The point about the scatterplots is valid, but if he thinks a correlation of .26 (same teacher, subject and year, but different grade classroom) is sufficient to justify VAM, which arguably has no validity let alone reliability, then it's impossible to take him seriously.