The state Education Department released statewide scores today for high school grades on the 2014 Partnership for Assessment of Readiness for College and Careers (PARCC) test of academic proficiency.
It was the first year of the new test. And also the last. PARCC is going to be replaced by another test next year — due partly, although not entirely, to PARCC’s association with the now-controversial Common Core state standards. Next year, the state switches to the ACT Aspire, which is also aligned to Common Core, although less intimately so. Arkansas educators are split on whether the move to ACT Aspire was the right one: PARCC was rigorous but time consuming, and many schools reported frustrating technical problems.
Nonetheless, Arkansas will have three different test results over three years, which makes year-to-year comparisons a little difficult.
The state gives these figures for students scoring a 3 or above on PARCC, a grade considered sufficient to do college work. (ELA stands for English Language Arts.)
The state said 23,376 students took non-required Grade 11 English and Algebra 2 tests. The results put 69 percent at level 3 and above in English and 39 percent at level 3 and above in Algebra 2. PARCC scores for grades 3 to 8 will be released in November. Scores for individual districts and schools are pending as well.
UPDATE FROM BENJI:
The most obvious question that arises from today’s data is this: How do these scores compare with 2014 scores? Can we draw any comparison at all with previous years’ statewide performance results on the old test, the Arkansas Benchmark? (I’m here using “Benchmark” as a shorthand to the old testing regime in general, including end-of-course tests for high school.)
In my opinion, yes — although it tells us more about the test than it does about the kids. I don’t think the PARCC scores say anything meaningful about Arkansas student growth over the past year. But they do suggest that the Benchmark, because of its lack of rigor, may have glossed over academic deficiencies in Arkansas when taken in aggregate.
Let me emphasize this is my opinion, not ADE’s. Hope Allen, the Director of Student Assessment at the Education Department, told me “you have to be careful when making comparisons of that nature, because they’re totally different tests with different standards.” The department is right now working to see “what correlations can we draw from Benchmark to PARCC in relation to the scores,” Allen said.
Nonetheless, I’m going to throw caution to the wind and hazard a couple of broad conclusions.
At today’s meeting, the Board of Education approved “cut scores” for PARCC, which were established by a panel drawn from all the states that participated in the common test; cut scores establish ranges defining student performance. PARCC sorts students into five groups:
The old Arkansas Benchmark, meanwhile, had four such performance-based categories: Advanced, Proficient, Basic and Below Basic.Those numbers are of tremendous importance to schools: For example, if fewer than 49.5 percent of a school’s students score Proficient or Advanced, the school is considered to be in academic distress. That opens the door to a state takeover of the district, potentially.
There’s no apples-to-apples to be found here, but I think it’s fair to say that a “Level 3 or above” on PARCC is roughly analogous to “Proficient or above” on the Benchmark. According to Allen, Levels 1 and 2 on PARCC indicate that a student is in need of an academic intervention of some kind; I’d argue that Basic and Below Basic indicate the same thing. Allen said that a Level 3 indicates that a student is “on track for college and career readiness,” from which I conclude that might be more accurate to think of a Level 3 as a range that straddles the line between Basic and Proficient.
That, however, only emphasizes the important point, which is that a significantly smaller percentage of Arkansas students, in aggregate, scored at a Level 3 or above in 2015 than scored Proficient or above in 2014.
Here’s ADE’s end-of-course testing data for 2014. There’s no data here for ELA in grades 9 and 10, so let’s focus on the subjects in which there are indeed scores for both 2014 Benchmark and 2015 PARCC: Algebra 1 and Geometry.
On the 2014 Benchmark, 32,014 students took the Algebra 1 test, and 75 percent scored Proficient and above. For Geometry, 33,744 students took the test and 74 percent scored Proficient or above.
On the 2015 PARCC, 34,433 students took the Algebra 1 test, and 60 percent scored a Level 3 or above. For Geometry, 32,611 students took the PARCC and 57 percent scored a Level 3 or above.
That is, if you accept my premise that “Proficient” and “Level 3” both essentially mean “an acceptable performance,” merely changing from the old test to the new test dropped the percentage of kids who performed at an acceptable level in Algebra by 15 percentage points and in Geometry by 17 percentage points. (I’m avoiding the Grade 11 ELA scores here since not all 11th graders took that exam this year.)
What does this mean? Well, as has already happened in other states, many Arkansan kids this year will get scores that indicate they’re not “on track for college and career readiness” — including, inevitably, a large number of kids who received a designation of Proficient or above last year (maybe always). This will be difficult for kids. It will be upsetting to parents and uncomfortable for schools.
None of this is surprising in the least. The Benchmark, by all accounts, was a pretty crappy test. PARCC was far more rigorous, despite various other problems with the test. It’s why the New York Times reported recently that some other states using PARCC (or other Common Core-aligned tests) have been subtly moving the goalposts on what actually does constitute proficiency.
Especially now that the state has moved on to another test, it’ll be easy to blame PARCC for the drop in scores. In a sense, that’s right. But if the ACT Aspire is as rigorous as it should be, its results next year will also show a significant drop from the 2014 Benchmark scores. The point of having a tougher test — tied to more rigorous standards — is that it should hold schools to a higher common standard.
Yet at the same time, this all shows the silliness of staking everything on a standardized test in the first place. I firmly believe that high standards and accountability matter, and that there must be some kind of quantitative assessment to gauge student and school performance effectively. But I also think it’s self-evident the testing regime has spiraled into an absurdist, cart-before-the-horse nightmare, often effectively penalizing schools and teachers for serving disadvantaged students rather than their more affluent (and thus higher performing, on average) peers. Not to mention the psychological effect on lower-performing students when their sense of self-worth in the school building is entirely reduced to a single metric. And, the inevitable incentive to cheat in ways large and small — including, at the state level, adjusting proficiency levels as politically necessary.
The preliminary PARCC scores illustrate why Arkansas absolutely needs higher standards. They also show the limitations of making those standards an object of worship.