Think of what you're saying. If it was so bimodal that the screened people "gained years" while the ones who didn't "gained 0", then how could the average possibly be 4 months?
If we take your theory at face value, on average the screened people would have an 8 month increase.
