Hi Roshan,
Since you have not mentioned the size of each group or the matching algorithm you're about to use, I cannot give you a precise answer. In general, it is OK. however, if the imbalance between sizes is considerable, it could bias the results.
Instead of concerning yourself with sizes, it is better to check for overlap and balance of covariates (after matching). Needless to say, ATT and ATE must also align with theories and be justifiable. If all these conditions are met, then I'm afraid there is no issue with sizes.
The treatment group is 333, and the control group is 24437, and I am using the nearest neighbor matching algorithm. So, is it ok to run a t-test and other statistics code to check the significance level of the treatment group on dependent variables?
- Badges
- Science topic
- Similar topics
- Psychology
- Mental Processes
- Learning