How do I set up contrasts in npar1way?

[Rick_SAS · Accepted solution]

You ask good questions, and unfortunately I don't have all the answers.

> I would very much like your opinion on the npar1way>multtest combination.

My opinion is that you should ask someone who knows more about those procedures than I do. 🙂

> something is wrong with the code you suggest. I ran the data using QuantReg and then with npar1way followed by multtest.

> The results were very very different: a p-value of 0.0542 from QuantReg vs a p-value of 0.0075 from multtest

Let me just say that these procedures are using different statistics and running different tests. For this very small data set (14 obs, joint levels have between 2 and 5 obs), the standard errors that PROC QUANTREG produces are larger than you are getting from your other method.  If you are happy with the other method and its assumptions, feel free to accept its conclusion. 

[normanbg]

Thank you for all the time you devoted to my problem.