Ordered Logit, Multinomial Logit- Interpretations

[Aparna Anand]

I am trying to run ordered logit for a dependent variable that takes values -1,0,1. Suppose education ( measured as ordered categorical variable) as one of the independent variable and the coefficient in a simple regression of educ on the ordered dependent variable is -.27 and the respective cut1 is -.42 and cut 2 is .32. I am having difficulty in interpreting these.
I am interpreting the following way:
The ordered logit for individuals with higher education is .42 less to be in the category of 1 as compared to individuals with lower amount of education.
The cut 1 of -.42 is observations which had value of -.42 or less will be in -1 category of dependent variable given they have lesser amount of education.
The cut 2 of .32 means observations that had value of .32 and higher will be in category value of 1 of dependent variable. The observations that had value between -.42 and .32 will be in the category of 0 of dependent variable.

Will this be right? Thanks

[Hong Il Yoo]

Given your modelling assumption and estimation results, there's a latent dependent variable y* = -0.27 * educ + standard logistic errors. If y* happens to be smaller than -0.42, your observed dependent variable is in category 1. If y* is between -0.42 and 0.32, your observed dependent variable is in category 2. If y* happens to be greater than 0.32, your observed dependent variable is in category 3.