(This might include working a baseline regression for equally sets of data individually, then producing comparisons of baseline-predicted consumption with true consumption for both of those sets of data individually, then combining the results together at the end. This article isn't going to rev