I am running Statgraphics Centurion version 16.2.04, 64-bit.

I need to run a multiple regression analysis in which the dependent variable is ordinal on a scale of 0 to 5, integers only. Apparently the appropriate model for this is the ordered logit model (aka ordered logistic regression aka proportional odds model) but I cannot figure out how to do it in Statgraphics. I have read through the manual and it only contains information for a standard logistic regression in which the outcome variable is binary. Does anyone know how to do this in Statgraphics?

I appreciate your help.

EDIT: to be clear, the outcome variable is a scale measuring the severity of a certain disease (5 is the worst). I want to evaluate the effects of different factors (such as age or tobacco use) on the change in the score after treatment compared to the score before treatment. In other words, the question is: do patients with variable x have a lower improvement in the severity of their condition compared to patients without variable x? I want to do this in a multiple regression fashion, controlling for many variables. My understanding is that since the dependent variable in the ordered logit model is ordinal categories, my outcome variable will be the "categories" of change in their score, ie by 1 point, 2 points, 3 points. etc. Thanks. ]]>

Hope someone can help. I have a repeated measures (4) ancova I need to run with no between subjects factor. I have 5 covariates. Im trying to get the effect of 1 covariate on the repeated measures when accounting for the other 4 covariates. How best to do this?

Thanks in advance. ]]>

= risk of following input x

= cost of following input x when in fact input x' occured

= known pdf of inputs evaluated for input x'

Objective : obtain a decent estimate of the integral

Challenge : is computationally expensive and can only be evaluated for a handful (less than 50) of x,x's.

Context : x is the input of a complex optimization program. To evaluate we have to find the optimal solution of the optimization program once for x and once for x'. optimal solutions are compared against each other to find their relative costs ,i.e., . ]]>