Hello all,
I should say that the question I posted here is more related to
statistical analysis than Splus itself. But I found that this email list
is a great source for interacting with people in the statistical field.
The question is how to assess the degree of variation of a response
variable is likely or not due to the variation of some explanatory
variables. The idea is that if you observed a response variable and saw
little variation, you probably will not pursue an analysis of this
response variable.
The case where I encountered is that we have more than 1000 response
variables and try to prioritize which one to analyze first. Any
thoughts, references are welcome.
Thanks in advance.
Ping Zhan
Biostatistician
Genaissance Pharmaceuticals
5 Science Park
New Haven, CT 06511
p.zhan@genaissance.com
|