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

Tuesday, July 7, 2015

Firing Black, Experienced Teachers In Camden (and Elsewhere)

A short while ago, the NJ Education Policy Forum published a brief of mine on the school "transformation" plans for Camden, NJ. Doing the sort of analysis I do often requires using a certain wonky, statistical language. But I think it's important to break these things down in a way laypeople can understand them. So let me explain what's happening in Camden, and why anyone who cares about education, social justice, or both should pay attention.

As I explain in the brief, five district schools have been targeted for "transformation" into charter schools run by the big three charter management organizations (CMOs) in Camden: KIPP, Uncommon, and Mastery.

As I've shown in my previous work on Newark, there is very little reason to believe these CMOs will get any better results than the district, all things being equal. That last phrase is the key: unless these CMOs are running schools that have the same student populations, the same resources, and the same student attrition rates, you can't really compare them. The best you can do is use a statistical model to tease out the effects that can be attributed to the schools (understanding that you can, at best, only partially succeed at this task).

When Bruce Baker did his analysis, he found the "great" charters that our media loves to laud aren't really that special after all. Oh, sure, they get better test scores, but only because these charters don't serve the same students, and they have more resources available to them. Which means the district schools are going to have to educate the students the charter won't be serving, and they'll have less money with which to do the job.

Nonetheless, State Superintendent Paymon Rouhanifard is going through with his plan to turn over these five Camden schools to the CMOs. His justification is that these schools are the "most struggling" in the district:
Every Camden student deserves an excellent education, in a safe and modern school near their house. That is why Camden City School District is partnering with non-profit organizations to transform five of our most struggling schools into renaissance schools. Each partner was carefully selected for their proven history of running successful schools, and will give Camden families more choices to access a safe, high-performing neighborhood public school. [emphasis mine]
Now, I saw this happen before in Newark: a state superintendent claims the schools targeted for intervention are the "worst" schools in the district, and therefore she has to take the radical action of firing staff and/or turning schools over to CMOs. But, as I showed, the schools targeted in Newark weren't the lowest-performing in the district; instead, they were the schools more likely to have black students, boys, free lunch-eligible students, and students with special education needs.

To make things worse, Newark's school "renewal" plan had a clear racially disparate impact on black teachers: in other words, Newark's black teachers were more likely to see their schools closed, reconstituted, or turned over to CMOs than Newark's white teachers.

So that led me to ask two questions about CPS's "transformation" plan:

1) Had CPS really identified the "most struggling" schools in Camden?

2) Would there be a racially disparate impact on black teachers, as in Newark?

Here's what I found:

- The Camden schools that the state is turning over to charter management organizations are not the "most struggling" schools in the city. Instead, they are the schools that are serving some of the most needy students in Camden -- and some of them are showing signs of being quite effective.

Let's look at the student populations first. The district schools are in blue, the charters are in green, and the schools slated to be "transformed" are in red. Since we don't have all the information on the so-called "Renaissance schools," which are run by the three big CMOs and only started up this past year, they aren't included in this analysis.

Free lunch is a proxy measure for economic disadvantage. Turns out three of the "transformation" schools have some of the highest FL rates in the city.

And the special education rates are quite high as well. Only one charter school has a classification rate over 12% -- but four of the "transformation" schools do. So this is our first clue something is amiss here: why do the "transformation" schools have relatively higher rates of FL and special education than the other schools?

Now, everyone agrees that schools with lower FL and special education rates have an advantage in test score outcomes. That's why the state developed Student Growth Percentiles (SGPs), which are supposed to account for differences in student populations. They don't, at least not fully, but they are arguably better measures of a school's effectiveness than simply looking at proficiency rates.

I averaged out three years of median SGPS for the schools in Camden to get an idea if the transformation schools really were struggling.

These are the growth scores for English Language Arts (ELA) over three years. Yes, Bonsall, Greenleaf, and Molina look to be struggling -- but not McGraw and East Middle. By this measure,  these schools are actually about average.

What about math?

Molina looks considerably better on this measure, and East Middle is, again, about average. But look at McGraw: it's crushing the mSGPs in math! Why would CPS want to turn over one of its best performers on math score growth measures to a charter operator?

I know the scatterplots drive some of you nuts, but indulge me for a minute, because this next part is important. Here are the mSGP math scores above, plotted against the percentage of students who are eligible for free lunch:

See the relationship? As the percentage of FL students goes up, growth measures go down. In other words, about one-quarter of the variation in growth scores here can be explained by how many FL students a school enrolls. In Camden, as everywhere, schools' test score outcomes are affected by the characteristics of the students they serve.

I don't believe I've fully caught all of the relative advantage or disadvantage a school has, based on its student population, in getting good test scores. But let's still try to account for the differences. The green line above is a "regression line," a kinda-sorta "average" of all the points in the graph. If you're above the line, you're beating prediction, based on how many FL kids you enroll. In the same way, if you're blow the line, you're not beating prediction.

There's no doubt -- Molina, Bonsall, and Whittier aren't "beating the odds." But they aren't alone:

This is a plot of how much each school is over or under prediction on their math scores. By this measure, the transformation schools are hardly the "most struggling" in Camden. Whittier and Bonsall are below prediction, but not nearly as much as several other schools. Molina and East Middle are about average, and McGraw, once again, is one of the top performers in the city.

So, no: the transformation schools are not the most struggling schools, at least by my analysis. So how did CPS come to such different results?

No one's saying.

- Camden's black and experienced teachers are more likely to have an employment consequence under "transformation" than its white and novice teachers.

By "employment consequence," I mean a teacher's school is being turned over to a CMO. According to CPS:

In the case of the school transitions, existing staff would have opportunities to re-apply for jobs but would not be guaranteed positions.” 
Let's start by looking at the staffs for each of Camden's schools. I have access to staffing files that allow me to break down staff experience and demographics; how do schools differ, for example, in the proportion of faculty who are black?

Remember how I showed that McGraw was actually, by at least one standard, a high-performing school? McGraw also has the highest proportion of black staff in the city -- yet, in spite of its relative effectiveness in showing growth on math scores, it's being targeted for "transformation," which means its staff are facing an employment consequence.

You'll notice the charters in Camden right now have small proportions of black staff members. There's every reason to believe this will also be the case at the three big CMOs' schools (I have more work coming out soon on this). Which means we are seeing a disturbing trend: black staff are being replaced by white staff in Camden, and for no apparent reason.

Here's a breakdown of average experience in Camden schools:

The charters hire relatively inexperienced staff; again, we have every reason to believe the large CMOs will follow the same trend.

Let's look across the entire Camden teaching corps, then, to see how being black and/or experienced affects a teacher's chances of suffering an employment consequence under "transformation." What I use is called a logistic regression; I won't get into all the math, but this is a fairly standard statistical technique for the social sciences.

What I'm doing here is saying: "Suppose we've got two teachers, one white, and one black (or Hispanic, or Asian). Suppose they teach at schools with the same SGPs in math and English Language Arts, and schools that have the same percentage of FL students. If we say the white teacher's odds of suffering an employment consequence under 'transformation' are even, how do those odds compare to a black teacher's odds?"

I've done the same thing for experience: holding FL and SGP scores constant, how do the odds for a teacher with 0 to 4 years of experience to have an employment consequence compare to a teacher with 5 to 9 years of experience?

The double stars (**) indicate a high level of statistical significance, which means the odds I give here are very unlikely to be found randomly. Black teachers are 1.64 times more likely to suffer an employment consequence than white teachers. Teachers with between 15 and 24 years of experience are over three times more likely to suffer that consequence compared to teachers with 0 to 4 years of experience.

Put simply: black and experienced teachers are more likely to have to reapply for their jobs under the Camden "transformation" plan than white and inexperienced teachers, even when taking into account their schools' student populations and growth scores.

I find this disturbing for two reasons. First -- it's just wrong. Yes, CPS's obligation, first and foremost, is to its students. But if there's no evidence to support "transformation" -- and I don't see that there is any -- then any plan with this kind of racially disparate impact is simply not justified.

Second, as I say in my brief, there is actually a good bit of evidence that aligning the races of students and teachers is helpful in improving test-based outcomes, particularly for students of color. Certainly, no one thinks black students should only have black teachers, and white students should only have white teachers, and the same for Hispanics and Asians and American Indians and so on.

But it's important for students of color to have teachers of color in their schools. What evidence, then, does CPS have that this plan, which almost certainly will replace experienced teachers of color with inexperienced white teachers, will be good for Camden's students?

One of the results of the reformy mindset -- which only ever wants to view policies in terms of whether or not test scores improve -- is that our debates about education rarely venture into exploring the consequences of policies outside of student test scores. Does anyone, for example, think it's a good idea to drive middle-class and middle-aged teachers of color out of work?

I could maybe see someone making a very strained argument that this is a necessary evil if there was any proof whatsoever that these sort of "transformations" lead to better educational outcomes for students. But there is no proof that this is the case. I looked at the "renewal" of Newark's schools and found that nothing got better; arguably, things actually got worse. Black and experienced faculty appear to have been driven from the renewal schools, and their growth measures went down.

If CPS is going to go ahead with this plan, they are going to have to come up with a much better justification than they have so far. They need to show the community evidence that this plan is worth the steep price that is being paid. They need to explain why the of firing black and experienced staff is necessary for improving Camden's schools.

If they can't, they should halt their plans immediately.

No comments: