If you're overinterpretting a drifting baseline, that would mean that the results you're getting are even MORE unusually aligned, not LESS. The drifting baseline should dampen your denominators, resulting in a lower scaling factor for each current period. The higher the scaling factor, the more the means of the distributions should shift apart, which lowers of the probability of distribution overlap (where the order between two states would shift).
But what if everything else was normal and you introduced a too low or too high baseline randomly across the states. Some would have a constant negative excess and some a constant positive.
I truly find this fascinating and am appreciative of your endeavours to investigate and understand the implications and consequences of the vaccination programme but alas my Stats ability simply isn’t good enough to really understand your research, other than in the broadest possible terms. However I hope you will persevere and discover whatever the facts may be.
If you're overinterpretting a drifting baseline, that would mean that the results you're getting are even MORE unusually aligned, not LESS. The drifting baseline should dampen your denominators, resulting in a lower scaling factor for each current period. The higher the scaling factor, the more the means of the distributions should shift apart, which lowers of the probability of distribution overlap (where the order between two states would shift).
But what if everything else was normal and you introduced a too low or too high baseline randomly across the states. Some would have a constant negative excess and some a constant positive.
I truly find this fascinating and am appreciative of your endeavours to investigate and understand the implications and consequences of the vaccination programme but alas my Stats ability simply isn’t good enough to really understand your research, other than in the broadest possible terms. However I hope you will persevere and discover whatever the facts may be.