Dear Statalist Team,
I have an ordered probit regression, where the outcome variable contains 4 possible outcomes. I have several explanatory variables (all of them are dummies or discrete except ln_hhincome which is continuous)
I have been reading on Statalist forum that the suitable command for marginal effects is:
margins, dydx(*) predict(outcome(1))
I would like to make sure that this is indeed true, as I have mostly dummy/discrete variables and got a bit confused when reading in the Forum.
Added below is the code and the output for my ordered probit regression:
oprobit life_satisfaction i.population i.female ln_hhincome i.age i.family_size i.pp_room i.employment i.religiosity i.health_status disability i.education i.marital_status i.hebrew_fluency i.volunteering i.region
Iteration 0: log likelihood = -5793.5337
Iteration 1: log likelihood = -5114.6857
Iteration 2: log likelihood = -5109.7377
Iteration 3: log likelihood = -5109.7363
Iteration 4: log likelihood = -5109.7363
Ordered probit regression Number of obs = 5,757
LR chi2(37) = 1367.59
Prob > chi2 = 0.0000
Log likelihood = -5109.7363 Pseudo R2 = 0.1180
----------------------------------------------------------------------------------------
life_satisfaction | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-----------------------+----------------------------------------------------------------
population | .4614532 .0546173 8.45 0.000 .3544052 .5685011
female | .0149484 .0322265 0.46 0.643 -.0482144 .0781112
ln_hhincome | .1626264 .0273989 5.94 0.000 .1089256 .2163272
|
age |
30-39 | -.1690995 .0568488 -2.97 0.003 -.2805211 -.0576779
40-49 | -.1634033 .0607061 -2.69 0.007 -.2823851 -.0444215
50-59 | -.1744873 .0661828 -2.64 0.008 -.3042032 -.0447713
60-74 | .0093207 .0710499 0.13 0.896 -.1299345 .1485758
|
family_size |
2 | .1266568 .0625239 2.03 0.043 .0041123 .2492013
3-5 | .0595687 .0666954 0.89 0.372 -.0711519 .1902893
6+ | .0515747 .0851436 0.61 0.545 -.1153037 .2184532
|
pp_room |
1-<2 | -.0394693 .0405352 -0.97 0.330 -.1189169 .0399783
2+ | -.1826042 .0823416 -2.22 0.027 -.3439908 -.0212176
|
employment |
employed | -.1243722 .0451227 -2.76 0.006 -.2128111 -.0359333
unemployed | -.3091662 .0927262 -3.33 0.001 -.4909062 -.1274261
|
religiosity |
traditional | .1777507 .0379347 4.69 0.000 .1034001 .2521013
religious | .3133971 .0513464 6.10 0.000 .21276 .4140342
very religious | .816839 .073988 11.04 0.000 .6718252 .9618528
|
health_status |
health is not so good | .4514483 .084381 5.35 0.000 .2860647 .616832
health is good | .7346467 .0881405 8.33 0.000 .5618945 .9073989
health is very good | 1.182111 .09201 12.85 0.000 1.001774 1.362447
|
disability | -.1145372 .0524588 -2.18 0.029 -.2173545 -.0117199
|
education |
educ BA max | .0562342 .0474255 1.19 0.236 -.0367181 .1491865
educ MA min | .0582294 .0645186 0.90 0.367 -.0682247 .1846835
|
marital_status |
married | .2259718 .0542659 4.16 0.000 .1196127 .332331
divorced/separated | -.0962337 .0716145 -1.34 0.179 -.2365955 .0441282
widow(er) | .1105644 .0888584 1.24 0.213 -.0635948 .2847236
|
hebrew_fluency |
weak | -.1509022 .1118857 -1.35 0.177 -.3701941 .0683897
moderate | -.1368649 .1067883 -1.28 0.200 -.3461662 .0724364
good | -.145148 .1048562 -1.38 0.166 -.3506623 .0603664
very good | .1088566 .104387 1.04 0.297 -.0957381 .3134514
|
volunteering |
volunteering | .2242197 .0402992 5.56 0.000 .1452348 .3032046
|
region |
jerusalem | -.2248031 .0625378 -3.59 0.000 -.3473749 -.1022312
haifa | -.2738027 .0583237 -4.69 0.000 -.3881151 -.1594902
center | -.1712356 .0524664 -3.26 0.001 -.274068 -.0684033
tel aviv | -.3260045 .0578918 -5.63 0.000 -.4394704 -.2125387
south | -.109568 .0597541 -1.83 0.067 -.2266839 .0075479
west bank | -.0483345 .1001799 -0.48 0.629 -.2446835 .1480145
-----------------------+----------------------------------------------------------------
/cut1 | .4896097 .2576887 -.0154509 .9946703
/cut2 | 1.5144 .2568057 1.01107 2.01773
/cut3 | 3.272835 .259405 2.76441 3.781259
----------------------------------------------------------------------------------------
Hope this is clear. Please let me know if you need any more clarification.
Many thanks
Related Posts with Marginal effect in an Ordered probit model with discrete variables
vce(conditional) or vce(delta) when predicting marginsIf I have used vce(robust) when estimating the IV probit model, do I need to use vce(unconditional) …
Calibration plots when using cross-validation for internal model validation - grateful for adviceI have developed a multivariable prediction model using logistic regression and am using cross-valid…
setting base level for margins with eydxHello. i'm running a maultivariable panel regression model with xtgee and a number of continuous and…
Extract last week row data from weekly dataHi. I'd like to ask for help due to difficulties during the stata work. I have weekly data. I'd lik…
Creating a Loop over a LoopGreetings I would like to run the following operation a 1000 times and store the coefficient estima…
Subscribe to:
Post Comments (Atom)
0 Response to Marginal effect in an Ordered probit model with discrete variables
Post a Comment