Florida

The Test-Based Evidence On The "Florida Formula"

** Reprinted here in the Washington Post

Former Florida Governor Jeb Bush has become one of the more influential education advocates in the country. He travels the nation armed with a set of core policy prescriptions, sometimes called the “Florida formula," as well as "proof" that they work. The evidence that he and his supporters present consists largely of changes in average statewide test scores – NAEP and the state exam (FCAT) – since the reforms started going into place. The basic idea is that increases in testing results are the direct result of these policies.

Governor Bush is no doubt sincere in his effort to improve U.S. education, and, as we'll see, a few of the policies comprising the “Florida formula” have some test-based track record. However, his primary empirical argument on their behalf – the coincidence of these policies’ implementation with changes in scores and proficiency rates – though common among both “sides” of the education debate, is simply not valid. We’ve discussed why this is the case many times (see here, here and here), as have countless others, in the Florida context as well as more generally.*

There is no need to repeat those points, except to say that they embody the most basic principles of data interpretation and causal inference. It would be wonderful if the evaluation of education policies – or of school systems’ performance more generally - was as easy as looking at raw, cross-sectional testing data. But it is not.

Luckily, one need not rely on these crude methods. We can instead take a look at some of the rigorous research that has specifically evaluated the core reforms comprising the “Florida formula." As usual, it is a far more nuanced picture than supporters (and critics) would have you believe.

What Florida's School Grades Measure, And What They Don't

A while back, I argued that Florida's school grading system, due mostly to its choice of measures, does a poor job of gauging school performance per se. The short version is that the ratings are, to a degree unsurpassed by most other states' systems, driven by absolute performance measures (how highly students score), rather than growth (whether students make progress). Since more advantaged students tend to score more highly on tests when they enter the school system, schools are largely being judged not on the quality of instruction they provide, but rather on the characteristics of the students they serve.

New results were released a couple of weeks ago. This was highly anticipated, as the state had made controversial changes to the system, most notably the inclusion of non-native English speakers and special education students, which officials claimed they did to increase standards and expectations. In a limited sense, that's true - grades were, on average, lower this year. The problem is that the system uses the same measures as before (including a growth component that is largely redundant with proficiency). All that has changed is the students that are included in them. Thus, to whatever degree the system now reflects higher expectations, it is still for outcomes that schools mostly cannot control.

I fully acknowledge the political and methodological difficulties in designing these systems, and I do think Florida's grades, though exceedingly crude, might be useful for some purposes. But they should not, in my view, be used for high-stakes decisions such as closure, and the public should understand that they don't tell you much about the actual effectiveness of schools. Let’s take a very quick look at the new round of ratings, this time using schools instead of districts (I looked at the latter in my previous post about last year's results).

Herding FCATs

About a week ago, Florida officials went into crisis mode after revealing that the proficiency rate on the state’s writing test (FCAT) dropped from 81 percent to 27 percent among fourth graders, with similarly large drops in the other two grades in which the test is administered (eighth and tenth). The panic was almost immediate. For one thing, performance on the writing FCAT is counted in the state’s school and district ratings. Many schools would end up with lower grades and could therefore face punitive measures.

Understandably, a huge uproar was also heard from parents and community members. How could student performance decrease so dramatically? There was so much blame going around that it was difficult to keep track – the targets included the test itself, the phase-in of the state’s new writing standards, and test-based accountability in general.

Despite all this heated back-and-forth, many people seem to have overlooked one very important, widely-applicable lesson here: That proficiency rates, which are not "scores," are often extremely sensitive to where you set the bar.

The Weighting Game

A while back, I noted that states and districts should exercise caution in assigning weights (importance) to the components of their teacher evaluation systems before they know what the other components will be. For example, most states that have mandated new evaluation systems have specified that growth model estimates count for a certain proportion (usually 40-50 percent) of teachers’ final scores (at least those in tested grades/subjects), but it’s critical to note that the actual importance of these components will depend in no small part on what else is included in the total evaluation, and how it's incorporated into the system.

In slightly technical terms, this distinction is between nominal weights (the percentage assigned) and effective weights (the percentage that actually ends up being the case). Consider an extreme hypothetical example – let’s say a district implements an evaluation system in which half the final score is value-added and half is observations. But let’s also say that every teacher gets the same observation score. In this case, even though the assigned (nominal) weight for value-added is 50 percent, the actual importance (effective weight) will be 100 percent, since every teacher receives the same observation score, and so all the variation between teachers’ final scores will be determined by the value-added component.

This issue of nominal/versus effective weights is very important, and, with exceptions, it gets almost no attention. And it’s not just important in teacher evaluations. It’s also relevant to states’ school/district grading systems. So, I think it would be useful to quickly illustrate this concept in the context of Florida’s new district grading system.

Do Value-Added Models "Control For Poverty?"

There is some controversy over the fact that Florida’s recently-announced value-added model (one of a class often called “covariate adjustment models”), which will be used to determine merit pay bonuses and other high-stakes decisions, doesn’t include a direct measure of poverty.

Personally, I support adding a direct income proxy to these models, if for no other reason than to avoid this type of debate (and to facilitate the disaggregation of results for instructional purposes). It does bear pointing out, however, that the measure that’s almost always used as a proxy for income/poverty – students’ eligibility for free/reduced-price lunch – is terrible as a poverty (or income) gauge. It tells you only whether a student’s family has earnings below (or above) a given threshold (usually 185 percent of the poverty line), and this masks most of the variation among both eligible and non-eligible students. For example, families with incomes of $5,000 and $20,000 might both be coded as eligible, while families earning $40,000 and $400,000 are both coded as not eligible. A lot of hugely important information gets ignored this way, especially when the vast majority of students are (or are not) eligible, as is the case in many schools and districts.

That said, it’s not quite accurate to assert that Florida and similar models “don’t control for poverty." The model may not include a direct income measure, but it does control for prior achievement (a student’s test score in the previous year[s]). And a student’s test score is probably a better proxy for income than whether or not they’re eligible for free/reduced-price lunch.

Even more importantly, however, the key issue about bias is not whether the models “control for poverty," but rather whether they control for the range of factors – school and non-school – that are known to affect student test score growth, independent of teachers’ performance. Income is only one part of this issue, which is relevant to all teachers, regardless of the characteristics of the students that they teach.

A Dark Day For Educational Measurement In The Sunshine State

Just this week, Florida announced its new district grading system. These systems have been popping up all over the nation, and given the fact that designing one is a requirement of states applying for No Child Left Behind waivers, we are sure to see more.

I acknowledge that the designers of these schemes have the difficult job of balancing accessibility and accuracy. Moreover, the latter requirement – accuracy – cannot be directly tested, since we cannot know “true” school quality. As a result, to whatever degree it can be partially approximated using test scores, disagreements over what specific measures to include and how to include them are inevitable (see these brief analyses of Ohio and California).

As I’ve discussed before, there are two general types of test-based measures that typically comprise these systems: absolute performance and growth. Each has its strengths and weaknesses. Florida’s attempt to balance these components is a near total failure, and it shows in the results.