On Tuesday, officials released dismal scores from the new Common Core-aligned tests students in the District took last spring. The next day, another set of scores showed DC students improving faster than those in the rest of the country. One thing that was consistent in the results was a large gap between rich and poor.

The first set of scores, on standardized tests known as PARCC, showed that only 25% of DC high school students were “college and career ready” in English. Even worse, only 10% met that bar on a test of high school geometry.

That looks like a huge drop from scores on the old DC test, known as the DC CAS. Last year, about 50% of 10th graders scored proficient or advanced on those reading and math tests. (PARCC scores for 3rd through 8th graders won’t be available until next month. They may be somewhat better than the high school scores, but will probably also decline significantly.)

You might conclude that students’ skills have suddenly plummeted, but in fact the two tests aren’t comparable. The PARCC tests, which are designed to measure whether students have the skills they’ll need to succeed in college, are far more rigorous. Instead of asking students to write an essay about their dream vacation, for example, a test might give them two sophisticated passages to read and ask them to make detailed comparisons.

In past years, DC’s education leaders and elected officials have celebrated incremental progress on the DC CAS. This year, they lamented the low PARCC scores and somberly declared they were prepared to do the hard but necessary work to improve them. But barely had the words left their mouths, or their press releases, when their lamentations turned to joy.

**DC’s growth in NAEP scores outpaces the rest of the country**

That’s because DC bucked a national trend on the other test: the NAEP, given to a sample of students across the country every two years. The NAEP is considered far more rigorous than most of the old state tests, including the DC CAS. This year, math scores for 4th and 8th graders declined nationwide. But in DC, 4th grade scores rose by three points in math and seven in reading, while 8th grade scores remained flat.

On the face of it, the PARCC and NAEP results appear contradictory. But the tests were differently constructed, and they were assessing different grade levels. And while it’s true DC’s NAEP scores have gone up, they’re still at or near the bottom compared to other states.

But of course, DC is more like a city than a state. And cities, which have higher concentrations of poverty, tend to have lower test scores. So it’s fairer to assess DC’s performance against another set of NAEP scores that compares large urban school districts to one another—although those scores include results only from DC Public Schools, not the charter sector.

On that measure, DCPS has improved—in fact, as DCPS Chancellor Kaya Henderson likes to point out, it’s the fastest-improving urban school district in the country. In 2007, DC was at the bottom of the NAEP list of districts. Now, for 4th grade results, it’s in the middle. But for 8th grade, it’s still near the bottom of 21 districts: higher than two others in reading and only one other (Detroit) in math.

**Both sets of scores reveal gaps between subgroups**

One thing that the NAEP and PARCC scores have in common is that they reveal the width and persistence of DC’s achievement gap. On the PARCC test, for example, School Without Walls—a selective DCPS school with a relatively affluent student body—saw 97% of its students reach the “college and career ready” bar in English and 76% in math. At the other end of the spectrum, three high-poverty DCPS schools had no students meet that bar in English, and eight had none who reached it in math.

Some previously high-performing charter schools saw their scores drop precipitously, as has happened elsewhere. Last year at KIPP DC College Prep, where most students are black and low-income, 95% of students scored proficient in math and 71% in reading on the DC CAS. This year, just under 20% of students met the PARCC bar in either subject.

Gaps between ethnic and socioeconomic groups loomed wide on the DC CAS, but PARCC has turned them into chasms. On the PARCC English test, for example, 82% of white students met “college and career ready” expectations, compared to 20% of black students, 25% of Hispanics, and 17% of economically disadvantaged students. On last year’s DC CAS in reading, an even higher percentage of whites scored proficient—92%—but so did 44% of black students, 50% of Hispanics, and 42% of the economically disadvantaged.

And despite the increases on the NAEP for DC as a whole, a demographic breakdown of DCPS’s scores reveals that gaps seen in previous years haven’t budged. In 8th grade reading, for example, 75% of white students scored proficient as compared to 11% of black students, 17% of Hispanics, and 8% of low-income students. In DC as a whole, the gaps between white students on the one hand and black and Hispanic students on the other are the largest in the nation on the NAEP, according to one calculation.

To be fair, one reason for the size of the gaps is that DC’s white and affluent students perform at an unusually high level. But to begin to close that gap, disadvantaged students need to improve faster than white ones. In fact, looking just at DCPS scores, proficiency rates for white students have either gone up or stayed the same on each of the four tests since 2013, while at least one other subgroup’s rate has gone down on all but one of the tests.

On the English and reading side, the root cause of the gap in scores is the relative lack of exposure low-income students have to knowledge and vocabulary, starting from birth—a deficit elementary schools usually reinforce by failing to focus on building students’ knowledge. (Frankly, I’m not sure how to explain the abysmally low PARCC math scores: even among white students, only 52% met “college and career ready” expectations.)

Some will see the test scores, and the gaps they reveal, as evidence that education reform hasn’t worked. Critics of the Common Core standards may use the PARCC results to argue the tests are unrealistically hard. Those on the other side will say the Common Core is revealing deficiencies that were masked by the DC CAS, and point out that this kind of change takes time.

There’s some truth to all those arguments. But the bottom line is that our schools are continuing to fail many students who enter with the greatest deficits, and we need to find a way to bring their knowledge and vocabulary closer to the level of their affluent peers. Test scores can tell us how far we still have to go, but they won’t tell us how to get there.

*Cross-posted at Greater Greater Washington.*