I will protect your pensions. Nothing about your pension is going to change when I am governor. - Chris Christie, "An Open Letter to the Teachers of NJ" October, 2009

Saturday, March 17, 2018

The Facts About NJ Charter Schools, Part II: Segregation By Special Education Need

In this series of posts, I'm breaking down a new report by myself and Julia Sass Rubin on New Jersey's charter schools. State data shows one incontrovertible truth:

New Jersey's charter schools enroll far fewer students proportionally who have learning disabilities, or who are Limited English Proficient, when compared to their hosting districts.

Here's a graph that shows this quite clearly:

Oh, sorry -- this graph isn't from our research. This graph is from a report published by the New Jersey Charter Schools Association, the state's biggest charter advocacy group.

Let me clean it up a bit for you...

It is, of course, completely inappropriate to use the same scale for measures that are as different as racial composition and special education classification. I would make my grad students resubmit their work if they ever tried to pull a stunt like this.

Still, you can clearly see that, according to the state's biggest charter cheerleaders, NJ charter schools enroll far fewer students proportionally who are classified with a learning disability, or who are English Language Learners.

Let's look at this in a more appropriate way. This graph is from our report (for real this time):


In our report, we compare all of the charter school students residing in a school district to the resident students who attend the public district schools. This method allows us to compare a community's charter students to its district students -- no matter where the charter students attend school. (I'll discuss this method in more detail later in this series.)

These findings are beyond question -- and they raise some serious issues. Even Chris Christie acknowledged that it costs more to educate a child with a learning disability; this particular fiscal burden falls hard on public district schools when charterization concentrates their proportion of classified children. It also makes comparisons between the academic outcomes of charters and district schools meaningless unless this disparity is accounted for.

The problem with most attempts to do this -- like the NJ CREDO study, which was commissioned by the state -- is that the statistical models employed use data wholly inadequate to the task. These data divide students into two groups: those with a learning disability, and those without. The problem is that classified students can have very different disabilities, and, consequently, very different educational needs.

As I've noted before, some disabilities, such as speech or "specific learning disabilities" (SLDs), are relatively low-cost. Others have a much higher cost. Guess which students are more likely to enroll in the charters?



The special needs students who are enrolled in NJ charters tend to have lower-cost disabilities the those in district schools. This analysis differs somewhat from above (see the report for details), but it matches our previous work. We're using 2016 data here; in that year, the state did not suppress data as they have done before.*

For a long time, charter cheerleaders have claimed -- with no empirical evidence -- that the reason their special education rates are lower is because their superior instruction and organization make special education classification necessary. The chart above directly refutes this. It's much less difficult to change the classification of a student with a speech or SLD disability than one with a traumatic brain injury, or blindness, or autism. If charters dissolve classified students' individualized education programs (IEPs) at higher rates than public district schools, it's only because the classified charter students have, on average, less profound disabilities than district students.

I've heard some make the case that school districts often place special needs students in specialized, out-of-district private schools, and that this is functionally no different than allowing students to enroll in charters. But that's a ridiculous argument on its face. When a school board makes a decision about an out-of-district placement, they make the decision, and they figure out how to pay for it. Charter school enrollments, on the other hand, are foisted upon school districts by the state with no ability for the district to approve or regulate the enrollment.

In other words: the state makes the decision to approve a charter school, but the district has to pay for it. Worse, if the district isn't where the charter is located, they don't even have the right to appeal the decision. If students in your town want to enroll in a charter school 20 miles away, you don't get any say in the matter -- your town has to pay for it, no matter the fiscal or educational harm.

And again: those students who enroll are less likely to have special education needs... most of the time:



This table shows the disparity between the charter population and the district population in the proportion of classified students for each population.** In North Plainfield, for example, 18.5 percent of the district's students are classified -- but none of the resident students who attend charters are listed as having a special education need. That disparity is the largest in the state.

But here's what's interesting: there are, in fact, districts where the charter and district student populations have similar proportions of special needs students. In fact, in New Brunswick, more classified students attend the charters, proportionally, than the public district schools. Keep in mind that, as we show above, the charter students in New Brunswick have less costly disabilities. This is a problem because the charter school funding formula treats all classified students, with the exception of students with a speech disability, the same in terms of the funds transferred to charters.

Still, New Brunswick shows that many of the other local charter school sectors could be enrolling more special needs students. So why don't they? Why are so many charter schools not stepping up and enrolling more special needs students -- even those with the least costly learning disabilities?

The charter sector has been promising for some time that it will start educating more children with special needs. Some charters do -- but many clearly do not. And why would they, when the state has refused for years to hold them to account? Why would they, when they could count on renewals and approvals for expansions even though it was obvious they were engaging in segregation by special education need?

During the Christie administration, the state turned a blind eye toward the segregation by special need that accompanies charter school expansion. Yet the data on this are so clear that not even the NJCSA doesn't dispute the truth. The Murphy administration, the NJ Legislature, and the NJDOE have got to start acknowledging this and come up with a plan to address it.

There's another student population NJ's charters have underserved, even more than special needs children: English language learners. We'll discuss that next.


* The data was suppressed in 2015 but not in 2014. I have no idea why. You can tell the data is not suppressed because there are many cells that have values between 1 and 9, even though in other years the cells were suppressed when less than 10.

** In the report, we limit the districts studied to those enrolling at least 50 students.

9 comments:

Stephen said...

As you likely know, when there was a more thoroughgoing analysis of classification practices and of student academic success, it appeared that the studied charter schools (here in Boston) were more substantially and effectively providing service to kids who would have been classified in traditional districts schools as having special needs than was apparent from a look at just the raw enrollment figures. Below are extracts from this article:
https://www.brookings.edu/research/the-importance-of-high-quality-general-education-for-students-in-special-education/

"A new study by one of us (Elizabeth Setren) follows students participating in special education at the time they apply to charter elementary, middle, or high schools in Boston. By comparing students who randomly receive offers to attend a charter school to those that randomly do not receive lottery offers, the study estimates the causal impact of enrolling in a charter school for both special education and general education students.

"The findings show that high quality general education can lead to remarkable progress for special education students. Attending a Boston charter school makes special education students 1.4 times more likely to score proficient or higher on their standardized tests, resulting in a 30 percent reduction of the special education achievement gap.

"Gains extend to beyond just test scores: Boston charter enrollment boosts special education students’ college preparation. These students are over 1.6 times more likely to meet a key graduation requirement, over three times more likely to be eligible for a state merit scholarship, and over 3.8 times more likely to take at least one AP exam in a charter school compared to their peers who do not receive charter lottery offers.

"Interestingly, special education students are less likely to receive special education services in Boston charters than their counterparts who enroll in the Boston Public Schools (BPS). Special education students who apply for a charter school lottery are two times less likely to keep their IEP and three times more likely to move to a more inclusive classroom setting if they randomly receive an offer to attend a charter.[4]

[...]

"The findings are robust to three common criticisms of charter schools: that high needs students do not enroll, that charters do not serve high needs students well, and that charters encourage lower performing students to leave. In recent years, the charter lottery special education applicants are representative of Boston Public Schools’ special education students in terms of level of services, disability type, and academic abilities. The enrollment numbers do not fully reflect this parity because charters are more likely to remove special education classifications and increase classroom inclusion. Achievement gains accrue even to the most disadvantaged students: students who had been in a substantially separate classroom in BPS and students who had the lowest test scores among special education students."

Duke said...

Stephen, I know Setren's piece very well, because Julie Mead and I reviewed it for NEPC:

http://nepc.colorado.edu/thinktank/review-charter-expansion

"What is most important to understand, however, is that any effects from the SEII study can only be generalized to the population of students who actually enter the lottery. The report only acknowledges this limitation in its very last paragraph: “It is worth noting that the results apply to charter lottery applicants.... [M]y estimates may not reflect the effects of expanding the number of seats in Boston’s charter sector or requiring charters to recruit more special needs students” (p. 19)."

This is a large limitation on Steren's work, especially since it's certain the population of special needs students applying to the Boston lotteries is NOT equivalent to the general special needs population.

Duke said...

In any case, I don't understand what MA data has to do with NJ data.

Stephen said...

You write: "it's certain the population of special needs students applying to the Boston lotteries is NOT equivalent to the general special needs population."

While "In recent years, the charter lottery special education applicants are representative of Boston Public Schools’ special education students in terms of level of services, disability type, and academic abilities," you do have a point that there is an additional group of special education students with exceptionally severe needs that are served in facilities that may be out of district rather than part of the traditional public school system. But your reference to school boards' need to "figure out how to pay for it" omits the fact that expenses for those students, at least here in Massachusetts, are omitted from the calculations that determine funding levels for charter schools, and that the school boards have access to resources such as these:

"Districts also receive a federal IDEA entitlement grant, which can be used to pay for special education services. In FY09, we will be distributing approximately $240 million under this program. The amount which each district receives is based on a complicated formula that includes a base amount (tied to the district's FY99 IDEA allocation) plus additional amounts reflecting the total number of students and the number of low income students in the district.

"The municipal Medicaid (MuniMed) program provides reimbursement for certain medical services provided by school districts to eligible students, many of whom are also receiving special education services. MuniMed reimbursements exceed $100 million annually. These are treated as general fund revenues by either the municipality (in the case of municipal school districts) or the regional school district. Although not required to do so, many municipalities appropriate all or part of their MuniMed reimbursements back to their local school district to help defray future special education costs.

"Circuit Breaker Program
"The state special education reimbursement program, commonly known as the circuit breaker program, was started in FY04 to provide additional state funding to districts for high-cost special education students. In FY08 approximately $200 million was paid to districts under this program.
[...]
"For children placed in a school district outside their home town by the Department of Transitional Assistance or the Department of Children and Families, and for children who have no parents or guardians in the commonwealth, the reimbursement is 100 percent above the threshold rather than 75 percent."
http://www.doe.mass.edu/finance/circuitbreaker/finance.html

Stephen said...

Jazzman: "In any case, I don't understand what MA data has to do with NJ data."

The Massachusetts research exposes the fact that a look at raw ongoing enrollment data can be severely misleading in respect to both special education and ELL status. Here in Massachusetts one is not supposed to classify a student as in need of special services and an IEP if he or she is making adequate academic progress in a regular classroom. And one is supposed to stop classifying a student as ELL once he or she becomes proficient in English.

Imagine a dozen kids entering a charter school classified as needing special education and another dozen classified as ELL. And within a few years all of the first dozen are making good progress in regular classrooms and the second dozen are all fluent in English. Zero SPED and zero ELL. While their identical twins in TPS are 24 SPED or ELL students.

I suppose one could then look at the enrollment figures and slam the charter schools for insufficiently serving SPED and ELL students. But it mightn't persuade all comers. In respect to NJ, I remain open-minded, neither assuming the situation is directly similar to Massachusetts, nor persuaded that your critique is well-founded.

Duke said...

Stephen, you accuse me of being "severely misleading" with no empirical evidence to back up your contention. All you ask me to do is "imagine" a hypothetical you have no evidence to support.

You also commit a logical fallacy: if charter students are, indeed, declassified "within a few years," there would be at least some students classified - not "zero."

Like your arguments against the evidence I've presented for cohort attrition in charters, you assume the least likely explanation on the basis of no evidence.

As I've shown previously, the fact that NJ charters spend far less on student support services, and have far fewer staff working in those service areas per pupil, backs up the simplest explanation: the charters don't enroll as many special needs students.

You can continue to argue this point repeatedly, but it doesn't change the evidence.

Stephen said...

Jazzman: "Stephen, you accuse me of being 'severely misleading' with no empirical evidence to back up your contention. "

To be more precise, I had stated that "a look at raw ongoing enrollment data can be severely misleading". Do you really dispute that? I think my extreme hypothetical illustrates how it can be misleading, and Setren's empirical evidence that I alluded to elucidates the fact that indeed in reality it sometimes is.

Jazzman: "You also commit a logical fallacy: if charter students are, indeed, declassified 'a few years,' there would be at least some students classified - not 'zero.'

The hyothetical alluded to a specific set of individuals eventually attaining zero SPED/ELL status.

Jazzman: "Like your arguments against the evidence I've presented for cohort attrition in charters, you assume the least likely explanation on the basis of no evidence. "

No evidence? Here, for example:
http://jerseyjazzman.blogspot.com/2016/10/charter-school-attrition-in-ma-reader.html
I provided conclusive evidence to demonstrate that comparing 12th grade enrollment to 9th grade enrollment three years earlier can be severely misleading:

"The 9th grade class enrollments you examine merge students who are newly starting out and others who are repeating the grade. That distorts any attempt to compare 9th grade and 12th grade enrollment figures to measure attrition.

Massachusetts grade-level retention reports can be found here:
http://www.doe.mass.edu/infoservices/reports/retention/

"Looking at Appendix B for 2015-2016, one finds for City on a hill the following grade-level retention rates:
9th 10th 11th 12th
27.8 3.3 3.2 7.4 "

Jazzman: "As I've shown previously, the fact that NJ charters spend far less on student support services, and have far fewer staff working in those service areas per pupil, backs up the simplest explanation: the charters don't enroll as many special needs students."

A properly cautious analysis would describe an adequately complete range of possible contributing factors for differing enrollment figures between schools, and hesitate to jump to conclusions, simple or otherwise.

Jazzman: "You can continue to argue this point repeatedly, but it doesn't change the evidence."

And perhaps we can agree that the evidence available to us is inadequate for firm, simple conclusion?

Duke said...

Stephen, I've given clear evidence based on universally reported data. You've put up in its place an "extreme hypothetical" (your own words.

There is no question that NJ charter schools have a far lower classification rate than comparable public district schools. There is no question the classified students charters do enroll tend to have less costly needs. Arguing around this is pointless

You seem to value having the last word. Take it. And consider getting your own blog.

Stephen said...

Thanks for the opportunity to discuss this with you, JJ.

In respect to your forthcoming piece re: LEP rates, will available NJ data allow you to compare durability of LEP status in the different sectors? And might you be able to construct graphs depicting the LEP/FLNE ratios at TPS and charter schools?