Thanks for catching this, that's even more straight-forward (and arguably less informative, but the best solution would be to have both plots of raw data and ma

Thanks for catching this, that's even more straight-forward (and arguably less informative, but the best solution would be to have both plots of raw data and marginal effects of course).