Doubt with Margins and mlogit

[Ruben Lopez]

Hello! Thank you for your help before anything!
r
I'm doing a mlogit in STATA and I want to know how to interpreter the margins command.

I have my mlogit regression and then I do the margins, and my question is:

I have my dependent variable Y1 where Y1 can take 4 different values: Y1.0 Y1.1 Y1.2 Y1.3, and after this, I have different independent variables such as X1 X2 X3 X4 X5... where for example, X1 can take 5 different values such as X1.1 X1.2 X1.3 X1.4 X1.5 ...

I put that X1.3 is the base outcome, so I have the following output when I do margins:

margins, dydx(*)

Average marginal effects Number of obs = 5,884
Model VCE : Robust

dy/dx w.r.t. : 1.X1 2.X1 4.X1 5.X1 2.SEX 2.X3 3.X4 1.X5
1.X6 1.X7 1.X8 1.X9
1._predict : Pr(Y1==0), predict(pr outcome(0))
2._predict : Pr(Y1==1), predict(pr outcome(1))
3._predict : Pr(Y1==2), predict(pr outcome(2))
4._predict : Pr(Y1==3), predict(pr
outcome(3))

------------------------------------------------------------------------------
| Delta-method
| dy/dx Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
1.X1 |
_predict |
1 | -.1512682 .0204401 -7.40 0.000 -.1913301 -.1112063
2 | .0592488 .0157732 3.76 0.000 .0283339 .0901637
3 | -.0428667 .0165899 -2.58 0.010 -.0753824 -.010351
4 | .134886 .0223512 6.03 0.000 .0910785 .1786936
-------------+----------------------------------------------------------------
2.X1 |
_predict |
1 | -.0652237 .0260257 -2.51 0.012 -.116233 -.0142143
2 | .0008017 .0153294 0.05 0.958 -.0292433 .0308468
3 | .0044417 .0211679 0.21 0.834 -.0370465 .04593
4 | .0599802 .0255664 2.35 0.019 .0098709 .1100894
-------------+----------------------------------------------------------------
3.X1 | (base outcome)
-------------+----------------------------------------------------------------
4.X1 |
_predict |
1 | .1358043 .0172819 7.86 0.000 .1019324 .1696762
2 | -.0245283 .0089672 -2.74 0.006 -.0421037 -.0069528
3 | -.0071706 .0133511 -0.54 0.591 -.0333382 .018997
4 | -.1041054 .0146178 -7.12 0.000 -.1327558 -.075455
-------------+----------------------------------------------------------------
5.X1 |
_predict |
1 | .1128886 .0239924 4.71 0.000 .0658644 .1599129
2 | -.0268576 .0121647 -2.21 0.027 -.0507001 -.0030151
3 | -.0166604 .0185078 -0.90 0.368 -.052935 .0196141
4 | -.0693706 .0205889 -3.37 0.001 -.1097241 -.0290171
-------------+----------------------------------------------------------------
1.SEX | (base outcome)
-------------+----------------------------------------------------------------
2.SEX |
_predict |
1 | .0329718 .0133317 2.47 0.013 .0068423 .0591014
2 | -.0121747 .0077655 -1.57 0.117 -.0273948 .0030453
3 | -.0356784 .0107789 -3.31 0.001 -.0568047 -.0145521
4 | .0148813 .0124139 1.20 0.231 -.0094495 .0392121
-------------+----------------------------------------------------------------

Should I interpret this like the probability of an individual who belongs to X1.5 to be in Y1.1 is 11% higher than someone who belongs to the group of reference X1.3???

Another example, if someone who belongs to X1.4, the probabilty to be in Y1.1 is 13,5% higher than someone who belongs to X1.3 (base outcome)?

And then, for example, if SEX == 1 is a man, if SEX == 2 is a woman so does the output tell me that, for example, a woman has -3.56% probabilty to belongs to Y1.3 than man?

Is this right? What is the correct interpretation?

Thank you for your help,
kind regards,
Rubén

[Richard Williams]

This handout might help:

https://www3.nd.edu/~rwilliam/xsoc73994/Margins05.pdf

[Ruben Lopez]

This handout might help:

https://www3.nd.edu/~rwilliam/xsoc73994/Margins05.pdf

Thank you very much, I think I have understood it well, It's the interpretation that I thought.

I have also found the mtable command and it's interesting.

Thank you very much Richard for your help.

Kind regards,
Rubén L