---------------------------------------------------------------------------------------------------------------------------

      name:  <unnamed>

       log:  C:\Users\mexmi\Documents\newer web pages\soc_meth_proj3\fall_2021_logs\class13.log

  log type:  text

 opened on:   1 Nov 2021, 09:57:31

 

 

. search medeff

 

. which medeff

c:\ado\plus\m\medeff.ado

 

.

. *class starts here. Let’s look at 2 models from the Excel sheet LRT page, Models M2 and M3.

 

. logit married age age_sq if age>16

 

Iteration 0:   log likelihood = -67581.342 

Iteration 1:   log likelihood = -59611.157 

Iteration 2:   log likelihood = -59606.232 

Iteration 3:   log likelihood = -59606.232 

 

Logistic regression                                   Number of obs =   99,215

                                                      LR chi2(2)    = 15950.22

                                                      Prob > chi2   =   0.0000

Log likelihood = -59606.232                           Pseudo R2     =   0.1180

 

------------------------------------------------------------------------------

     married | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]

-------------+----------------------------------------------------------------

         age |   .2399034   .0021659   110.76   0.000     .2356583    .2441486

      age_sq |  -.0022193   .0000215  -103.23   0.000    -.0022614   -.0021771

       _cons |  -5.263343   .0493725  -106.60   0.000    -5.360112   -5.166575

------------------------------------------------------------------------------

 

. logit married age age_sq yrsed if age>16

 

Iteration 0:   log likelihood = -67581.342 

Iteration 1:   log likelihood = -59455.156 

Iteration 2:   log likelihood = -59450.943 

Iteration 3:   log likelihood = -59450.942 

 

Logistic regression                                   Number of obs =   99,215

                                                      LR chi2(3)    = 16260.80

                                                      Prob > chi2   =   0.0000

Log likelihood = -59450.942                           Pseudo R2     =   0.1203

 

------------------------------------------------------------------------------

     married | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]

-------------+----------------------------------------------------------------

         age |     .23473   .0021882   107.27   0.000     .2304412    .2390188

      age_sq |  -.0021575   .0000218   -99.05   0.000    -.0022002   -.0021148

       yrsed |   .0404459   .0022919    17.65   0.000     .0359539    .0449379

       _cons |  -5.698349   .0557695  -102.18   0.000    -5.807655   -5.589043

------------------------------------------------------------------------------

 

. *calculating the LRT between the models:

 

. display 16260.80-15950.22

310.58

 

 

* The chisquare stat is 310.58 on 1 df. And the probability of this is:

 

. display chi2tail(1,310.58)

1.633e-69

 

. *calculating the BIC between the models. First calculate the BIC.

 

. display -310.58+(1*ln(99215))

-299.07496

 

. * BIC test strongly also prefers the second model with the extra term because BIC <-10.

 

. *Bayes factor of M3 compared to M2:

 

. display exp(-(-299.07)/2)

8.754e+64

 

. * To generate a BIC P value of the kind we are used to, in this case would be P value of M2 compared to M3:

 

. display 1/(8.754e+64+1)

1.142e-65

 

. clear all

 

. use "C:\Users\mexmi\Documents\current class files\intro soc methods\GSS_subset_200610_panel.dta"

 

* A quick look at the GSS panel data for HW4:

 

. tabulate year marhomo_approve, row col

 

+-------------------+

| Key               |

|-------------------|

|     frequency     |

|  row percentage   |

| column percentage |

+-------------------+

 

   year_1: |

  gss year |

  for this |       homosexuals should have the right to marry

respondent | strongly    disagree  neither a      agree  strongly  |     Total

-----------+-------------------------------------------------------+----------

      2006 |       478        222        180        259        208 |     1,347

           |     35.49      16.48      13.36      19.23      15.44 |    100.00

           |     44.42      42.05      40.54      39.24      39.54 |     41.65

-----------+-------------------------------------------------------+----------

      2008 |       336        168        153        207        169 |     1,033

           |     32.53      16.26      14.81      20.04      16.36 |    100.00

           |     31.23      31.82      34.46      31.36      32.13 |     31.94

-----------+-------------------------------------------------------+----------

      2010 |       262        138        111        194        149 |       854

           |     30.68      16.16      13.00      22.72      17.45 |    100.00

           |     24.35      26.14      25.00      29.39      28.33 |     26.41

-----------+-------------------------------------------------------+----------

     Total |     1,076        528        444        660        526 |     3,234

           |     33.27      16.33      13.73      20.41      16.26 |    100.00

           |    100.00     100.00     100.00     100.00     100.00 |    100.00

 

* To download the medeff package which is required for one part of HW4 Q3, type this and then click on the top link and then the link to install:

 

search medeff

 

 

. log close

      name:  <unnamed>

       log:  C:\Users\mexmi\Documents\newer web pages\soc_meth_proj3\fall_2021_logs\class13.log

  log type:  text

 closed on:   1 Nov 2021, 13:57:58

---------------------------------------------------------------------------------------------------------------------------