Hello all, I am working on a research project and trying test an issue pre and post an event. I'm concerned that the distribution of my sample may be posing a problem with my final conclusions. I would like to see if my concerns are legitimate.
I am comparing two samples pre and post an event. The two samples are Incr and Decr. The sample distributions (# of observations) are as follows: Sample: Incr Decr Pre-period: 13 6 Post-period: 26 18 I am running an OLS regression in SAS and I am also using a dummy to denote 1 for the post period and 0 for the pre period to increase the overall sample size when running the regression rather than comparing the two samples (pre and post) separately. Adding a dichotomous variable may or may not be a good idea. My regression looks like the following: Y = D*X1 for sample Incr where D equals 1 if post and zero otherwise Y = D*X2 for sample Decr where D equals 1 if post and zero otherwise **Is there a way to statistically control for this imbalance in the data for the pre and post periods?? Any comments/suggestions would be greatly appreciated. Thanks. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
