02-28-2013 09:13 AM
I have an odd situation I am trying to figure out. I have two tables containing summarized data at a dealer level, treatment versus control.
Let's say I have Dealer_Num, Net_Revenue, and Customer_Count for a time period summarized by dealer. I have 127 dealers in my treatment group and 380 dealers in my control group.
Since not all the Control dealers have similar volume to the Treatment dealers, I want to select those most alike.
I currently have this running in Excel Solver, using SUMPRODUCT and a binary flag to "turn on or off" a dealer, and minimize the difference between the two groups. Is there a way (presumably quicker) to run this in SAS?
02-28-2013 09:23 AM
This is a common problem in epidemiology for case-control matching. A typical problem there might be matching by nearest age within 5 years. It is also called proximity matching. A generalization to multiple variables is often done by propensity score matching.
If you search for those terms on sas.com, you will find a number of examples to choose from.
Need further help from the community? Please ask a new question.