After I read this, I immediately checked WKSSTAT.OGS and found that - sure enough - the LabTalk checks to make sure there are no more than nine Y datasets selected. I tried changing the LabTalk to allow more datasets and the Multiple Regression routine happily chugged away and spit out a table with more than nine datasets, but it only had results for the first nine. This means the limitation is in the source code so it can't be corrected without a new build.
P.S. How many datasets is reasonable? As you increase the number, the math gets indistinguishable from noise.
Edited by - greg on 05/07/2002 11:03:20 |
stat.mr with more than 9 Datasets?
Posted - 04/30/2002 : 08:01:44 AM
