Mike-

So I have figured out how to do a significance test on the new NAEP data explorer, and have poked around a bit. It looks like it is rare for a two year difference in scores to hit statistical significance at the state level, with it being much easier for smaller trends to pass significance tests at the national level. So for instance a national two point decline was statistically significant at the national level in 2015, but three point gains at the state level did not make standard significance tests.

State level results can and do however get significant over longer time periods. Usually this takes the form of multiple incremental gains that add up to statistical significance over time. I will be happy if we see a lot of this in AZ, IN and NC.

]]>Question – can you characterize the “meaningfulness” of the NAEP gain sizes? Perhaps as SD or something?

]]>