Standardized Tests

Schools Aren't The Only Reason Test Scores Change

In all my many posts about the interpretation of state testing data, it seems that I may have failed to articulate one major implication, which is almost always ignored in the news coverage of the release of annual testing data. That is: raw, unadjusted changes in student test scores are not by themselves very good measures of schools' test-based effectiveness.

In other words, schools can have a substantial impact on performance, but student test scores also increase, decrease or remain flat for reasons that have little or nothing to do with schools. The first, most basic reason is error. There is measurement error in all test scores - for various reasons, students taking the same test twice will get different scores, even if their "knowledge" remains constant. Also, as I've discussed many times, there is extra imprecision when using cross-sectional data. Often, any changes in scores or rates, especially when they’re small in magnitude and/or based on smaller samples (e.g., individual schools), do not represent actual progress (see here and here). Finally, even when changes are "real," other factors that influence test score changes include a variety of non-schooling inputs, such as parental education levels, family's economic circumstances, parental involvement, etc. These factors don't just influence how highly students score; they are also associated with progress (that's why value-added models exist).

Thus, to the degree that test scores are a valid measure of student performance, and changes in those scores a valid measure of student learning, schools aren’t the only suitors at the dance. We should stop judging school or district performance by comparing unadjusted scores or rates between years.

The Unfortunate Truth About This Year's NYC Charter School Test Results

There have now been several stories in the New York news media about New York City’s charter schools’ “gains” on this year’s state tests (see here, here, here, here and here). All of them trumpeted the 3-7 percentage point increase in proficiency among the city’s charter students, compared with the 2-3 point increase among their counterparts in regular public schools. The consensus: Charters performed fantastically well this year.

In fact, the NY Daily News asserted that the "clear lesson" from the data is that "public school administrators must gain the flexibility enjoyed by charter leaders," and "adopt [their] single-minded focus on achievement." For his part, Mayor Michael Bloomberg claimed that the scores are evidence that the city should expand its charter sector.

All of this reflects a fundamental misunderstanding of how to interpret testing data, one that is frankly a little frightening to find among experienced reporters and elected officials.

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).

How Often Do Proficiency Rates And Average Scores Move In Different Directions?

New York State is set to release its annual testing data today. Throughout the state, and especially in New York City, we will hear a lot about changes in school and district proficiency rates. The rates themselves have advantages – they are easy to understand, comparable across grades and reflect a standards-based goal. But they also suffer severe weaknesses, such as their sensitivity to where the bar is set and the fact that proficiency rates and the actual scores upon which they’re based can paint very different pictures of student performance, both in a given year as well as over time. I’ve discussed this latter issue before in the NYC context (and elsewhere), but I’d like to revisit it quickly.

Proficiency rates can only tell you how many students scored above a certain line; they are completely uninformative as to how far above or below that line the scores might be. Consider a hypothetical example: A student who is rated as proficient in year one might make large gains in his or her score in year two, but this would not be reflected in the proficiency rate for his or her school – in both years, the student would just be coded as “proficient” (the same goes for large decreases that do not “cross the line”). As a result, across a group of students, the average score could go up or down while proficiency rates remained flat or moved in the opposite direction. Things are even messier when data are cross-sectional (as public data lmost always are), since you’re comparing two different groups of students (see this very recent NYC IBO report).

Let’s take a rough look at how frequently rates and scores diverge in New York City.

The Busy Intersection Of Test-Based Accountability And Public Perception

Last year, the New York City Department of Education (NYCDOE) rolled out its annual testing results for the city’s students in a rather misleading manner. The press release touted the “significant progress” between 2010 and 2011 among city students, while, at a press conference, Mayor Michael Bloomberg called the results “dramatic." In reality, however, the increase in proficiency rates (1-3 percentage points) was very modest, and, more importantly, the focus on the rates hid the fact that actual scale scores were either flat or decreased in most grades. In contrast, one year earlier, when the city's proficiency rates dropped due to the state raising the cut scores, Mayor Bloomberg told reporters (correctly) that it was the actual scores that "really matter."

Most recently, in announcing their 2011 graduation rates, the city did it again. The headline of the NYCDOE press release proclaims that “a record number of students graduated from high school in 2011." This may be technically true, but the actual increase in the rate (rather than the number of graduates) was 0.4 percentage points, which is basically flat (as several reporters correctly noted). In addition, the city's "college readiness rate" was similarly stagnant, falling slightly from 21.4 percent to 20.7 percent, while the graduation rate increase was higher both statewide and in New York State's four other large districts (the city makes these comparisons when they are favorable).*

We've all become accustomed to this selective, exaggerated presentation of testing data, which is of course not at all limited to NYC. And it illustrates the obvious fact that test-based accountability plays out in multiple arenas, formal and informal, including the court of public opinion.

Colorado's Questionable Use Of The Colorado Growth Model

I have been writing critically about states’ school rating systems (e.g., Ohio, Florida, Louisiana), and I thought I would find one that is, at least in my (admittedly value-laden) opinion, more defensibly designed. It didn't quite turn out as I had hoped.

One big starting point in my assessment is how heavily the systems weight absolute performance (how highly students score) versus growth (how quickly students improve). As I’ve argued many times, the former (absolute level) is a poor measure of school performance in a high-stakes accountability system. It does not address the fact that some schools, particularly those in more affluent areas, serve  students who, on average, enter the system at a higher-performing level. This amounts to holding schools accountable for outcomes they largely cannot control (see Doug Harris' excellent book for more on this in the teacher context). Thus, to whatever degree testing results can be used to judge actual school effectiveness, growth measures, while themselves highly imperfect, are to be preferred in a high-stakes context.

There are a few states that assign more weight to growth than absolute performance (see this prior post on New York City’s system). One of them is Colorado's system, which uses the well-known “Colorado Growth Model” (CGM).*

In my view, putting aside the inferential issues with the CGM (see the first footnote), the focus on growth in Colorado's system is in theory a good idea. But, looking at the data and documentation reveals a somewhat unsettling fact: There is a double standard of sorts, by which two schools with the same growth score can receive different ratings, and it's mostly their absolute performance levels determining whether this is the case.

Louisiana's "School Performance Score" Doesn't Measure School Performance

Louisiana’s "School Performance Score" (SPS) is the state’s primary accountability measure, and it determines whether schools are subject to high-stakes decisions, most notably state takeover. For elementary and middle schools, 90 percent of the SPS is based on testing outcomes. For secondary schools, it is 70 percent (and 30 percent graduation rates).*

The SPS is largely calculated using absolute performance measures – specifically, the proportion of students falling into the state’s cutpoint-based categories (e.g., advanced, mastery, basic, etc.). This means that it is mostly measuring student performance, rather than school performance. That is, insofar as the SPS only tells you how high students score on the test, rather than how much they have improved, schools serving more advantaged populations will tend to do better (since their students tend to perform well when they entered the school) while those in impoverished neighborhoods will tend to do worse (even those whose students have made the largest testing gains).

One rough way to assess this bias is to check the association between SPS and student characteristics, such as poverty. So let’s take a quick look.

Gender Pay Gaps And Educational Achievement Gaps

There is currently an ongoing rhetorical war of sorts over the gender wage gap. One “side” makes the common argument that women earn around 75 cents on the male dollar (see here, for example).

Others assert that the gender gap is a myth, or that it is so small as to be unimportant.

Often, these types of exchanges are enough to exasperate the casual observer, and inspire claims such as “statistics can be made to say anything." In truth, however, the controversy over the gender gap is a good example of how descriptive statistics, by themselves, say nothing. What matters is how they’re interpreted.

Moreover, the manner in which one must interpret various statistics on the gender gap applies almost perfectly to the achievement gaps that are so often mentioned in education debates.

A Game Of Inches

One of the more telling episodes in education I’ve seen over the past couple of years was a little dispute over Michelle Rhee’s testing record that flared up last year. Alan Ginsburg, a retired U.S. Department of Education official, released an informal report in which he presented the NAEP cohort changes that occurred during the first two years of Michelle Rhee’s tenure (2007-2009), and compared them with those during the superintendencies of her two predecessors.

Ginsburg concluded that the increases under Chancellor Rhee, though positive, were less rapid than in previous years (2000 to 2007 in math, 2003 to 2007 in reading). Soon thereafter, Paul Peterson, director of Harvard’s Program on Educational Leadership and Governance, published an article in Education Next that disputed Ginsburg’s findings. Peterson found that increases under Rhee amounted to roughly three scale score points per year, compared with around 1-1.5 points annually between 2000 and 2007 (the actual amounts varied by subject and grade).

Both articles were generally cautious in tone and in their conclusions about the actual causes of the testing trends. The technical details of the two reports – who’s “wrong” or “right” - are not important for this post (especially since more recent NAEP results have since been released). More interesting was how people reacted - and didn’t react - to the dueling analyses.

We Should Only Hold Schools Accountable For Outcomes They Can Control

Let’s say we were trying to evaluate a teacher’s performance for this academic year, and part of that evaluation would use students’ test scores (if you object to using test scores this way, put that aside for a moment). We checked the data and reached two conclusions. First, we found that her students made fantastic progress this year. Second, we also saw that the students’ scores were still quite a bit lower than their peers’ in the district. Which measure should we use to evaluate this teacher?

Would we consider judging her even partially based on the latter – students’ average scores? Of course not. Those students made huge progress, and the only reason their absolute performance levels are relatively low is because they were low at the beginning of the year. This teacher could not control the fact that she was assigned lower-scoring students. All she can do is make sure that they improve. That’s why no teacher evaluation system places any importance on students’ absolute performance, instead focusing on growth (and, of course, non-test measures). In fact, growth models control for absolute performance (prior year’s test scores) so it doesn't bias the results.

If we would never judge teachers based on absolute performance, why are we judging schools that way? Why does virtually every school/district rating system place some emphasis – often the primary emphasis – on absolute performance?