Today is Part 3 of our special three-part newsblast stemming from NJ’s Joint Committee on Public Schools hearing on student achievement and school performance. You can read Part 1 here and dive into Part 2 here.
On Tuesday, we started with the Good. Yesterday highlighted the Bad. Today we get into….
The Ugly
Some testimony is a difference of opinion. Some is a difference of fact. At the October 13 hearing, superintendent Dr. Jamil Maroun told the committee that the NJSLA and NJGPA are graded on a curve. In his telling, test makers sort statewide scores into a bell curve, then draw "arbitrary" cut lines that guarantee roughly 40% of students fail no matter what they know. His conclusion: headlines saying New Jersey graduates can't do basic math are "fundamentally flawed."
If that were true, it would be a scandal. It isn't true.
Norm-referenced vs. criterion-referenced: the 30-second version
A norm-referenced test ranks students against each other. Think percentiles: being in the 80th percentile means you beat 80% of test takers, and by definition only 20% can be above you.
A criterion-referenced test measures students against a fixed standard. If every student masters the material, every student can pass. If none do, none pass.
The NJSLA is criterion-referenced. Its performance levels are set by panels of New Jersey educators who match test items to written descriptions of what students at each grade should know. NAEP works the same way. The National Center for Education Statistics directly states that NAEP uses a criterion-referencing approach.
Of course, scores across a million students tend to cluster in the middle. That's true of height, too. A bell-shaped spread of scores is not the same as grading on a curve. What matters is where the bar is set — and on the NJSLA, it's set against the standards, not against other kids.
Exhibit A: proficiency went up
If a curve locked 40% of students below the line every year, proficiency could never move much. Yet here is what the state's own data show for PARCC/NJSLA, grades 3–10, from 2015 to 2019:
Year | ELA % met/exceeded | Math % met/exceeded |
|---|
2015 | 47.2 | 37.5 |
2016 | 51.5 | 42.1 |
2017 | 54.4 | 42.9 |
2018 | 56.6 | 44.7 |
2019 | 57.8 | 44.0 |
ELA proficiency rose more than 10 points in four years. That's tens of thousands of additional students clearing a bar that didn't move. A curve doesn't do that. Learning does.
Exhibit B: then it fell
Starting in 2019 and then fueled by COVID, the same tests showed the drop everyone knows happened. NJ students’ Math and ELA proficiencies fell almost 10 percentage points between 2019 and 2022, and remained substantially below 2019 levels by 2025. Again, a curve would have held the share steady. It didn't, because students lost ground over this time, something empirically studied and widely reported on.
Dr. Maroun's own figures prove the point. He read out the share of students below proficiency since 2015 — 39%, 41%, 42%, 33%, 32% and so on. Those numbers swing by as much as 10 points. A fixed curve doesn't swing.
Exhibit C: the Nation's Report Card
NAEP tells the same story. New Jersey's 8th-grade math average climbed from 281 in 2003 to 296 in 2013. It slipped to 292 by 2019, then crashed to 281 in 2022 and sat at 282 in 2024. The share of NJ 8th graders proficient in math fell from 44% in 2019 to 33% in 2022, with only a partial rebound to 37% in 2024. Fourth-grade math dropped from 246 to 239 over the pandemic and sits at 240 today — below the 249 New Jersey posted in 2007, as Assemblywoman Dawn Fantasia reminded the committee.
Scores rise, scores fall, and the reasons are real: instruction, attendance, a pandemic. That's what a criterion-referenced test is supposed to capture.
About that cut score
Dr. Maroun pointed out that when New Jersey lowered the passing score from 750 to 725 (on the NJGPA), the pass rate jumped about 10 points overnight. True — and it proves the opposite of his claim. Lower the bar, more kids clear it. The real lesson is that where Trenton sets the bar matters enormously — which is exactly why parents, and all of us, should be engaged when performance levels are changed.
So, while disagreements are a bedrock of our democracy, they should always be based on facts and evidence. Especially when discussing the future of our state’s most precious asset, our children.
