Charter Schools: The New Battle of Trenton

So you guys have all seen this:

On Christmas Day, 1775, George Washington and his troops crossed the Delaware River, starting at what is known now as Upper Makefield, PA. The general led his forces south along the banks of the river, eventually sneaking up on the Hessians, the German mercenary troops of the British who were occupying Trenton. The Germans had been partying to celebrate the season, and weren't quite in their best fighting shape when George attacked. It was an important victory.

The reason I know this is that Upper Makefield is my hometown. Every year I would get dragged out by my parents on December 25th to watch our neighbors dress up like Continental Army soldiers and reenact Washington's crossing at the very same spot. 

I always suspected that my parents' friends who volunteered for this duty were celebrating the season before the reenactment in much the same way as the Hessians did. I say this because, year after year, the Durham boats used in the pageant would wind up taking extremely circuitous routes to the New Jersey side of the river, betraying a lack of naval coordination and river-faring acumen that was likely precipitated by the consumption of copious amounts of distilled beverages.

I could be wrong, of course. But I'm not...

There's a way-too-narrow bridge now where the crossing occurred (famously described in Kurt Vonnegut's Slaughterhouse Five) whose pylons make for a few small islands in the otherwise landless expanse between the two shores of the Delaware. In my memory, the few years when the reenactors managed to accurately recreate Emmanuel Leutze's famous painting are offset by the many years when the boats would run aground against these man-made islands as the current swept the boats downriver.

Why am I telling you this? Well, friends, sometimes I have to give you more background than you deserve before I spring my latest, desperate, reformy metaphor on you:

TRENTON — Just as the city’s school board closed a $10.5 million budget gap last month by cutting administrative staff — a gap that some board members blamed in part on charter schools — two more charter applicants submitted proposals to enter the district in 2015. 

The state Department of Education announced last week that it received applications from Rising Star Preparatory Charter School and ASPIRE Academy Charter School to open schools that could eventually attract 900-plus students from Trenton. 

If approved, they would be in addition to four existing charter schools — Foundation Academy Charter School, International Charter School of Trenton; Paul Robeson Charter School for the Humanities and Village Charter School — and two that have been approved to open this fall, STEM-to-Civics and International Academy of Trenton. 

While officials at the charters said that they anticipate the new schools will have a positive effect in the area because they will provide what they believe will be quality options for parents and students, a number of school board members have shown resistance to the growing charter market in Trenton. 

During school board and city council meetings, members of the board and the superintendent of the district, Francisco Duran, have laid much of the blame for the district’s $10.5 million budget deficit at the feet of the city’s charters. 

School Board President Sasa Olessi-Montano has said the charter movement in Trenton has only created chaos for the district’s budget process and for students.

Yes, the charter wars have come to Trenton. Thanks to Former Acting Commissioner of Education Chris Cerf and his former boss, formerly viable presidential candidate Chris Christie, New Jersey has seen a proliferation of charter school applications. Which means the school districts which are forced to host these charters have to come up with more money to fund more of them, often at the expense of their own programs and with no say in their governance.

You'll often hear the charter cheerleaders cry poverty, claiming that they don't get the money they really deserve. Conveniently, they neglect to mention two things: first, the public school districts usually have expenses on their books that can't be shared with charters, so the comparisons of per student spending aren't really relevant. Related to this are issues of economies of scale: for example, if you pull one hundred kids out of a public school district, you still have to light and heat the buildings where they would have gone to school.

Second -- and this is the one that the mandarins of the local press keep ignoring -- charter schools do not serve the same students as the public schools.

I'm guessing, between Bruce Baker and myself, we've now got over two dozen variations on this same graph, featuring schools and districts from all over the state... but sure, let's add one more. Here are the traditional public schools (TPSs) and charters in the Trenton area; the charters are in red. Once again, the charters do not serve the same population of free lunch-eligible students (a proxy measure for poverty) as most of the Trenton public schools.

How about special education students?

Still on the low end. Both of these comparisons need a few caveats: first, I'm mixing grade levels, which is tricky for both free lunch-eligibility and special education percentages, which can have a tendency to shift as grade levels move from elementary to secondary. Second, those special education percentages lump together kids who have mild needs (like speech therapy) with those who have severe impairments (like low-functioning autism or emotional disabilities). The tendency is for New Jersey's charters to take fewer (if any) of the children with more serious, more costly needs, but I don't know if that is the case in Trenton (more later, hopefully).

Finally: having a free lunch-eligible percentage of over 60 means having a lot of kids in poverty, even if the public schools have more. To illustrate, let's pull things back a little and look at all of Mercer County, the home of Trenton and quite possibly one of the most economically unequal areas in the entire United States.

I know some of you let your eyes glaze over when you see these scatterplots, but they really are good at getting to the heart of the matter. Let me break it down:

The vertical axis shows the proportion of students who passed the cutoff for "proficiency" on the Grade 8 NJASK language arts test in 2013. Yes, all of Princeton Charter School's eighth graders passed the test, as did the vast majority in Princeton, East Windsor, Robbinsville, and the other "outer" school districts (more on that in a sec). Of the three Trenton area charters for which we have data, only Foundation Academy had a proficiency rate over 70 percent.

So, can we replicate Foundation on a larger scale? Clearly, the answer is "no," because Foundation (and Village) have a substantially lower percentage of students who qualify for free lunch than the public schools of Trenton. We just can't make every school in Trenton have a lower-than-average poverty rate. 

As I've pointed out many times on this blog, peer effect is real and undoubtedly contributes substantially to Foundation's success. That's not say the school isn't doing good work and may have developed some best practices that are worth emulating; I'm only stating that there is very little evidence here that Foundation can be scaled up across all of Trenton.

Some of you might look at the variation in Trenton's public schools and see that while they all have the highest poverty rates in Mercer County, they also vary widely in their outcomes. So why can't all of Trenton's schools score at least as well as Village? Well, there are likely differences between these schools' populations that can't be reflected in the data, but wind up influencing test scores. And I'm not saying school differences don't matter; they almost certainly do, although we'd have to study this situation a lot more closely to figure out what those differences are (falling back on reformy bromides like "no excuses!" or "more class time" is not helpful in this regard).

What I'm saying is this: even if every Trenton school could emulate the results of its best schools with equivalent rates of poverty, those schools still wouldn't "close the gap" with low-poverty schools in the same county.

Which brings me to my final point (and back to my tortured metaphor):

Growing up across the river, I know Mercer County pretty well. You can think of it as three concentric semicircular rings growing out from the intersection of Route 1 and the Delaware. The first ring is Trenton itself. The second is from roughly the city line to the loop made made I-95/295: the "inner" suburbs of Ewing, Hamilton, and Lawrence. The final ring extends to the county border: Hopewell, Princeton, the Windsors, Robbinsville, etc.

This is a pattern you see in cities of all sizes all over the country: a poor urban core (with maybe some gentrifying), surrounded by working-class inner suburbs, surrounded by relatively wealthy exurbs. If you look at the scatter plot above, you'll see the free-lunch eligible percentage slide closer to zero the further out you go from the core of Trenton. And you'll see the proficiency rate slide closer to 100 percent as that free-lunch rate decreases.

This is the most obvious thing in the world: standardized test outcomes correlate very tightly with economic measures. I know all the reformy types love to claim this is "making excuses," but their protestations don't change the fact that this correlation exists, and that there is little debate the cause is poverty's effect on test-based outcomes. In Mercer County, almost 90 percent of the variation in 8th Grade proficiency rates can be explained by free-lunch eligibility rates. As the poverty rate goes down, the proficiency rate goes up, with a nearly perfect relationship. Only the willingly obtuse or the outright mendacious would ever try to deny this truth or what it is telling us: poverty matters.

But Mercer County shows us something else:

Look at the gap in free-lunch eligibility between the highest inner suburban schools and the lowest Trenton public schools: it's over 50 percentage points. Like the cold water that runs between the shores of the Delaware, this poverty gap between the 'burbs and the city only has a few small islands in between: Trenton's charter schools. The only schools in Trenton that have free-lunch rates low enough to even approach the suburban schools are charters.

What are we to make of this? Well, if we start with the acknowledgment that charter schools are privately managed and do not have to grant students, those students' families, or their teachers the same rights and commitments to transparency as public schools, we have to conclude that charters are offering an exchange: students in poor areas can gain a peer effect if their families are willing to give up the governance structure of public schools.

You'll notice that there is no such deal being offered in the outer exurbs: those kids get to enjoy their peer effects while their families get self-determination and democratic representation in the running of their school system. You'll also notice that even the most economically segregated school in Trenton can only compete with the inner, working-class suburbs. The "gap" in proficiency with the outer exurbs still exists. Some may say this is a triumph, but we have to be honest about its cost, and we have to be honest about its cause.

Because Foundation hasn't assembled its student population by mixing kids from the 'burbs with kids from Trenton; rather, it's culled a number of less-poor children with fewer special needs from Trenton's public schools and concentrated them. This is just not a scalable formula for achieving educational equity. There is no evidence here that we will be able to close the "achievement gap" by expanding charter schools when those schools have different populations than the public schools in poor urban areas.

One other thing: remember that lone charter in the upper left corner, with 100 percent proficiency and the smallest percentage of free lunch-eligible kids in the county? You can make a very good case that Princeton Charter School is, for all intents and purposes, a private school that is being subsidized by the taxpayers of Mercer County (Bruce Baker has, in fact, made just that case).

But imagine, instead, if PCS made a commitment to living in the Delaware Poverty Gap. Imagine a school -- maybe a network of schools -- that drew students from Trenton and the outer 'burbs, allowing students to enjoy peer effects without simply concentrating the less-poor Trenton students in a few stand alone schools. Imagine a series of schools that mixed affluent and less-affluent kids through Mercer County. What might happen?

I'd still argue that there was no need for a charter school structure to create schools like this. But there would at least be the justification that those charters had a possibility of being scalable. I don't see any way that continuing to expand charters in Trenton, populated only with Trenton's children, can ever get the district to a point where a significant number of students could be served in schools with low enough populations of less-affluent students such that they could really close the "achievement gap."

In other words: we can keep opening charters, or we can start trying to develop solutions that address the real problem. What's it going to be?

ADDING: Inter-district choice may be a good first step, but I still think it suffers from the same problem: it's just not scalable. I'll try to get to this at some point in the future.

ADDING MORE: Even more typos than usual. Thanks, Giuseppe.