As far as I can tell, the author breaks the universe of teachers into 5% groups and then proceeds to show that the percentile groups are ordinally stable betwee

As far as I can tell, the author breaks the universe of teachers into 5% groups and then proceeds to show that the percentile groups are ordinally stable between these two years. In other words, the top scoring group in year 1 beat all other groups in year 2, and so forth, with perhaps some slight shuffling in the absolute middle of the distribution [hard to see on the graph]. This is supposed to be evidence that the correlation is weak??