Tomorrow, December 4th, the New Jersey State Board of Education will release statewide scores from the 2023-2024 New Jersey Student Learning Assessments (NJSLA, the tests NJ uses to be compliant with federal education law). During the public meeting the State Board will look through two different lenses to examine how NJ students are progressing: whether they’re meeting grade-level expectations (also called “Performance Levels” or “proficiency”) and whether that proficiency is growing at an appropriate pace, also called “Student Growth Percentiles.” These two very different ways of quantifying students progress can be confusing to residents.

You’re not alone! Remember, in 2017 President Donald Trump’s nominee for Education Secretary, Betsy DeVos, was stymied when then-Senator Al Franken asked her to explain the difference between proficiency (another way of saying Performance Level) and growth.  But if one wants to get a relatively unvarnished look at the quality of our state school system, it is important to understand the ways the NJ DOE gauges student learning: 

Let’s dig in. 

How did we come up with these “Performance Levels?”

The New Jersey Student Learning Assessments are given in the spring in third through eighth grade and once in high school. Students take these tests in English Language Arts (reading) and math, as well as periodic science tests. Each testing year students receive a Performance Level based on whether they’ve mastered expectations delineated for their grade in the NJ Student Learning Standards. NJSLA questions, which are the same for all students (with alternatives for some multilingual students and students with disabilities) test whether students have mastered the necessary skills to succeed in the next grade. If they do, then they are deemed “proficient.”

Who decides on the K-12 standards school districts must adopt and the state must assess? The State Board of Education approves these standards, which are driven by analyses of what a typical student needs to master in order to successfully advance through their K-12 school years and be prepared for colleges and/or careers.

(Click for NJ’s math standards and ELA standards.)

Who decides what “proficient” means? Again, the State Board, which has approved the designations, based on a score range from 650 to 850. 

790-850 is Level 5, or “Exceeded Expectations.” 

750-789 is Level 4, or “Met Expectations.”

725–749 is Level 3, or “Approached Expectations.”

700-749 is  Level 2, or “Partially Met Expectations.”

650-699 is  Level 1, or “Did Not Yet Meet Expectations.”

The DOE database called “School Performance Reports” (see here for an Explainer) includes proficiency levels school-by-school, grade-by-grade, and district-by-district. As per federal law, these levels are broken down by “subgroups”: African-American students, Hispanic students, white students, low-income students, etc. 

(One side note: The NJ State Board has, over time, lowered scaled scores so that more students either meet or exceed expectations. Many other states have done so as well, which some say is a well-meaning attempt to grant grace to struggling students and others say is a false gloss on declining student achievement due to loosened federal accountability laws  since the end of the unforgiving No Child Left Behind. Think of it like the way airlines currently pad arrival times so passengers think their plane is on time but it is actually late.)

What is Student Growth and how is that different from meeting grade-level expectations?

In addition to looking at scale scores (which the DOE says “provides the best generalized information about overall performance”), the Board and the public will also look at the other way the DOE measures progress for elementary and middle school students: Student Growth Percentiles, or SGPs. While Performance Levels are mathematical conversions of how many questions students answer correctly, SGPs look at a student’s academic progress over time compared to academic peers. New Jersey introduced this method when the federal education law, currently called Every Student Succeeds Act, was modified to grant more flexibility to states, a backlash to No Child Left Behind. In short, SGPs calculate a  trend line across a series of data points to estimate a student’s growth rate. 

Another way to phrase the difference between proficiency and growth is absolute academic performance vs. relative academic performance. Proficiency levels measure what a student knows at the end of the school year; SGPs measure how much they’ve learned.

SGP scores are plotted on a range from 1 to 99. Below 35 is “low growth,” 35-65 is “typical growth,” and 65-99 is “high growth,” in all cases compared to academic peers.

But SGPs are more than a number; we can, to some extent, extrapolate from them to determine instructional quality in a school, district, and state. Let’s say a student has been getting high scores on NJSLA and, thus, “Exceeded Expectations.” In other words, this student is at the ceiling of the tested material. Now let’s look at this student’s Growth Percentile: It’s a 34, which means this student’s learning curve is slower than her academic peers. This can be a red flag for parents and educators: Is the student bored? Are there problems at home? Is there an undiagnosed learning disability? Does the student need more challenging material? Is the school/district appropriately challenging high-performing students?

Indeed during Senator Franken’s famous conversation with Education Secretary-to-be Betsy DeVos, he argued that just looking at proficiency encourages teachers to ignore students who, while scoring highly on the proficiency metric, aren’t making adequate growth compared to their peers. [See: educational triage.]

On the other hand, if a student has low proficiency levels but high growth, he or she is (relatively) doing really well.

Let’s look at a school. The DOE database shows that Robert Treat Academy, a public charter school in Newark, has an overall SGP of 60 in reading (ELA) and 64 in math, just a smidgeon below “high growth.” In comparison, Newark Public Schools’ overall SGP is 48 in reading and 47 in math, a bit lower. Robert Treat is serving its students better than NPS, regardless of where they start out. (These are data based on 2022-2023, the most recent available in the DOE database.)

As a state (again from 2022-23), NJ students scored a median 50 SGP, firmly within the “typical growth” band, despite efforts (and tons of federal money) to accelerate student learning in the aftermath of Covid learning loss. 

What’s the take-away?

Of course, no one test can pinpoint the academic performance of a student/school/district/state Yet NJSLA results give us important information, not only for parents but also for state policymakers. For example, low reading proficiency rates (and growth trends) in NJSLA reading tests motivated the State Legislature to mandate teachers use the science of literacy when teaching reading, coalesced the groups behind the NJ Tutoring Corps, and brought more heft to arguments (currently in litigation) that tying district assignments to ZIP codes/housing costs disenfranchises low-income students. 

With so much at stake, it is worth taking the time to understand how well New Jersey is serving its 1.3 million schoolchildren.