Multinomial Logistic Regression - Missing Category

[Giacomo Puccio]

Greetings fellow Statalists,

I'm a university student currently working on an investigation about smallfarmer market participation and i'm using a multinomial logistic approach in order to fulfill this. However i'm having a problem in terms of the output i'm getting.

My dependent variable categories are:
0 if the farmer does not participate in markets
1 if the farmer sells his/her products on the local market
2 if the farmer sells his/ her products on the external market

However, when i run the model i get that the base category (0) is missing, the (1) category is being used as a base and i only get the output of the (2) category taking (1) as a base.

Since this was not the desired output i tried to force Stata into taking (0) as a base category by using ,base(0) command. However, when i did that i got an error that said "baseline category missing".

I don't really understand why this is happening since i do get the three categories as an output when i tabulate the variable. I was wondering if any expert could help me to solve this issue.

Thank you in advance!

[Weiwen Ng]

In any regression command*, Stata will drop all observations with missing data on any of the variables (dependent or independent) specified in the regression. I suspect that the people who don't participate in markets have missing independent variables. I'd suggest checking that.

* it's a bit more complicated than this, but this is generally true.

[Giacomo Puccio]

I cheked and there were missing values on the regressors as you said. I'll recode them into 0 since it would represent no participation.

Thank you for the advice!