North Carolina education leaders are looking at changing how they calculate one of the two main measures of school quality — the latest effort to reevaluate the system that determines how a school is perceived publicly and whether a school receives extra oversight or assistance.
Each fall, the North Carolina Department of Public Instruction releases test score data for every school — an accountability approach that also includes what are called “growth scores,” which are designed to indicate whether schools met or exceeded expectations for their standardized test scores.
Other WRAL Top Stories
It’s an important metric used to help decide a school’s performance letter grade and whether a school is low-performing, a designation that can subject a school to higher scrutiny and also provide a school with additional assistance and flexibility through school improvement programs, such as the Restart Schools program. It’s also a designation that matters for many parents wondering about the quality of their child’s school.
But leaders and experts say the growth-score metric isn’t well understood by parents or the broader school community, and it doesn’t really show how much a school’s test scores “grew” or improved, if at all. Instead, it uses an algorithm to predict students’ test scores and then measures their actual scores against those predictions. Then, it calculates how every school did against those predictions and plots them on a bell curve. The result, every year, is that roughly three-quarters of all schools meet or exceed growth
In other words, “growth” is based on relative performance against a prediction model, not year-over-year changes in proficiency.
“Growth is actually far more complicated than I think folks think,” said Michael Maher, the state Department of Public Instruction’s chief accountability officer.
DPI currently assigns letter grades to schools — much like a student would get for a test. About 80% of that comes from how many students test on grade level, and the other 20% comes from the “growth” score.
The grades, required by federal law, are intended to help the public know how well a school is doing and to hold schools accountable for their students’ performance.
For years, education leaders have recommended changes to the school performance letter grades to state lawmakers, including highlighting the growth measure more. Lawmakers haven’t taken up the proposals, though a bipartisan legislative committee recommended making changes in 2024.
Superintendent of Public Instruction Mo Green and the State Board of Education have started a task force to again examine school accountability.
Those leaders primarily think North Carolina’s school performance letter grades, which are established in state law, are based on too few measures, with too much emphasis on proficiency rates that often correlate with a school’s poverty level.
Now, education leaders are rethinking the popular growth metric — the only metric used alongside the proficiency rates to calculate the performance letter grades — entirely, though they could need General Assembly help to make any major changes.
How NC calculates test score growth
“Growth” scores are intended to provide an extra layer to school accountability beyond the raw test scores themselves. They attempt to level all schools, which could face different socioeconomic or funding challenges, by measuring how their students’ test results changed from the year before.
In North Carolina, calculating growth starts by measuring how well an individual student did against how well an algorithm predicted they would do. The idea is to identify effective teachers even in schools where grade-level proficiency rates are lower to start with.
To present a school-level score, all of the school’s student results are combined. Then, the school is placed on a bell curve next to all other schools. Where the school falls on that curve determines whether it didn’t meet growth expectations, met growth expectations or exceeded them. Because it’s on a bell curve, roughly the same percentages of schools fall into each of those categories every year, regardless of how much better or worse schools did overall on actually passing the tests.
State leaders still consider it a good metric for measuring teacher quality, but they’re now questioning whether they should change how it’s calculated at the school level, in part because of that curve.
“It's kind of fairer to schools, because you're compared to … what's happening in that year across the state,” said Peter Halpin, an associate professor in the School of Education at the University of North Carolina at Chapel Hill. “But it's less informative to parents because parents don't often have these questions about relative growth.”
In a news release last week, the Wake County Public School System noted that 79.4% of its schools met or exceeded growth during the 2025-26 school year. Durham Public Schools noted that 38 of its 55 schools met or exceeded growth. Chapel-Hill Carrboro City Schools, Chatham County Schools and Orange County Schools all highlighted how many schools met or exceeded growth.
In any given year, more than 40% of schools meet growth, nearly 30% of schools exceed growth, and nearly 30% don’t meet growth, with some variance by a few percentage points. While schools celebrated widespread gains in proficiency rates last week, the share of schools meeting or exceeding growth didn’t really budge.
Maher thinks the current growth model — the proprietary Education Value-Added Assessment System (EVAAS), provided by SAS Institute — is a good measure for teachers but doesn’t work as well when applied at the school level. DPI isn’t looking to drop EVAAS for teachers but is wondering if there’s another way to measure growth at the school level and is working with SAS and an accountability task force on that.
“It is, I think, a more challenging measure to use for [school] accountability,” Maher said. EVAAS was the best tool when the state needed to come up with a growth measure for schools, but Maher notes most other states are using other approaches.
SAS’s EVAAS tool is still the growth measurement tool required by North Carolina law.
“There are opportunities for us now to look at some different tools when we think about growth,” Maher said. A model that looked at each student’s growth, rather than placing them on a curve, would make more sense to people, including parents, he said.
“We really are leaning to this idea of individual student projections and then measuring growth based on the individual student,” Maher said. “And then we can do some things like aggregate that up and roll it up into a district level, school level, things like that.”
How other states calculate test score growth
About a dozen states use a form of EVAAS or another value-added model to measure school growth, according to the PRiME Center at St. Louis University. SAS tailors EVAAS models to different customers based on their preferences.
Another 23 states use student growth percentiles, which measure students’ growth against students who performed similarly the year before. That approach levels out extremes, where lower-performing students can make bigger gains than higher-performing ones. But Halpin notes that it still places students on a curve, just several different ones.
“Both of those have the disadvantage of not really communicating to parents what they might intuitively think of as growth,” Halpin said.
Another dozen states use growth-to-proficiency models, which measure a student’s progress toward meeting proficiency. It doesn’t account for external factors.
Three states use gain score measures that more straightforwardly show year-to-year gains, not adjusted for any external factors
A criticism of those two approaches is that failing to adjust for external factors makes it tough to fairly compare schools that have very different student bodies.
DPI and the State Board of Education is discussing possible changes during monthly meetings of the State Board’s Task Force on Accountability for Public School Units. SAS will present some new ideas for calculating growth during the task force’s upcoming September meeting, Maher said.
A SAS representative, when reached by WRAL for comment, said it would defer to DPI on policy matters.