Ordinal_statement_for_Ordinal

Ordinal statement for Ordinal Responses

The ordinal statement is a variation of the multi statement (see the “Multi statement for Categorical models” section) and its syntax is given by:

    ordinal(observedVariable, inverseLinkFunction, 
       explanatoryVariable, beta, alpha0, alpha1, …)

For example:

   ordinal(Y, ilogit, C, slope, intercept0, intercept1, …)

where the intercepts are in numerically ascending order. It fits the model with the probability for Y >= i defined by:

   P(Y >= i | C)=ilogit(C*slope + intercepti)

And if there are m values of Y, they are 0, 1, … m – 1, and there are m – 1 intercepts.

This is equivalent to the multi statement:

   multi(Y, ilogit, -(C*slope+intercept0), -(C*slope+intercept1), …)

Observation statement action codes

Action codes may be given, within an observation statement, which specify computations to make either before or after an observation is made. For instance, if there is a urine compartment that is emptied after each observation, part of the model may resemble this:

   deriv(A0 = gfr*Cp)
   observe(UrineObs = A0 + UrineEps, doafter={A0=0})