[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: RE: Two questions about ordinal regression using ologit and ocratio
> Hello there,
> I have two *quick* questions.
> 1) Am I right in thinking that ologit is the 'cumulative odds model'
> for ordinal response variables (as described in Armstrong and Sloan,
After locating a set of manuals it appears that I am.
Sorry about that. Thought the help menus and web would give me
all the same info.
> 2) Is there available a generalisation of the ocratio function
> that doesn't assume a constant slope?
> I have a 3-level outcome and want to compare A versus B and C
> along with B versus C. However, I want to end up with two sets
> of coefficients rather than the one set that ocratio will give me.
> I have experimented with mlogit but am so far unable to have
> two different reference groups using this method.
* For searches and help try: