What School Rankings Get Wrong — And What We Can (and Can't) Fix

We built SchoolScope’s score, so we’ll open ours first. It should face the same questions as anyone else’s, including who pays for it and where it falls short.

What goes into a school rating like ours?

Ratings like ours start the same way. Someone picks a few results from public files and decides how much each one counts. Ours, the Scope Score, uses California Department of Education files. For elementary and middle schools, it combines test results, growth between two grades, chronic absence, suspensions and English learners’ ELPAC results. High schools swap growth for graduation and college readiness.

Each result is compared with other schools at the same level before it’s added in. Small schools’ results are pulled toward the middle, so one unusual class swings them less. Open the number and you can see every part.

The scored layers fall into two of our three lenses. Academic Performance holds the test results, growth, and for high schools, graduation and college readiness. School Climate holds attendance, suspensions and English learners’ results. The third lens, Community Profile, covers who attends, the gaps between student groups and spending. It sits beside the score on a school’s profile, but none of it goes into the number.

Not every school has every result. When a school has no usable growth or ELPAC result, that result’s share is spread across the school’s other results instead of counting as zero. The “Take out a result” buttons in the figure show how, and how many schools were scored that way.

On a profile, the first things you see are the school’s percentile and a band word: Strong, Solid, Developing or Needs Support. Both come from where the school ranks among California schools at its level. We lead with them because they’re the part a parent can act on. We wouldn’t worry about a gap of a point or two between two scores.

The band matters for another reason. A score places a school only among California schools at its own level. An elementary school’s score and a high school’s score come from two separate comparisons. So we’d read the band and percentile beside the score, and compare a school only with schools at its own level.

What is a Scope Score made of?

The Scope Score is SchoolScope’s rating for California public and charter schools, on a scale from 50 to 100, built from California Department of Education public files. On that scale, 75 is the typical California school at its level and 100 is the best result in the state. For elementary and middle schools, it combines the average share of test results at the top level, the average share meeting the state standard or higher, growth for one class from one grade to a later one (families move, so not exactly the same children), chronic absence, suspensions, and English learners’ results on the ELPAC test. High schools replace growth with graduation and the state’s College/Career Indicator. Each school is compared only with schools at the same level. Its band (Strong, Solid, Developing or Needs Support) comes from its statewide percentile among them. Continuation and some alternative schools are ranked only against schools like them. When a school has no usable growth or ELPAC result, that result’s weight is spread across its other results instead of counting as zero. Private schools are listed without a score, since there’s no state testing data for them. Schools whose enrollment is assigned, such as court schools, get no score either. The weights, as of August 2026, are published on the methodology page.

Who pays for school ratings, and does it matter?

Once you know what’s in a number, it’s fair to ask who makes it and how they’re paid. Here is what GreatSchools and Niche say in public.

GreatSchools is a nonprofit. Its 2023 tax return reports $9.07 million in revenue, including $5.72 million in contributions and grants and about $1.97 million labelled “Licensing.” The return doesn’t say what that licensing covers. Its ratings also travel with home listings. When we checked on October 12, 2026, a Zillow home listing showed its nearby schools with a “GreatSchools Rating,” and Zillow’s home search let you filter by that rating. Zillow said in a 2013 announcement that its school ratings came from GreatSchools. A 2019 Chalkbeat report said the ratings appeared in home listings on Zillow, Realtor.com and Redfin. When a rating sits next to a price, we think it’s easy to read it as part of what the house is worth.

Niche is a private company. Its partnerships page offers schools a premium profile and “enrollment marketing solutions.” It also invites real estate sites to “add Niche school grades to your real estate content.” So a school you’re comparing may also be a Niche customer. That alone doesn’t show that payment changes a grade.

Being paid some other way wouldn’t settle the question either. A site funded by parents could still profit from their worry. So instead of asking you to trust how we’re paid, here are five commitments you can check:

• Schools cannot buy a rank, badge, or place in a list.
• Factual corrections can change the data; payment cannot change a score.
• Commercial relationships do not enter the scoring code.
• Our methodology and sources are public.
• We document corrections when we get something wrong.

The other services also do things we don’t. GreatSchools covers all 50 states. Niche has parent and student reviews, and it rates private schools. SchoolScope lists private schools but never scores them, because there’s no state testing data for them.

Like GreatSchools, we compare a school only with schools in its own state. Niche first ranks each school’s test results within its state, then compares schools across states. Each state gives its own tests, so a California score and an Ohio score aren’t on the same ruler.

What does a single number hide?

Whoever builds it, a rating squeezes several different stories into one number.

Four of them are worth pulling apart.

The first is the top of the test. “Met or exceeded” lumps together students who reached the state standard and students who reached the highest level. California renamed the levels starting with 2025–26 results, so the top two are now Advanced (exceeded) and Proficient (met). Two schools with the same combined rate can have very different shares at the top. Our exceeded-versus-met guide shows real schools side by side, and the label guide maps the old names to the new ones.

The second is growth against level. Test levels tell you where students are. For an elementary school, our growth measure compares its grade 3 class with its grade 5 class two years later. For a middle school, it’s grade 6 and grade 8. It uses scale scores, the numbers behind the level labels. Families move in and out, so the two groups aren’t exactly the same children. Growth shows which way a school’s results moved; it can’t tell you what caused the change. California’s own growth model follows matched students in grades 4 to 8. We show it on profiles where it’s available and leave it out of the score, and our methodology explains why. When growth and test levels disagree, that’s worth asking about.

The third is attendance and discipline. Chronic absence means missing at least 10% of enrolled school days. A high rate often reflects the needs of the families a school serves: health, housing, a long ride to school. On its own, it doesn’t show that the school caused the absences. A suspension rate tells you what share of students were suspended at least once that year. It can’t tell you whether the school’s response was fair. Both are small parts of our score, and both work better as questions than as verdicts.

The fourth is how English learners do on the ELPAC. ELPAC is California’s English test for students who are learning English. The share of a school’s English learners at the top level depends partly on how recently those children started learning English. Under California’s current guidance, a Summative ELPAC Overall Level 4 is now the sole criterion for reclassifying a student as fluent. For the Alternate ELPAC, it’s Level 3. We count the school’s share as a small part of the score, and we don’t treat it as proof of teaching quality.

How much does a rating reflect the families a school serves?

A lot, and ours is no exception. In California, schools with fewer students from low-income families tend to post higher test results, and test results are the biggest part of our score at every level. So the Scope Score is closely tied to family income, which we measure by the share of students who qualify for free or reduced-price meals.

Other ratings have shown the same pattern. A 2019 Chalkbeat analysis found that GreatSchools generally gave schools serving mostly low-income students significantly lower ratings than schools serving wealthier families. GreatSchools has changed its formula since then, so that finding describes the ratings as they were in 2019.

Growth is the part of our score least tied to income. In August 2026, we cut its share anyway, and the last section explains that trade. Our methodology says what we have and haven’t done to correct for income.

For a parent, this changes how to read a high or low score. A score tells you how a school’s students did. It can’t separate what the school added from what children brought with them. That’s why a profile shows who attends on the same page as the score.

What should you check that a rating can’t see?

Some of what matters most never reaches a state data file. These are the four empty layers at the bottom of the figure.

Start with your child’s teacher. A school average can’t tell you how the teacher your child will actually have teaches. Even watching that teacher once tells you less than it feels like. In a study for the Gates Foundation’s Measures of Effective Teaching project, principals and peer teachers in Hillsborough County, Florida, scored videos of lessons by 67 teachers. The researchers found that “a single observation by a single observer is a fairly unreliable estimate of a teacher’s practice.” They added that “having more than one observer really does matter.” We read that as a reason to hear from several families whose children had the teacher before trusting any one impression, including your own.

Then there are the arts, clubs and daily life. A music program, a quiet place to eat lunch, or the club your child can’t stop talking about can matter a great deal and never move our score.

Special-education services are harder still to see. California publishes test results for students with disabilities where the group is big enough, and we show them. A test score isn’t an IEP, though. It can’t tell you whether the services in your child’s Individualized Education Program happen in every classroom, or whether accommodations follow your child from room to room. Ask the school how it delivers those services, and talk with families who use them.

The last is belonging and fit. A school that suits a child who needs structure might not suit a child who needs room to wander. High scores can come with a high-pressure culture, and middling scores can come with a calm classroom that’s right for your child. A visit and a few conversations show you the fit far better than a number can.

Niche’s high school rankings do add parent and student surveys about teachers and clubs. Those give a school-wide view, not a view of your child’s classroom. None of these four is in our score or in the state files it’s built from, so plan to check them yourself.

How should we use ratings well?

With those limits in mind, a rating works best for narrowing a list and sharpening your questions.

First, compare schools at the same level. Elementary, middle and high school scores answer different questions, and we never mix them in a ranking. Also check whether a school is ranked within its own group. Continuation schools and some alternative schools get a score. It’s ranked only against schools like them, not against all California schools at that level. Their profiles say so. Schools whose enrollment is assigned, such as court schools, show their results with no score at all.

Second, look at several years, by percentile and band. One year can be unusual. Scores built with different formulas don’t make one clean trend, and older and current high school scores can use different inputs. So compare where a school ranked each year, rather than whether its score went up.

Third, look at the schools around it. The same score can mean different things in different districts. The two schools below show the same score on their profiles and have the same band. One sits near the top of its district, and the other near the bottom.

If you’re choosing between addresses, look at the district’s other schools as well as the one score.

Fourth, bring one concrete question. If absences are high, find out who helps families struggling to get children to school. Where growth and test levels disagree, ask what changed and how many children the numbers cover. And if suspensions are high, it’s fair to ask what staff try before suspending a child.

Last, trust your tour for the things a tour can show you. You’ll see how adults talk to children, how the office treats you and how new families are welcomed. Those are real, and they’re about fit. Teaching is harder to judge in one visit. In the video study, one trained observer watching one lesson got a shaky read on how that teacher usually teaches.

How do we hold SchoolScope to account?

We’ve asked you to read every rating with care, so here is where ours is weakest.

Of the parts of our score we could check, growth is the least steady. In August 2026, we compared each school’s growth for two different groups of children. The two agreed at about 0.56–0.58 for elementary and middle schools. A school’s share of test results at the top level, checked the same way, agreed at about 0.97. A correlation measures how closely two sets of numbers move together, and 1 means they move in lockstep. We can’t tell from that gap how much of growth is noise and how much is real change, but it’s reason enough not to let one year of growth decide anything.

So we cut growth’s share of the score. At the old weight, more than a quarter of elementary schools could land in a different band depending on which group of children we happened to measure. The cut made bands steadier. It also made our score track family income a little more closely. We took that trade on purpose: a band that shifts with the luck of one class doesn’t help a parent. Reasonable people could weigh it the other way.

We also haven’t shown that our growth measure predicts how children do later, such as in high school or after graduation.

What we can fix is the record. When we get something wrong, we correct it and log the change in our methodology’s change log. If our numbers don’t match what you know about a school, tell us. It may mean we’re missing context, or that we made a mistake.

Corrections, August 14, 2026: this post previously said California doesn’t publish student-level growth data. It also described competitor methodologies as not fully public, which is wrong — both publish weights. And it argued our incentives were clean because parents fund us; we’ve replaced that with five commitments you can check.

Corrections, August 15, 2026, after a line-by-line source check of this post. We began importing California’s growth model and showing available results, so the text above says that rather than “it’s on our list.” We had described our own growth measure two different ways in the same article; the scored number is the cohort comparison, and only that description survives. We claimed growth and exceeded rates were “close to independent” at r = +0.21 — that’s a weak positive link, not independence, and the “94% unexplained” line that followed it was doing work correlation can’t do. We said roughly half of a year of growth is noise; a correlation can’t make that split, so we cut the claim and kept the actionable part. We said a high absence rate means the school “has a problem” — the evidence says a high rate signals a high-need student body at least as often, and the section now says that. Two spending paragraphs cited an NCES file that does not contain the figures we quoted. We corrected them using our own SACS calculations; those paragraphs were removed in the October shortening. The classroom-observation study we cited covers early-childhood centers for ages 0 to 6, not K-12 tours, and we replaced it with one that observed school-age classrooms. That study discussion was removed in the October shortening. We asserted Niche has a firewall between payment and grades; we can’t find Niche saying that, so the claim is gone and not replaced with its opposite. And we described GreatSchools’ licensing revenue as the reason its scores appear on Zillow — the tax return doesn’t say that, so we now report only what it does say.

Correction, October 12, 2026: the October 11 version said one lesson “cannot establish teaching quality.” The study behind that point is narrower, so the post now quotes it: in Hillsborough County, Florida, a single observation by a single observer was “a fairly unreliable estimate of a teacher’s practice.” An earlier version had also credited that study with “more than 23,000 videotaped lessons,” a figure the paper doesn’t contain; it doesn’t appear here. This version brings the classroom-observation study back, with its population named.

See what your schools really look like