log type: text
opened on: 10 Oct 2005, 11:01:55
. table wed hed, contents(mean endogdm)
----------------------------------
| hed
wed | 1 2 3 4
----------+-----------------------
1 | 1 0 0 0
2 | 0 1 0 0
3 | 0 0 1 0
4 | 0 0 0 1
----------------------------------
. table wed hed, contents (mean endog)
----------------------------------
| hed
wed | 1 2 3 4
----------+-----------------------
1 | 1 0 0 0
2 | 0 2 0 0
3 | 0 0 3 0
4 | 0 0 0 4
----------------------------------
. desmat: poisson count wed hed endogdm
------------------------------------------------------------------------------------------
Poisson regression
------------------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 16
Initial log likelihood: -221501.223
Log likelihood: -41944.565
LR chi square: 359113.316
Model degrees of freedom: 7
Pseudo R-squared: 0.811
Prob: 0.000
------------------------------------------------------------------------------------------
nr Effect Coeff s.e.
------------------------------------------------------------------------------------------
count
wed
1 2 0.979** 0.005
2 3 0.608** 0.005
3 4 0.081** 0.005
hed
4 2 0.740** 0.005
5 3 0.414** 0.005
6 4 0.216** 0.005
endogdm
7 1 1.115** 0.003
8 _cons 9.067** 0.005
------------------------------------------------------------------------------------------
* p < .05
** p < .01
. set linesize 79
. desmat: poisson count wed hed endog
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 16
Initial log likelihood: -221501.223
Log likelihood: -24059.274
LR chi square: 394883.898
Model degrees of freedom: 10
Pseudo R-squared: 0.891
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
wed
1 2 1.372** 0.007
2 3 1.020** 0.007
3 4 -0.278** 0.008
hed
4 2 1.134** 0.007
5 3 0.819** 0.006
6 4 -0.017* 0.007
endog
7 1 1.722** 0.009
8 2 0.676** 0.007
9 3 0.537** 0.008
10 4 2.487** 0.009
11 _cons 8.652** 0.008
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 47932.55
Prob > chi2(5) = 0.0000
. display 83703-47932
35771
. display chi2tail(3,35771)
0
. display chi2tail(3,100)
1.554e-21
. lincom _x_7-_x_9
( 1) [count]_x_7 - [count]_x_9 = 0
------------------------------------------------------------------------------
count | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
(1) | 1.184561 .0123912 95.60 0.000 1.160275 1.208847
------------------------------------------------------------------------------
. *Note: We know because the model with 4 endogamy terms fits much better than
the model with 1 endogamy term, that educational endogamy varies across educational groups.
. *In order to say which level of educational endogamy is the highest, we can compare the coefficients, and test them.
. table hed wed, contents (mean eddiff3)
----------------------------------
| wed
hed | 1 2 3 4
----------+-----------------------
1 | 0 0 0 1
2 | 0 0 0 0
3 | 0 0 0 0
4 | 1 0 0 0
----------------------------------
. desmat: poisson count wed hed endog eddiff3
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 16
Initial log likelihood: -221501.223
Log likelihood: -17940.195
LR chi square: 407122.056
Model degrees of freedom: 11
Pseudo R-squared: 0.919
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
wed
1 2 1.132** 0.007
2 3 0.815** 0.007
3 4 -0.276** 0.008
hed
4 2 0.942** 0.007
5 3 0.667** 0.007
6 4 0.009 0.007
endog
7 1 1.410** 0.010
8 2 0.796** 0.007
9 3 0.583** 0.007
10 4 2.147** 0.010
eddiff3
11 1 -1.947** 0.023
12 _cons 8.964** 0.008
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 35694.39
Prob > chi2(4) = 0.0000
. *That's an improvement in the LRT goodness of fit on 1 df from the
. *previous model, but it is still far, far away from the actual data
. desmat: poisson count wed hed endog eddiff3 eddiff2
-------------------------------------------------------------------------------
Poisson regression
-------------------------------------------------------------------------------
Dependent variable count
Optimization: ml
Number of observations: 16
Initial log likelihood: -221501.223
Log likelihood: -145.628
LR chi square: 442711.189
Model degrees of freedom: 12
Pseudo R-squared: 0.999
Prob: 0.000
-------------------------------------------------------------------------------
nr Effect Coeff s.e.
-------------------------------------------------------------------------------
count
wed
1 2 0.817** 0.008
2 3 0.461** 0.007
3 4 -0.142** 0.009
hed
4 2 0.627** 0.008
5 3 0.355** 0.007
6 4 0.180** 0.008
endog
7 1 0.763** 0.011
8 2 0.779** 0.007
9 3 0.601** 0.008
10 4 1.195** 0.011
eddiff3
11 1 -2.749** 0.024
eddiff2
12 1 -1.068** 0.006
13 _cons 9.611** 0.009
-------------------------------------------------------------------------------
* p < .05
** p < .01
. poisgof
Goodness-of-fit chi2 = 105.2568
Prob > chi2(3) = 0.0000
. table hed wed, contents (mean eddiff2)
----------------------------------
| wed
hed | 1 2 3 4
----------+-----------------------
1 | 0 0 1 0
2 | 0 0 0 1
3 | 1 0 0 0
4 | 0 1 0 0
----------------------------------
. *let's take a look at how this model fits.
. predict M6
(option n assumed; predicted number of events)
. table hed wed, contents (sum count sum M6) row col
------------------------------------------------------------
| wed
hed | 1 2 3 4 Total
----------+-------------------------------------------------
1 | 32016 33374 8407 988 74785
| 32016 33801.72 8138.919 828.3573 74785
|
2 | 28370 137876 43783 8446 218475
| 27942.28 137876 44327.54 8329.189 218475
|
3 | 7051 48766 61633 18195 135645
| 7319.081 48221.46 61633 18471.45 135645
|
4 | 984 13794 28635 51224 94637
| 1143.643 13910.81 28358.55 51224 94637
|
Total | 68421 233810 142458 78853 523542
| 68421 233810 142458 78853 523542
------------------------------------------------------------
. *One way to look at this is the simple residuals
. gen M6_resid=M6-count
. table hed wed, contents (sum M6_resid)
------------------------------------------------------
| wed
hed | 1 2 3 4
----------+-------------------------------------------
1 | 0 427.7227 -268.0806 -159.6427
2 | -427.7227 0 544.5352 -116.8105
3 | 268.0806 -544.5352 0 276.4531
4 | 159.6427 116.8105 -276.4531 0
------------------------------------------------------
. table hed wed, contents (sum M6_resid) row col
-----------------------------------------------------------------
| wed
hed | 1 2 3 4 Total
----------+------------------------------------------------------
1 | 0 427.7227 -268.0806 -159.6427 -.0006104
2 | -427.7227 0 544.5352 -116.8105 .0019531
3 | 268.0806 -544.5352 0 276.4531 -.0014648
4 | 159.6427 116.8105 -276.4531 0 .0001221
|
Total | .0006104 -.0019531 .0014648 -.0001221 0
-----------------------------------------------------------------
. *one simple way to standardize the residuals is to use the Pearson residuals
. poisgof, pearson
Goodness-of-fit chi2 = 105.9502
Prob > chi2(3) = 0.0000
. gen M6_pearson_resid= (M6_resid^2)/ M6
. table hed wed, contents (sum M6_pearson_resid) row col
------------------------------------------------------------
| wed
hed | 1 2 3 4 Total
----------+-------------------------------------------------
1 | 0 5.412347 8.830066 30.76666 45.00908
2 | 6.547307 0 6.689263 1.638179 14.87475
3 | 9.819156 6.149098 0 4.137537 20.10579
4 | 22.28475 .9808705 2.695002 0 25.96062
|
Total | 38.65121 12.54232 18.21433 36.54238 105.9502
------------------------------------------------------------
. * Here is a clue to where we might try to apply another degree of freedom into our model- the two furthest out cells, where education =(1,4) and (4,1) fit poorly and contribute a lot to our pearson chisquare test. One additional term would fit both exactly, because we already have one term for these two cells (eddiff3) in the model.
. log close
log type: text
closed on: 10 Oct 2005, 11:56:25
-------------------------------------------------------------------------------