st: Formal question: transformation of variables and diff-in-diff
Date
Mon, 25 Jul 2011 16:47:51 -0300
Hi all
I would like to ask you a simple but formal question:
I observe a variable X at two points in time to treatment and control
groups, which were determined by common support (pscore algorithm). I
got an estimate of the diff-in-diff W.
If I do Y = X/100, match again treated and control within the common
support, will I find the same groups? If so, the estimate of
diff-in-diff in this new scenario will be simply W/100?