Date: Tue, 23 Jan 2007 13:00:07 -0800
Reply-To: Joep <joeparts@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Joep <joeparts@GMAIL.COM>
Subject: Re: Creating a new group id by combining two other group id's via
Content-Type: text/plain; charset="us-ascii"
I apologize for the confusion.
I am doing a meta-analysis.
So basically a group is a set of correlations between two variables. As
I focus on 5 dependent and 20 independent variables, there are 100
combinations. For each combination, I want to calculate the average
correlation (and a whole long list of other statistics).
Now groupd orders the dependent variables into 1 to 5; groupi orders
the independent variables into 1 to 20. Combining both into group would
lead to 100 combinations. Each combination is then used in the next
steps of my analyses.
Does this help?