To add more fuel to the fire on M&M's version of NQP calculations, what happens to the 100th percentile (observation #20 of 20, in Daren's case...) point on the plot? To reconcile the differences, I presented yet a third method to my students, which they worked out as a minitab lab. (I'm going out on a limb that this method would be valid with the experts, at least in terms of getting across the ideas of an NQP...we divided the area into n regions of equal area, and plotted the "Table A" vs. observed values for the midpoints of each region.)
As they were able to compare their numbers to NSCORES calculations, it stressed the differences in different methods. But it also taught them the ideas behind these types of calculations, my feeling being that the "big picture" concepts for this topic were the important thing for them to understand.
Sarah Latham St. Albans School Washington DC email@example.com