Neither here nor there but isn’t Dunning Kruger mostly a statistical artefact? Like the bottom quartile on x will have less extreme values on y unless x and y are highly correlated.