#### The Maryland State Department of Education today released state-level results from the PARCC tests taken by high school students for algebra I, algebra II, and English 10 last spring.

The three tests, each of which was given in two parts last spring but will be whittled down to just one in 2016, are typically taken by students after completing most of the associated class in high school. For example, if a student takes algebra II and English 10 during his sophomore year, he would take both the algebra II and English 10 tests near the conclusion of that year. PARCC stands for Partnership for Assessment of Readiness for College and Careers. It’s a multi-state consortium that develops, administers, scores, and reports results from large-scale standardized tests. Schools are required to administer the tests, or a suitable equivalent, under federal law.

PARCC also develops high school-level tests in geometry, English 9, and English 11, but those tests were not given in Maryland in the spring. This year, students will have the option to take these other tests as well.

## Explanation of PARCC performance levels

PARCC states have agreed on a scale to describe a student’s performance on the assessments ranging from 1 to 5. Then, Maryland educators decided that they would consider, at the high school level, a score of 4 or 5 to indicate that a student had demonstrated college- and career-ready performance in the subject on the standardized test given. A score of 1 would mean the student showed no or minimal command of the material.

In algebra I, for example, a student performing at Level 3 in PARCC’s scoring system would also be at Level 3 in Maryland’s scoring system. He would have demonstrated, for example, that he can reliably “calculate sums and products of two rational and/or irrational numbers.” A student performing at Level 4, however, can reliably “calculate sums and products of two rational and/or irrational numbers and determine whether the sums and products are rational or irrational.”

Now, when I say “reliably,” what I actually mean is that a student would get about half the problems right and half wrong that measure that particular knowledge, assuming a perfect random distribution of difficulty level. That is, a student who gets a 4 or 5 on the algebra I test would get more than half the problems correct in which he had to “calculate sums and products of two rational and/or irrational numbers and determine whether the sums and products are rational or irrational.” A student who got a 3 or lower would get more than half of those problems wrong. A score of 1 could mean the student can’t multiply irrational numbers at all.

(If the difficulties of problems aren’t perfectly distributed, the designation of “half” *might* refer to the percentage of problems half the kids answered correctly rather than half the problems on the test, but how that percentage is determined exactly depends on the actual raw scores, which PARCC doesn’t release.)

A student’s final performance level on the test, then, is computed as some function of all the different standards that are assessed on the test. In other words, if a student demonstrates understanding at Level 4 on the multiplication and addition of rational and irrational numbers but Level 2 on the graphing of quadratic functions, his score might be Level 3, although the amount of weight different standards have on the final score isn’t equal in all cases.

There are also sweeping assumptions incorporated into the score, including a consistent test environment and good student effort on the test. I’ve observed high school and elementary students taking the PARCC tests, and I can say, without a moment of hesitation, that student effort is much greater at the elementary school level than it is in high school, where students know the test doesn’t count toward their grade, college prospects, or job prospects.

That’s why when I read editorials in newspapers, like this one in the *Baltimore Sun* that say the result that fewer than half of Maryland’s high school students are meeting expectations is news, I’m completely speechless at the lack of understanding.

**In an editorial entitled “PARCC results reveal a sobering truth,” the Sun writes:**

The truth is, [test scores] tell us something we need to know, something that has been obscured by years of steadily rising scores on our old tests and seemingly perennial top rankings for Maryland’s schools. Disturbing percentages of Maryland’s high school graduates are unprepared for college or careers, and that unwelcome distinction falls much more heavily on African-Americans, Hispanics, special education students and English language learners.

First question: Who needs to be sobered up and what have they been drinking? The “truth” is, good editors, it’s not news that special education students perform worse on standardized tests than kids who don’t have special needs. Yet the *Sun* hams it up with an open jaw. And is the news organization actually suggesting we should expect students who don’t speak English as well as the average American kid to perform as well on a test that measures English?

Or that African-Americans should, on average, rise above the higher percentage who live below the poverty line to perform with their more affluent peers, on average, on a standardized test on which results are more closely correlated with the socioeconomic level of the students than with anything resembling a learning standard? We weren’t born yesterday, but maybe the editors of the *Sun* have been drinking too much Kool-Aid. A good report of this information would at least consider the poor performance of students at Title I schools for Algebra I:

Subgroup | % at Level 1 | % at Level 2 |
---|---|---|

All Students | 13.0 | 28.6 |

White | 6.2 | 19.6 |

African-Americans | 22.0 | 39.9 |

Title I | 28.5 | 40.9 |

At the state level, which is the level of the current data release, information about Title I schools tells us more than any information about race, special education, or limited English proficiency. We can see that more kids performed poorly on the algebra I test at Title I schools than in the state overall. Maybe we can look into why kids at Title I schools aren’t getting as strong an education in algebra I as other kids in the state, but it is also entirely possible that there are other reasons that have nothing to do with the quality of education schools and teachers are providing why the performance at Title I schools is worse, on average, than that in the state.

The high school results on the PARCC test, because of the assumptions built into the score, don’t mean more than half of Maryland’s students and three-fourths of the state’s African-American students are unprepared for college or careers. Being “prepared” for college or career is such a big concept that the PARCC tests measure, at most, a tiny part of it—and probably not even that since the test is still undergoing huge development efforts.

Rather, the results just mean, without considering any other data points, that students did poorly on the PARCC tests, administered here for the first time amidst a battlefield of politics and community members telling kids who have no vested interest in the results that testing is evil and they should opt out, if only Maryland allowed opt-outs.

So, let’s get our story straight: The results are out. We’ve got work to do. Teachers, parents, and reporters won’t learn anything from these results they didn’t already know. When do we get rid of this blasted law and get back to learning in our classrooms instead of working toward a test?

MSDE will release more detailed data and information for elementary and middle schools before the end of November. We’ll cover that as well, but let’s not pretend these data points tell us anything the highly qualified and caring teachers of Maryland didn’t already know.

Handy little dashboard that I produced that I thought might help folks who are interested in this topic: https://public.tableau.com/profile/andrew.hammond#!/vizhome/PARCC_2015/PARCC2015-MeetorExceed