[R] [Not R question]: Better fit for order probit model

[Robert A LaBudde]

At 03:17 AM 6/16/2007, adschai wrote:
>Thank you so much Robert. I haven't thought about the idea of 
>clumping categories together. One of the reason is because these 
>categories are bridge condition rating scores. They indeed represent 
>different meaning and serviceability conditions. They vary from 0-9. 
>I have about 300,000 data in which the first 5 labels, i.e. 0-4, are 
>bad condition bridge and there are only less than 1000 instances in 
>total. The worst case, is for example, score 0 (meaning the bridge 
>is not operatable), I have 60 instances. Score 1 I have about 100.
>
>I would appreciate if you could provide some opinion as to how you 
>would make the order probit fits better in this case? Thank you so 
>much in advance.

You certainly appear to have enough data to populate these 
categories. Your problems in a getting a good fit may relate to other problems.

You need to supply more information in order to say more.

What are the definitions of each category?

Is the ordering consistent, or are there really two different scales, 
one for bridge with essentially no problems, and another for those 
with serious damage?

What evidence do you have that your fit is poor?

What model are you fitting?

Etc.

================================================================
Robert A. LaBudde, PhD, PAS, Dpl. ACAFS  e-mail: ral at lcfltd.com
Least Cost Formulations, Ltd.            URL: http://lcfltd.com/
824 Timberlake Drive                     Tel: 757-467-0954
Virginia Beach, VA 23464-3239            Fax: 757-467-2947

"Vere scire est per causas scire"