The Accessibility Conundrum In Accountability Systems

One of the major considerations in designing accountability policy, whether in education or other fields, is what you might call accessibility. That is, both the indicators used to construct measures and how they are calculated should be reasonably easy for stakeholders to understand, particularly if the measures are used in high-stakes decisions.

This important consideration also generates great tension. For example, complaints that Florida’s school rating system is “too complicated” have prompted legislators to make changes over the years. Similarly, other tools – such as procedures for scoring and establishing cut points for standardized tests, and particularly the use of value-added models – are routinely criticized as too complex for educators and other stakeholders to understand. There is an implicit argument underlying these complaints: If people can’t understand a measure, it should not be used to hold them accountable for their work. Supporters of using these complex accountability measures, on the other hand, contend that it’s more important for the measures to be “accurate” than easy to understand.

I personally am a bit torn. Given the extreme importance of accountability systems’ credibility among those subject to them, not to mention the fact that performance evaluations must transmit accessible and useful information in order to generate improvements, there is no doubt that overly complex measures can pose a serious problem for accountability systems. It might be difficult for practitioners to adjust their practice based on a measure if they don't understand that measure, and/or if they are unconvinced that the measure is transmitting meaningful information. And yet, the fact remains that measuring the performance of schools and individuals is extremely difficult, and simplistic measures are, more often than not, inadequate for these purposes.

Multiple Measures And Singular Conclusions In A Twin City

A few weeks ago, the Minneapolis Star Tribune published teacher evaluation results for the district’s public school teachers in 2013-14. This decision generated a fair amount of controversy, but it’s worth noting that the Tribune, unlike the Los Angeles Times and New York City newspapers a few years ago, did not publish scores for individual teachers, only totals by school.

The data once again provide an opportunity to take a look at how results vary by student characteristics. This was indeed the focus of the Tribune’s story, which included the following headline: “Minneapolis’ worst teachers are in the poorest schools, data show." These types of conclusions, which simply take the results of new evaluations at face value, have characterized the discussion since the first new systems came online. Though understandable, they are also frustrating and a potential impediment to the policy process. At this early point, “the city’s teachers with the lowest evaluation ratings” is not the same thing as “the city’s worst teachers." Actually, as discussed in a previous post, the systematic variation in evaluation results by student characteristics, which the Tribune uses to draw conclusions about the distribution of the city’s “worst teachers," could just as easily be viewed as one of the many ways that one might assess the properties and even the validity of those results.

So, while there are no clear-cut "right" or "wrong" answers here, let’s take a quick look at the data and what they might tell us.

The Great Teacher Evaluation Evaluation: New York Edition

A couple of weeks ago, the New York State Education Department (NYSED) released data from the first year of the state's new teacher and principal evaluation system (called the “Annual Professional Performance Review," or APPR). In what has become a familiar pattern, this prompted a wave of criticism from advocates, much of it focused on the proportion of teachers in the state to receive the lowest ratings.

To be clear, evaluation systems that produce non-credible results should be examined and improved, and that includes those that put implausible proportions of teachers in the highest and lowest categories. Much of the commentary surrounding this and other issues has been thoughtful and measured. As usual, though, there have been some oversimplified reactions, as exemplified by this piece on the APPR results from Students First NY (SFNY).

SFNY notes what it considers to be the low proportion of teachers rated “ineffective," and points out that there was more differentiation across rating categories for the state growth measure (worth 20 percent of teachers’ final scores), compared with the local “student learning” measure (20 percent) and the classroom observation components (60 percent). Based on this, they conclude that New York’s "state test is the only reliable measure of teacher performance" (they are actually talking about validity, not reliability, but we’ll let that go). Again, this argument is not representative of the commentary surrounding the APPR results, but let’s use it as a springboard for making a few points, most of which are not particularly original. (UPDATE: After publication of this post, SFNY changed the headline of their piece from "the only reliable measure of teacher performance" to "the most reliable measure of teacher performance.")

A Quick Look At The ASA Statement On Value-Added

Several months ago, the American Statistical Association (ASA) released a statement on the use of value-added models in education policy. I’m a little late getting to this (and might be repeating points that others made at the time), but I wanted to comment on the statement, not only because I think it's useful to have ASA add their perspective to the debate on this issue, but also because their statement seems to have become one of the staple citations for those who oppose the use of these models in teacher evaluations and other policies.

Some of these folks claimed that the ASA supported their viewpoint – i.e., that value-added models should play no role in accountability policy. I don’t agree with this interpretation. To be sure, the ASA authors described the limitations of these estimates, and urged caution, but I think that the statement rather explicitly reaches a more nuanced conclusion: That value-added estimates might play a useful role in education policy, as one among several measures used in formal accountability systems, but this must be done carefully and appropriately.*

Much of the statement puts forth the standard, albeit important, points about value-added (e.g., moderate stability between years/models, potential for bias, etc.). But there are, from my reading, three important takeaways that bear on the public debate about the use of these measures, which are not always so widely acknowledged.

Differences In DC Teacher Evaluation Ratings By School Poverty

In a previous post, I discussed simple data from the District of Columbia Public Schools (DCPS) on teacher turnover in high- versus lower-poverty schools. In that same report, which was issued by the D.C. Auditor and included, among other things, descriptive analyses by the excellent researchers from Mathematica, there is another very interesting table showing the evaluation ratings of DC teachers in 2010-11 by school poverty (and, indeed, DC officials deserve credit for making these kinds of data available to the public, as this is not the case in many other states).

DCPS’ well-known evaluation system (called IMPACT) varies between teachers in tested versus non-tested grades, but the final ratings are a weighted average of several components, including: the teaching and learning framework (classroom observations); commitment to the school community (attendance at meetings, mentoring, PD, etc.); schoolwide value-added; teacher-assessed student achievement data (local assessments); core professionalism (absences, etc.); and individual value-added (tested teachers only).

The table I want to discuss is on page 43 of the Auditor’s report, and it shows average IMPACT scores for each component and overall for teachers in high-poverty schools (80-100 percent free/reduced-price lunch), medium poverty schools (60-80 percent) and low-poverty schools (less than 60 percent). It is pasted below.

The Proportionality Principle In Teacher Evaluations

Our guest author today is Cory Koedel, Assistant Professor of Economics at the University of Missouri.

In a 2012 post on this blog, Dr. Di Carlo reviewed an article that I coauthored with colleagues Mark Ehlert, Eric Parsons and Michael Podgursky. The initial article (full version here, or for a shorter, less-technical version, see here) argues for the policy value of growth models that are designed to force comparisons to be between schools and teachers in observationally-similar circumstances.

The discussion is couched within the context of achieving three key policy objectives that we associate with the adoption of more-rigorous educational evaluation systems: (1) improving system-wide instruction by providing useful performance signals to schools and teachers; (2) eliciting optimal effort from school personnel; and (3) ensuring that current labor-market inequities between advantaged and disadvantaged schools are not exacerbated by the introduction of the new systems.

We argue that a model that forces comparisons to be between equally-circumstanced schools and teachers – which we describe as a “proportional” model – is best-suited to achieve these policy objectives. The conceptual appeal of the proportional approach is that it fully levels the playing field between high- and low-poverty schools. In contrast, some other growth models have been shown to produce estimates that are consistently associated with the characteristics of students being served (e.g., Student Growth Percentiles).

Revisiting The Widget Effect

In 2009, The New Teacher Project (TNTP) released a report called “The Widget Effect." You would be hard-pressed to find too many more recent publications from an advocacy group that had a larger influence on education policy and the debate surrounding it. To this day, the report is mentioned regularly by advocates and policy makers.

The primary argument of the report was that teacher performance “is not measured, recorded, or used to inform decision making in any meaningful way." More specifically, the report shows that most teachers received “satisfactory” or equivalent ratings, and that evaluations were not tied to most personnel decisions (e.g., compensation, layoffs, etc.). From these findings and arguments comes the catchy title – a “widget” is a fictional product commonly used in situations (e.g., economics classes) where the product doesn’t matter. Thus, treating teachers like widgets means that we treat them all as if they’re the same.

Given the influence of “The Widget Effect," as well as how different the teacher evaluation landscape is now compared to when it was released, I decided to read it closely. Having done so, I think it’s worth discussing a few points about the report.

Matching Up Teacher Value-Added Between Different Tests

The U.S. Department of Education has released a very short, readable report on the comparability of value-added estimates using two different tests in Indiana – one of them norm-referenced (the Measures of Academic Progress test, or MAP), and the other criterion-referenced (the Indiana Statewide Testing for Educational Progress Plus, or ISTEP+, which is also the state’s official test for NCLB purposes).

The research design here is straightforward – fourth and fifth grade students in 46 schools across 10 districts in Indiana took both tests, their teachers’ value-added scores were calculated, and the scores were compared. Since both sets of scores were based on the same students and teachers, this is allows a direct comparison of how teachers’ value-added estimates compare between these two tests. The results are not surprising, and they square with similar prior studies (see here, here, here, for example): The estimates based on the two tests are moderately correlated. Depending on the grade/subject, they are between 0.4 and 0.7. If you’re not used to interpreting correlation coefficients, consider that only around one-third of teachers were in the same quintile (fifth) on both tests, and another 40 or so percent were one quintile higher or lower. So, most teachers were within a quartile, about a quarter of teachers moved two or more quintiles, and a small percentage moved from top to bottom or vice-versa.

Although, as mentioned above, these findings are in line with prior research, it is worth remembering why this “instability” occurs (and what can be done about it).

The Year In Research On Market-Based Education Reform: 2013 Edition

In the three most discussed and controversial areas of market-based education reform – performance pay, charter schools and the use of value-added estimates in teacher evaluations – 2013 saw the release of a couple of truly landmark reports, in addition to the normal flow of strong work coming from the education research community (see our reviews from 2010, 2011 and 2012).*

In one sense, this building body of evidence is critical and even comforting, given not only the rapid expansion of charter schools, but also and especially the ongoing design and implementation of new teacher evaluations (which, in many cases, include performance-based pay incentives). In another sense, however, there is good cause for anxiety. Although one must try policies before knowing how they work, the sheer speed of policy change in the U.S. right now means that policymakers are making important decisions on the fly, and there is great deal of uncertainty as to how this will all turn out.

Moreover, while 2013 was without question an important year for research in these three areas, it also illustrated an obvious point: Proper interpretation and application of findings is perhaps just as important as the work itself.

A Few Additional Points About The IMPACT Study

The recently released study of IMPACT, the teacher evaluation system in the District of Columbia Public Schools (DCPS), has garnered a great deal of attention over the past couple of months (see our post here).

Much of the commentary from the system’s opponents was predictably (and unfairly) dismissive, but I’d like to quickly discuss the reaction from supporters. Some took the opportunity to make grand proclamations about how “IMPACT is working," and there was a lot of back and forth about the need to ensure that various states’ evaluations are as “rigorous” as IMPACT (as well as skepticism as to whether this is the case).

The claim that this study shows that “IMPACT is working” is somewhat misleading, and the idea that states should now rush to replicate IMPACT is misguided. It also misses the important points about the study and what we can learn from its results.