/cut1 without z score using Ologit Command

[sara jh]

Dear All,

I would like to kindly ask you about why there is not z value for the intercept when I use the Ologit?

Thank you,

[Clyde Schechter]

You can always calculate it yourself as the coefficient divided by its standard error. But what would you use it for if you had it? There is almost no circumstance in which it makes any sense to test the hypothesis that a cutpoint in this kind of model equals 0.

[Leonardo Guizzetti]

Clyde is correct. There is no scientific interest in testing hypotheses on the cut point values themselves. They are simply a mathematical abstraction of convenience. The cut points also can't be compared across models, so there again it's not interesting.

You can derive the z-statistics also by using -test-, and then taking the square-root of the resulting chi-2 statistic with 1 degree of freedom.

[sara jh]

Thank you very much.