An Education Hearing I'd Like To See

At the end of February, the District of Columbia Council’s Education Committee held its annual hearing on the performance of the District’s Public Schools (DCPS). The hearing (full video is available here) lasted over four hours, and included discussion on a variety of topics, but there was, inevitably, a block of time devoted to the discussion of DCPS testing results (and these questions were the focus of the news coverage).

These exchanges between Council members and DCPS Chancellor Kaya Henderson focused particularly on the low-stakes Trial Urban District Assessment (TUDA).* Though it was all very constructive and not even remotely hostile, it’s fair to say that Ms. Henderson was grilled quite a bit (as is often the case at these kinds of hearings). Unfortunately, the arguments from both sides of the dais were fraught with the typical misinterpretations of TUDA, and I could not get past how tragic it is to see legislators question the superintendent of a large urban school district based on a misinterpretation of what the data mean - and to hear that superintendent respond based on the same flawed premises.

But what I really kept thinking -- as I have before in similar contexts -- was how effective Chancellor Henderson could have been in answering the Council’s questions had she chosen to interpret the data properly (and I still hold out hope that this will become the norm some day). So, let’s take a quick look at a few major arguments that were raised during the hearing, and how they might have been answered.

When Growth Isn't Really Growth, Part Two

Last year, we published a post that included a very simple graphical illustration of what changes in cross-sectional proficiency rates or scores actually tell us about schools’ test-based effectiveness (basically nothing).

In reality, year-to-year changes in cross-sectional average rates or scores may reflect "real" improvement, at least to some degree, but, especially when measured at the school- or grade-level, they tend to be mostly error/imprecision (e.g., changes in the composition of the samples taking the test, measurement error and serious issues with converting scores to rates using cutpoints). This is why changes in scores often conflict with more rigorous indicators that employ longitudinal data.

In the aforementioned post, however, I wanted to show what the changes meant even if most of these issues disappeared magically. In this one, I would like to extend this very simple illustration, as doing so will hopefully help shed a bit more light on the common (though mistaken) assumption that effective schools or policies should generate perpetual rate/score increases.

When Checking Under The Hood Of Overall Test Score Increases, Use Multiple Tools

When looking at changes in testing results between years, many people are (justifiably) interested in comparing those changes for different student subgroups, such as those defined by race/ethnicity or income (subsidized lunch eligibility). The basic idea is to see whether increases are shared between traditionally advantaged and disadvantaged groups (and, often, to monitor achievement gaps).

Sometimes, people take this a step further by using the subgroup breakdowns as a crude check on whether cross-sectional score changes are due to changes in the sample of students taking the test. The logic is as follows: If the increases are found when comparing advantaged and more disadvantaged cohorts, then an overall increase cannot be attributed to a change in the backgrounds of students taking the test, as the subgroups exhibited the same pattern. (For reasons discussed here many times before, this is a severely limited approach.)

Whether testing data are cross-sectional or longitudinal, these subgroup breakdowns are certainly important and necessary, but it's wise to keep in mind that standard variables, such as eligibility for free and reduced-price lunches (FRL), are imperfect proxies for student background (actually, FRL rates aren't even such a great proxy for income). In fact, one might reach different conclusions depending on which variables are chosen. To illustrate this, let’s take a look at results from the Trial Urban District Assessment (TUDA) for the District of Columbia Public Schools between 2011 and 2013, in which there was a large overall score change that received a great deal of media attention, and break the changes down by different characteristics.

Select Your Conclusions, Apply Data

The recent release of the National Assessment of Educational Progress (NAEP) and the companion Trial Urban District Assessment (TUDA) was predictably exploited by advocates to argue for their policy preferences. This is a blatant misuse of the data for many reasons that I have discussed here many times before, and I will not repeat them.

I do, however, want to very quickly illustrate the emptiness of this pseudo-empirical approach – finding cross-sectional cohort increases in states/districts that have recently acted policies you support, and then using the increases as evidence that the policies “work." For example, the recent TUDA results for the District of Columbia Public Schools (DCPS), where scores increased in all four grade/subject combinations, were immediately seized upon supporters of the reforms that have been enacted by DCPS as clear-cut evidence of the policy triumph. The celebrators included the usual advocates, but also DCPS Chancellor Kaya Henderson and the U.S. Secretary of Education Arne Duncan (there was even a brief mention by President Obama in his State of The Union speech).

My immediate reaction to this bad evidence was simple (though perhaps slightly juvenile) – find a district that had similar results under a different policy environment. It was, as usual, pretty easy: Los Angeles Unified School District (LAUSD).

Is Selective Admission A School Improvement Plan?

The Washington Post reports that parents and alumni of D.C.’s Dunbar High School have quietly been putting together a proposal to revitalize what the article calls "one of the District's worst performing schools."

Those behind the proposal are not ready to speak about it publicly, and details are still very thin, but the Post article reports that it calls for greater flexibility in hiring, spending and other core policies. Moreover, the core of the plan – or at least its most drastic element - is to make Dunbar a selective high school, to which students must apply and be accepted, presumably based on testing results and other performance indicators (the story characterizes the proposal as a whole with the term “autonomy”). I will offer no opinion as to whether this conversion, if it is indeed submitted to the District for consideration, is a good idea. That will be up to administrators, teachers, parents, and other stakeholders.

I am, however, a bit struck by two interrelated aspects of this story. The first is the unquestioned characterization of Dunbar as a “low performing” or “struggling” school. This fateful label appears to be based mostly on the school’s proficiency rates, which are indeed dismally low – 20 percent in math and 29 percent in reading.

Being Kevin Huffman

In a post earlier this week, I noted how several state and local education leaders, advocates and especially the editorial boards of major newspapers used the results of the recently-released NAEP results inappropriately – i.e., to argue that recent reforms in states such as Tennessee and D.C. are “working." I also discussed how this illustrates a larger phenomenon in which many people seem to expect education policies to generate immediate, measurable results in terms of aggregate student test scores, which I argued is both unrealistic and dangerous.

Mike G. from Boston, a friend whose comments I always appreciate, agrees with me, but asks a question that I think gets to the pragmatic heart of the matter. He wonders whether individuals in high-level education positions have any alternative. For instance, Mike asks, what would I suggest to Kevin Huffman, who is the head of Tennessee’s education department? Insofar as Huffman’s opponents “would use any data…to bash him if it’s trending down," would I advise him to forego using the data in his favor when they show improvement?*

I have never held any important high-level leadership positions. My political experience and skills are (and I’m being charitable here) underdeveloped, and I have no doubt many more seasoned folks in education would disagree with me. But my answer is: Yes, I would advise him to forego using the data in this manner. Here’s why.

ESEA Waivers And The Perpetuation Of Poor Educational Measurement

Some of the best research out there is a product not of sophisticated statistical methods or complex research designs, but rather of painstaking manual data collection. A good example is a recent paper by Morgan Polikoff, Andrew McEachin, Stephani Wrabel and Matthew Duque, which was published in the latest issue of the journal Educational Researcher.

Polikoff and his colleagues performed a task that makes most of the rest of us cringe: They read and coded every one of the over 40 state applications for ESEA flexibility, or “waivers." The end product is a simple but highly useful presentation of the measures states are using to identify “priority” (low-performing) and “focus” (schools "contributing to achievement gaps") schools. The results are disturbing to anyone who believes that strong measurement should guide educational decisions.

There's plenty of great data and discussion in the paper, but consider just one central finding: How states are identifying priority (i.e., lowest-performing) schools at the elementary level (the measures are of course a bit different for secondary schools).

Are There Low Performing Schools With High Performing Students?

I write often (probably too often) about the difference between measures of school performance and student performance, usually in the context of school rating systems. The basic idea is that schools cannot control the students they serve, and so absolute performance measures, such as proficiency rates, are telling you more about the students a school or district serves than how effective it is in improving outcomes (which is better-captured by growth-oriented indicators).

Recently, I was asked a simple question: Can a school with very high absolute performance levels ever actually be considered a “bad school?"

This is a good question.

Underlying Issues In The DC Test Score Controversy

In the Washington Post, Emma Brown reports on a behind the scenes decision about how to score last year’s new, more difficult tests in the District of Columbia Public Schools (DCPS) and the District's charter schools.

To make a long story short, the choice faced by the Office of the State Superintendent of Education, or OSSE, which oversees testing in the District, was about how to convert test scores into proficiency rates. The first option, put simply, was to convert them such that the proficiency bar was more “aligned” with the Common Core, thus resulting in lower aggregate proficiency rates in math, compared with last year’s (in other states, such as Kentucky and New York, rates declined markedly). The second option was to score the tests while "holding constant" the difficulty of the questions, in order to facilitate comparisons of aggregate rates with those from previous years.

OSSE chose the latter option (according to some, in a manner that was insufficiently transparent). The end result was a modest increase in proficiency rates (which DC officials absurdly called “historic”).

The Great Proficiency Debate

A couple of weeks ago, Mike Petrilli of the Fordham Institute made the case that absolute proficiency rates should not be used as measures of school effectiveness, as they are heavily dependent on where students “start out” upon entry to the school. A few days later, Fordham president Checker Finn offered a defense of proficiency rates, noting that how much students know is substantively important, and associated with meaningful outcomes later in life.

They’re both correct. This is not a debate about whether proficiency rates are at all useful (by the way, I don't read Petrilli as saying that). It’s about how they should be used and how they should not.

Let’s keep this simple. Here is a quick, highly simplified list of how I would recommend interpreting and using absolute proficiency rates, and how I would avoid using them.

Subscribe to Testing Data