Dear Statalist,
I am using gologit2 for predicting the impact of mainly educational level (low, mid, high) and employment status (employed, unemployed, housewife, other) as well as the interactions of these on happiness (very unhappy, unhappy, neither nor, happy, very happy). I have some questions, and would be grateful if you could help.
At the end of my message, I am sharing the full model (the model is for married women between 25 and 40), and as a snapshot right below is the gologit2 result for the fourth level of happiness (happy) (fifth level (very happy) is the ref category), using edulevel##employmentstatus as the indep variable (low edu (1) and in paid work (1) are ref. categories):
mid edu x unemployed .69(.70)
mid edu x housewife .20(.25)
mid edu x other emp. status .53(.43)
high edu x unemployed -12.76(683.44)
high edu x housewife 1.17(.58)*
high edu x other emp. status 1.25 (.64)*
constant -2.49 (.14)***
My interpretation for interactions based on my research and reading other relevant papers is:
Controlling for educational level and employment status, rather than having low education, married women with medium-level education who are unemployed are statistically significantly more likely be in the first four categories of happiness relative to being very happy. (I am still not sure where to situate the ref category of employment in the interactions (which is "in paid work"))
- Is this interpretation right, and is employment status not a part of any interpretations because the variable is edu#emp and not emp#edu?
- Is it ok to use a separate model only for interactions (edu#employment rather than edu##employment) -this makes the coefficients more stat. significant?
- Is using "margins" adequate for predicted probabilities after gologit2?
Best regards,
A.
-------
Full model with the command:
gologit2 newhappiness i.edu2##i.employment2 if women==1 & married==1 & age>24 | age<41
Generalized Ordered Logit Estimates Number of obs = 4,778
LR chi2(44) = 199.60
Prob > chi2 = 0.0000
Log likelihood = -5892.8282 Pseudo R2 = 0.0167
-------------------------------------------------------------------------------
> ---
newhappiness | Coef. Std. Err. z P>|z| [95% Conf. Interv
> al]
-----------------+-------------------------------------------------------------
> ---
1 |
edu2 |
2 | .1547906 .358403 0.43 0.666 -.5476663 .8572
> 476
3 | 1.49217 1.032369 1.45 0.148 -.5312365 3.515
> 576
|
employment2 |
2 | -1.651732 .3934039 -4.20 0.000 -2.42279 -.8806
> 747
3 | -.0204169 .2829476 -0.07 0.942 -.574984 .5341
> 501
4 | -1.045575 .3869348 -2.70 0.007 -1.803953 -.2871
> 967
|
edu2#employment2 |
2 2 | 1.621811 .7507637 2.16 0.031 .1503407 3.09
> 328
2 3 | -.1515169 .4913804 -0.31 0.758 -1.114605 .8115
> 711
2 4 | 1.38689 .6868248 2.02 0.043 .040738 2.733
> 042
3 2 | 13.05099 903.986 0.01 0.988 -1758.729 1784.
> 831
3 3 | 11.7874 1086.223 0.01 0.991 -2117.17 2140.
> 745
3 4 | -.8174096 1.480582 -0.55 0.581 -3.719298 2.084
> 479
|
_cons | 3.666567 .2456152 14.93 0.000 3.18517 4.147
> 964
-----------------+-------------------------------------------------------------
> ---
2 |
edu2 |
2 | .1511321 .1644236 0.92 0.358 -.1711323 .4733
> 964
3 | .6260354 .309864 2.02 0.043 .0187132 1.233
> 358
|
employment2 |
2 | -1.458335 .2324866 -6.27 0.000 -1.914 -1.002
> 669
3 | .1839836 .1337072 1.38 0.169 -.0780776 .4460
> 449
4 | -.8204532 .2066869 -3.97 0.000 -1.225552 -.4153
> 542
|
edu2#employment2 |
2 2 | .4322724 .3254839 1.33 0.184 -.2056643 1.070
> 209
2 3 | .1121248 .2466386 0.45 0.649 -.3712781 .5955
> 276
2 4 | 1.317152 .3370311 3.91 0.000 .6565834 1.977
> 721
3 2 | .6097216 .7300508 0.84 0.404 -.8211516 2.040
> 595
3 3 | -.5666551 .8122474 -0.70 0.485 -2.158631 1.025
> 321
3 4 | .8627773 .8150215 1.06 0.290 -.7346355 2.46
> 019
|
_cons | 1.896557 .1136762 16.68 0.000 1.673756 2.119
> 358
-----------------+-------------------------------------------------------------
> ---
3 |
edu2 |
2 | .1165318 .1087664 1.07 0.284 -.0966465 .32
> 971
3 | .6586714 .1833428 3.59 0.000 .2993261 1.018
> 017
|
employment2 |
2 | -1.302325 .2399155 -5.43 0.000 -1.772551 -.8320
> 991
3 | .2517154 .0895231 2.81 0.005 .0762534 .4271
> 774
4 | -.2636517 .1689614 -1.56 0.119 -.5948099 .0675
> 065
|
edu2#employment2 |
2 2 | .4137267 .3083412 1.34 0.180 -.1906109 1.018
> 064
2 3 | -.0994364 .1539574 -0.65 0.518 -.4011874 .2023
> 146
2 4 | .345244 .2246619 1.54 0.124 -.0950852 .7855
> 733
3 2 | .3085669 .5443729 0.57 0.571 -.7583843 1.375
> 518
3 3 | .5337038 .6568673 0.81 0.417 -.7537324 1.82
> 114
3 4 | .291513 .48085 0.61 0.544 -.6509356 1.233
> 962
|
_cons | .2297448 .0770898 2.98 0.003 .0786516 .380
> 838
-----------------+-------------------------------------------------------------
> ---
4 |
edu2 |
2 | .1632639 .1962506 0.83 0.405 -.2213803 .547
> 908
3 | .265104 .2932042 0.90 0.366 -.3095657 .8397
> 737
|
employment2 |
2 | -1.00188 .6034976 -1.66 0.097 -2.184714 .1809
> 531
3 | .2835846 .1622066 1.75 0.080 -.0343345 .6015
> 036
4 | -.3208984 .3560915 -0.90 0.367 -1.018825 .3770
> 282
|
edu2#employment2 |
2 2 | .6861652 .7077328 0.97 0.332 -.7009655 2.073
> 296
2 3 | .198249 .2522358 0.79 0.432 -.2961241 .692
> 622
2 4 | .5332706 .4295388 1.24 0.214 -.30861 1.375
> 151
3 2 | -12.76436 683.437 -0.02 0.985 -1352.276 1326.
> 747
3 3 | 1.172605 .5788301 2.03 0.043 .0381194 2.307
> 092
3 4 | 1.251019 .6356624 1.97 0.049 .0051435 2.496
> 894
|
_cons | -2.494466 .1442841 -17.29 0.000 -2.777258 -2.211
> 675
-------------------------------------------------------------------------------
> ---
Comment