(Guest Post by Matthew Ladner)
When Doug Harris took to the pages of the New York Times to denounce Michigan charter schools as “the biggest school reform disaster in the country” he might have profited from an honest reading of the research literature and 10 minutes looking at NAEP data. For example, it took about 10 minutes to calculate the above chart. Michigan charters must be the fastest improving “disaster” yet seen.
Harris went on in the piece not just to trash Michigan, but to boast of the wonders of New Orleans charters. Harris used New Orleans to contrast as the Happy Hunting Grounds of charter schooling in contrast to the Detroit hellscape:
The New Orleans results have been impressive. In the decade after the reforms, the city’s standardized test scores have increased by eight to 15 percentile points and moved the district from the bottom to almost the state average on many measures. High school graduation and college entry rates also seem to have improved significantly, even while suspensions, expulsions and the rate of students switching schools have all dropped. Detroit and New Orleans represent radically different versions of school choice — and the one that seems to work is the one that uses the state oversight that Ms. DeVos opposes.
Well that sounds awfully impressive, but it perhaps less so when you check the only common metric testing data available in both Michigan and Louisiana charters-NAEP. When you plot the cohort gains of Louisiana and Michigan charter schools against state averages, it looks something like:
NAEP cohort gain calculations have their limitations, but the reader should note that similar findings to these were found in research on trends in state test scores. Perhaps someone will be kind enough to provide a link to the Credo dot chart showing nearly indistinguishable growth performance between Detroit and New Orleans in the comment section, but it was reminiscent of this chart as I recall.
Ironically, Harris also denounced Detroit charters as a “Wild West.” Just for the record- here is what a real Wild West charter sector’s results looked like between 2011 and 2015:
Jay P. Greene's Blog 
I have to say, your use of highly . . . interpretive data point labels is something I’d like to see become widely adopted.
Any chance the size of the data points could be reduced? The huge blue marbles reduce both precision and readability.
Those are reduced- down to 50 from 100. I will try out a 25 next time.