Michigan Charter Schools a Smashing Success

Michigan Capitol Confidential ^ | 1/22/2013 | Michael Van Beek

[MichCapCon]

The new study by Stanford University’s Center for Research on Education Outcomes is the most thorough analysis ever done of Michigan’s public charter schools, and the results are almost entirely positive.

Of the 56 different outcomes tested, CREDO found 52 positive results for public charter schools and their students when compared to conventional public schools.

First for its rigor: The study matched individual charter school students to students attending the conventional schools those charter school students would have attended. Among other variables, students were paired based on grade level, gender, race, income status and prior test scores. The study then compared the annual gains in reading and math test scores between these charter school students and their “virtual twins” in conventional public schools.

This was done for 85,650 students from 273 different charter schools over a five-year period. No other study to date has used this level of detail or this amount of data to analyze school performance in Michigan, with the second-best settling for comparing school-level averages.

The positive results hold for nearly every category of school and student CREDO analyzed.

For instance, it didn't matter if a charter school was located in an urban, suburban, town or rural area — students from each of these geographical categories averaged better learning gains than their peers in conventional schools. Rural charter school students did the best, making gains equivalent to more than three months of learning compared to their non-charter counterparts.

The trend held across school type as well. Charter schools students in elementary, middle and K-12 (multi-level) schools all outperformed conventional school students. Elementary school students, making up 72 percent of all charter school enrollees, averaged the greatest learning gains of these groups.

Another encouraging finding was that average learning growth for charter school students increased the longer those students were enrolled in the charter school of their choice. Students start demonstrating superior average learning gains after just one year in a charter school, but those gains grow over time. By their fifth year, students in charter schools averaged annual learning gains equivalent to about three and half months of learning (and about a full year’s worth of learning gains over the entire period).

The positive learning gains for charter school students were consistent across the achievement spectrum as well. Both low-achieving and high-achieving students in charter schools had better learning gains in reading and math compared to their counterparts in conventional schools.

In the study’s own words: "[Charter schools] move low-performing students ahead with larger gains, and preserve more of the learning for high-performing students."

Due to its scope and rigor, this new Stanford study is the most definitive analysis of Michigan public charter schools to date. Its methodology isn’t perfect, but it's the best that's ever been done, and it clearly shows that overall, charter schools in Michigan are an enormous success.

[cripplecreek]

Freedom to learn

If anyone wants to be added to the Michigan Cap Con ping list, let me know.

[buwaya]

I dont doubt that this is a real result of charter vs public.

However, the critics will reply that this study design does not eliminate the self-selection problem, in that the parents of the charter school kids selected the charter, and that the act of selection indicates that they were different in some unquantified or unquantifiable factors than the average of those who persisted in the public schools.

This does not mean that the kids in question would have done as well in the public schools, the prior test metric controls for that to a degree, but it does leave open the question of whether the remaining public school students would have benefited from charter schools.

A more rubust test model would have used a lottery system, where only a proportion of charter school applicants are accepted, and their subsequent performance is compared to their fellow-applicants. This eliminates the self-selection problem. There have been several cases like this, unfortunately with rather small sample sizes.