More than half of all Colorado students in third through eighth grades continue to fall below state expectations in reading, writing and math, according to results of state tests students took this spring. That has been the case since Colorado switched to more rigorous tests four years ago.

In literacy, 44.5 percent of students in those grades statewide met expectations. In math, 34.1 percent did. It is difficult to compare this year’s scores to scores from previous years because of changes in requirements for which students take which tests.

However, the percentage of students meeting expectations in literacy went up at least slightly this year in every grade, three through eight. The math results were mixed.

Results in both subjects show a persistent and troubling reality mirrored across the country: White and Asian students continue to score higher than black and Hispanic students, and students from middle- and high-income families outperform students from low-income families.

The gaps between students from higher- and lower-income families are about 30 percentage points. For example, 45 percent of sixth-graders from middle- and high-income families met expectations on the state math test, but only 14 percent of sixth-graders from low-income families did.

“As a society and a state, this is unacceptable,” Colorado Education Commissioner Katy Anthes said in a statement. “And every effort must continue to be made to reverse this course.”

About 550,000 students across Colorado were tested in the spring. Students in third through eighth grades took literacy and math tests that are Colorado’s modified version of the PARCC tests. (The state refers to the tests as the Colorado Measures of Academic Success, or CMAS, tests.) High school students took well-known college entrance exams: Ninth- and 10th-graders took the PSAT, and 11th-graders took the SAT.

The percentage of students meeting expectations on the literacy and math tests varied by grade. In third grade, for example, 40 percent of students met expectations on the literacy test and 39 percent met expectations on the math test. Both represent a 2 percentage-point increase from 2015, the first year Colorado gave the PARCC tests.

Joyce Zurkowski, who oversees testing for the state education department, said that while the upward trends are encouraging, “the change is not happening as quickly as we’d hope.”

At the high school level, this spring marked the second year Colorado 11th-graders took the SAT, and the third year 10th-graders took the PSAT. Ninth-graders also took the PSAT this year.

Scores on those exams were similar to last year, with Colorado students continuing to do better than national averages. For example, Colorado 11th-graders scored an average of 513 on the SAT reading and writing section, and 501 on the math. The average score of students who took the SAT on the same day nationwide was 497 in reading and writing, and 489 in math.

As in previous years, the data shows girls in grades three through eight scored better on state literacy tests than did boys. The gap between the genders increased the older students got: 54 percent of eighth-grade girls met expectations in literacy, while only 34 percent of boys did.

The reverse was true in math, at least in the lower grades. Boys in grades three through seven scored higher than girls, but eighth-grade girls did slightly better than eighth-grade boys.

Girls also scored higher than boys on the PSAT and SAT, though by 11th grade the gap narrowed to a single point: The average score for girls was 1,015; for boys, it was 1,014.

Some of the biggest gaps are between students with and without disabilities. For example, just 6 percent of eighth-graders with disabilities met expectations in literacy, compared with 48 percent of eighth-graders without disabilities, a whopping 42-point difference.

Measuring academic progressThe state also calculates the progress students make on the tests year to year. This calculation, known as the “median growth percentile,” measures how much students improve in an academic year compared with other students with similar scores in the previous year.

The state – and many school districts – consider this measurement just as important, if not more important, than raw test scores, which often correlate to students’ level of societal privilege. Growth scores, on the other hand, measure the improvement students make in a year – and provide insight into how effective their teachers and schools are in teaching them.

Because of that, growth scores make up a big portion of the ratings the state gives to schools and districts. Low-rated schools and districts are subject to state sanctions.

A student’s growth is ranked on a scale of 1 to 99. A score of 99 means a student did better on the test than 99 percent of students who scored similarly to him the year before.

Students who score above 50 are considered to have made more than a year’s worth of academic progress in a year’s time, whereas students who score below 50 are considered to have made less than a year’s worth of progress.

Statewide data shows white students, students from higher-income families and students without disabilities had growth scores above 50. Students of color, students from low-income families and students with disabilities had scores below 50.

For example, elementary students who do not qualify for subsidized lunches had a growth score of 54 in both literacy and math. Elementary students who do qualify had a growth score of 47. Having a lower growth score means it may be harder for those students to reach grade level.

The state also compares the scores of students learning English as a second language to the scores of students who are not. When the data is cut in that way, the differences are minimal in elementary and middle school. For example, the overall growth score in math for elementary-aged English learners was 49, while the score for non-English learners was 51.

However, the difference in growth scores between those two groups was bigger in high school – a trend that holds true for several other student groups, as well.

Difficult to discernThe reason educators and state officials focus on how different groups of students do on the tests is to ensure schools are educating all students – not just those with the most privilege.

Of all the groups, it can be most difficult to tell how well schools are serving students learning English as a second language. That is because of the way the state categorizes students.

English language learners who attain fluency score very well on the state tests, especially in literacy. But whether they score on par with – or perhaps even better than – native English speakers remains an open question because that category includes other students as well.

That’s not the only reason it can be hard to draw conclusions about the academic progress of different student groups. Colorado has strict student privacy rules that, for example, obscure the growth scores of any group with fewer than 20 students, officials said.

Education advocacy groups have called on the state to release more information that would provide a fuller picture of whether schools and districts are serving all students well.

Participation rates upColorado was once a hotbed of the testing opt-out movement, with tens of thousands of fed-up parents excusing their children from taking the state assessments. But participation has been rising, and it was up again this past spring for students in grades three through 10.

It is likely that part of the increase is because of the passage of a bill in 2015 paring back the amount of time Colorado students spend taking standardized tests.

But there was another factor this year, too: Zurkowski attributed a bump in ninth-grade participation, in particular, to a switch in tests. Ninth-graders took the PSAT this past spring instead of the PARCC tests. Whereas just 76 percent of Colorado ninth-graders participated in the PARCC literacy test last year, nearly 94 percent of ninth-graders took the PSAT, a preparatory test for college-entrance exams and a qualifying test for National Merit scholarships.

“I believe students and parents are recognizing the relevance of the PSAT test,” Zurkowski said.

The state is set to make another switch next year. Instead of administering the PARCC tests to students in grades three through eight, Colorado is developing its own literacy and math tests.

But state officials said they don’t anticipate a significant change in participation or the ability to compare student scores from year to year. The Colorado-developed test questions will be based on the same academic standards as the PARCC questions, Zurkowski said.

Chalkbeat is a nonprofit news site covering educational change in public schools.