0

I have estimated an ordinal probit model in Stata. The dependent variable is walkability. The main independent variables are on a Likert scale (1=agree, 2=partially agree, 3=disagree). The other variables are dummies.

After estimation I have calculated margins for each level of the outcome. Stata reports results for 2,3 options and use 1st option of the independent variable as base category.Thus at outcome 1 I have prob for 2 and 3 options of the independent variable and the 1st option is missed (baseline). While I do understand the interpretation of margins of factor variables, I am confused now with the dummy variables like gender, age, uni. Stata reports their margins. How will they be interpreted?

Please quote an example of ordered probit when both dependent and independent are factor variables.

Nick Cox
  • 56,404
  • 8
  • 127
  • 185
numra
  • 51
  • 1
    There was an example in the previous question you asked. – dimitriy May 02 '15 at 19:29
  • Yes i did reviewed the previous question. however now i am more concerned about interpretation. While it does explain how to interpret margins of factor indep variable, how will you interpret margins for other indep variables that are dummies let say gender, age? – numra May 03 '15 at 11:55
  • All examples of ordinal probit have the outcome as an ordered factor variable; that follows from the definition of the model. – Nick Cox May 03 '15 at 12:05
  • @numra The sentence with the number .5727857 gives an example of that. – dimitriy May 03 '15 at 15:17
  • I have got that, and i did tried to explain that i am not asking for the main indep variable. i am asking about other indep variables like gender (0 female 1 male), how will you intrepet its margins? – numra May 03 '15 at 18:34

0 Answers0