Interpretation of Probit-Model

[Marc Almand]

Hey guys, I need your help. I want to run a probit model with following variables: y=healthstatus which has 4 categories (very bad, bad, good, very good) and x=age.

I used the following command:

oprobit healthstatus c.age, r

How do I interpret the coefficient of age (=0.0232125)? If age increases by one unit, then on average the probability of being in a high (health) category ('good' or 'very good') increases, ceteris paribus.

[Fei Wang]

Marc, you need to use -margins- after -oprobit- to obtain interpretable results.

oprobit healthstatus c.age, r
margins, dydx(*)

You may find each category of healthstatus owns a coefficient -- it means if age increases by one unit, how much the probability of being in a specific category of healthstatus will increase by.

[Marc Almand]

Marc, you need to use -margins- after -oprobit- to obtain interpretable results.

oprobit healthstatus c.age, r
margins, dydx(*)

You may find each category of healthstatus owns a coefficient -- it means if age increases by one unit, how much the probability of being in a specific category of healthstatus will increase by.

Thanks. Is it possible to interpret the sign of the coefficient? Is it fair to say that if age goes up, then you're more likely to end up in a higher category, so you're healthier?

[Fei Wang]

Thanks. Is it possible to interpret the sign of the coefficient? Is it fair to say that if age goes up, then you're more likely to end up in a higher category, so you're healthier?

A positive coefficient means an increase in probability associated with rising age, while a negative coefficient means a decrease in probability associated with rising age. You may make corresponding conclusions based on the actual results.