I believe this is just a plotting problem. The positive population in the workFlow appears to be the correct one, despite the median values. Carrying-on through many iterations of the plots confirms this. thanks, Aric