200) : logit foreign mpg
(running logit on estimation sample)

Bootstrap replications (200)
----+--- 1 ---+--- 2 ---+--- 3 ---+--- 4 ---+--- 5
.................................................. 50
.................................................. 100
.................................................. 150
.................................................. 200

Logistic regression Number of obs = 74
Replications = 200

command: logit foreign mpg
_bs_1: e(ll)

------------------------------------------------------------------------------
| Observed Bootstrap Normal-based
| Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_bs_1 | -39.28864 4.084001 -9.62 0.000 -47.29314 -31.28414
------------------------------------------------------------------------------

. estat bootstrap, all

Logistic regression Number of obs = 74
Replications = 200

command: logit foreign mpg
_bs_1: e(ll)

------------------------------------------------------------------------------
| Observed Bootstrap
| Coef. Bias Std. Err. [95% Conf. Interval]
-------------+----------------------------------------------------------------
_bs_1 | -39.28864 .9831143 4.0840012 -47.29314 -31.28414 (N)
| -45.72757 -29.04522 (P)
| -46.91825 -31.80264 (BC)
------------------------------------------------------------------------------
(N) normal confidence interval
(P) percentile confidence interval
(BC) bias-corrected confidence interval

.

Thank you so much Carlo.

I used the syntax that you provided.

bootstrap (e(ll)), reps(200) : traj if gender==1, var(bmi*) indep(age*) model(beta) order(2 2 2 )
bootstrap (e(ll)), reps(200) : traj if gender==1, var(bmi*) indep(age*) model(beta) order(2 2 )

However, because I am dealing with latent variables class analysis , specifically the group based trajectory modeling. I don't know how to interpret that, specifically how to compare what model is better with 3 or 2 latent variables. There is no P value to compare the both models.


Do you have any feedback about that.

Regards.

[Carlo Lazzaro]

No, I do not.
Sorry for that.