How we score schools

SchoolScope scores every California public school 0–100 from official CDE data. Elementary and middle schools: 38% exceeded standard, 20% met+exceeded,27% growth, 5% absenteeism, 5% suspension,5% ELPAC proficiency. High schools replace growth with graduation rate (25%) and college readiness (20%).

Last updated: March 2026 · Data source: California Department of Education · About SchoolScope

Three lenses, one score

The Scope Score draws from two of three lenses we use to understand a school. The third lens provides context you can see — but it never feeds the score.

Lenses 1 and 2 feed the Scope Score. Lens 3 is displayed on every school profile so you can see the full context — but it is never part of the score. Why? Read our full reasoning below.

How is the Scope Score calculated?

Every scored school gets a single Scope Score from 0–100, built from weighted signals. The weights differ by school level because the available data — and what matters most — changes as students progress.

When a signal is unavailable for a school (e.g., no chronic absenteeism data), the weight is redistributed proportionally across available signals. Schools without sufficient data are not ranked.

How scores are computed

Each dimension is measured relative to the California state average for schools at the same level. The size of the gap matters — not just whether a school is above or below average.

A school with 70% of students exceeding the standard ranks meaningfully higher than one at 55% exceeded, even though both are well above average. The score reflects how far above average each school lands, not just its position in a sorted list.

The final Scope Score is centered around 50 — the state average school scores roughly 50. Most schools fall between 30 and 70. A score of 75 means the school is well above average across its weighted dimensions; a score of 35 means it consistently sits below average. The scale is consistent, so a 10-point difference at the top of the range means the same thing as a 10-point difference in the middle.

This approach is different from systems that rank schools by position (1st, 2nd, 3rd) and then compress everything to a 1–10 scale. With position-based ranking, the school ranked 1st and the school ranked 200th might be nearly identical in actual performance — the rank just doesn't tell you that. Our scoring preserves the real distance between schools.

What each metric means

% Exceeded Standard — the ceiling

Most rating sites report "% proficient" — students who met or exceeded the standard. This hides a crucial difference.

Consider two schools, both "70% proficient":

• School A: 50% exceeded, 20% met — a school actively pushing students to excel
• School B: 10% exceeded, 60% met — a school where most students barely clear the bar

The exceeded rate is one of the strongest signals we've found of a school that challenges students, which is why it gets the highest academic weight in our Scope Score formula. Read: Why the exceeded vs. met split changes everything →

% Met + Exceeded — the floor

This is baseline proficiency: what percentage of students clear the bar? It ensures we don't ignore schools that reliably get students to grade level, even if they aren't pushing many past it. The ceiling and floor together paint the full picture.

Growth Trajectory — two signals, not one

We measure growth two ways, using whichever is available:

Cohort tracking (preferred): We follow the same school's cohort across years — for example, tracking 2023's 3rd graders to see how they perform as 5th graders in 2025. This uses SBAC scale scores, which are designed for cross-year, cross-grade comparison. 98% of elementary schools and 93% of middle schools have cohort growth data.

Cross-sectional (fallback): When historical data isn't available, we compare this year's 5th graders to this year's 3rd graders using scale scores. This is less precise than cohort tracking because it compares different students, but it's still informative.

Both approaches are measured relative to the state average for schools at the same level — how far above or below average matters, not just rank order. A school whose cohort gained 60 scale score points when the state average is 50 gets meaningfully more credit than one that gained 52.

High schools don't have a growth component because California only tests at grade 11 — there's no earlier grade to compare against.

Chronic Absenteeism — a culture signal

Chronic absenteeism measures the percentage of students who miss 10% or more of school days. It's one of the strongest predictors of academic outcomes and reflects school culture, family engagement, and community stability.

It's inverted in our formula — lower absenteeism means a higher score. It carries more weight at elementary and middle (5%) than high school (5%) because engagement patterns are established early.

Suspension Rate — discipline philosophy

Suspension rate measures the percentage of students suspended at least once. High suspension rates often indicate a discipline-heavy culture rather than a supportive one, and disproportionately affect underserved communities. Also inverted — lower is better.

Graduation Rate — the biggest high school weight

Graduation rate carries 25% of the high school Scope Score — the single largest weight — because it's the most consequential outcome. A high school that doesn't graduate its students isn't delivering on its most basic promise, regardless of test scores. We use the Adjusted Cohort Graduation Rate (ACGR) published by the California Department of Education.

College Readiness (CCI) — what happens after graduation

California's College/Career Indicator (CCI) measures the percentage of graduates who completed the requirements for college or career readiness — things like completing A-G coursework, passing AP exams, earning career technical education certifications, or meeting other state-defined criteria. At 20% of the high school Scope Score, it rewards schools that prepare students for what comes next, not just for the diploma.

ELPAC Proficiency — English Learner progress as a school climate signal

ELPAC is California's English Language Proficiency Assessments for California. It measures the English proficiency of students identified as English Learners across four domains: listening, speaking, reading, and writing. We use the percentage of English Learner students at a school who reach Level 4, "Well Developed" — the highest proficiency tier.

We classify ELPAC in the School Climate lens rather than the Academic Performance lens. The reasoning: a school where English Learner students are progressing toward full proficiency is delivering on its instructional mission for every student it serves — that's a climate and culture signal, not just a test score.

ELPAC carries 5% of the Scope Score at all three levels. When a school has no English Learner students, this weight redistributes proportionally across the remaining dimensions. No school is penalized for not having an EL population — we simply work with the data that exists.

Data sources

All data is publicly available. No restrictions on commercial use. Our current test year is 2025.

• CAASPP Smarter Balanced AssessmentELA and Math test scores, grades 3–8 and 11
• CDE SACS Current Expense of EducationDistrict current expense per student, annual — context only, never scored
• CDE Chronic Absenteeism DataSchool-level chronic absenteeism rates
• CDE Suspension DataSchool-level suspension rates
• CDE Graduation Rate (ACGR)Adjusted Cohort Graduation Rate by school
• CDE College/Career Indicator (CCI)College and career readiness rates
• CDE College-Going RatePercentage of graduates enrolling in postsecondary education
• NCES Common Core of Data (CCD)Enrollment counts, student-teacher ratios, free/reduced lunch
• CDE Census Day EnrollmentRace/ethnicity and gender breakdown by school (GN_F / GN_M / GN_X reporting categories)
• NCES School Attendance Boundary Survey (SABS)Attendance boundaries used for feeder pattern estimation

The sources above are the nine inputs to the Scope Score formula. For our complete data source inventory across all 20+ public datasets — including the UC admissions, College Board AP authorization, IB programme, and federal CRDC data shown on school profiles as context (never scored), plus district pages and community context — see our About page.

How we correct for biases

Raw data has built-in biases that make comparisons unfair. We apply four corrections before computing any ranking. We make judgment calls. Here's exactly what they are.

These four corrections mean our rankings are more stable year-over-year, fairer to schools with harder-to-test populations, and more honest about uncertainty than raw-average approaches.

How rankings work

The scoring pipeline runs in this order: grade-subject normalization → cohort or cross-sectional growth → Bayesian smoothing → weighted composite → Scope Score (centered at 50). See bias corrections and the formula table for details.

• Schools are ranked within their own level — elementary vs. elementary, middle vs. middle, high vs. high. Rankings are never mixed across levels.
• Percentile tells you what percentage of same-level schools score below this one. A school at the 90th percentile outscores 90% of schools at its level statewide.
• The maximum percentile displayed is 99th. We don't show 100th percentile.
• Schools without sufficient data to compute a Scope Score are not ranked. No data means no score — we don't manufacture numbers.

School Archetypes

Every California public school receives one of seven archetypes based on its performance profile. Archetypes describe a school's character — what makes it distinctive — rather than reducing it to a single number. They are classified automatically from Scope Score data at render time.

Classification is evaluated in priority order — a school that qualifies as both a Finisher and a Balanced school will be labeled as a Finisher. Private schools do not receive archetypes because they lack CAASPP test data.

What each archetype means for you

A label is a starting point, not a verdict. Here's what each one should make you ask.

• ⭐ High Ceiling: Ask what happens for a kid who shows up behind — a high ceiling for the top can sit right next to thin support at the bottom, and the score alone won't show you which.
• 🚀 Growth Engine: Ask how the school tracks each kid's growth year to year, and what happens when a kid stalls instead of gaining.
• 💪 Strong All-Around: Ask about what we don't score — class size, electives, the feel of a classroom — strong numbers everywhere on paper don't mean every classroom feels the same.
• 📈 On the Rise: Ask how many years the improvement has held, not just the latest one — one strong cohort and a real turnaround can look identical in a single year of data.
• 🏫 Solid Base: Ask what the school does for a kid who's already meeting the standard with room to spare — steady proficiency across the board doesn't tell you whether the strongest kids are being pushed further.
• 🤝 Culture First: Ask whether the low absenteeism and suspension numbers come from what the school does or from the community it serves — both can produce the same numbers, and only a visit tells you which.
• 🌱 Building Momentum: This label means the data doesn't clearly sort the school one way or the other — ask what the school is proud of, and what it's working on.

Performance Bands

Rather than asking parents to interpret a precise number, we group Scope Scores into four performance bands. The band tells you what matters: is this school in the right zone?

Precise Scope Scores (0–100) are always shown alongside band labels for transparency. The band communicates the zone; the number provides the detail.

How we write our verdicts

Every school profile opens with a written verdict — a headline, a summary, and an honest read of the strongest signal and the biggest catch in that school's data. Those sentences are not written by hand for each school, and they are not generated by a chatbot free-associating. They are assembled from a fixed catalog of sentence templates by deterministic rules: the same school always produces the same verdict — there is no randomness, and every number in a verdict appears next to the comparison that gives it meaning.

How a verdict is chosen

• Signals and catches. We compute the school's standing on each scored dimension — exceeded rate, proficiency, growth, attendance, suspension, ELPAC, plus graduation and college readiness for high schools — and on subgroup performance versus the state. Standings well above typical become candidate signals; standings meaningfully behind become candidate catches.
• A school is never introduced by its subgroup. Subgroup performance is often the most informative number on a page, and it stays prominent — in the equity panel, in the opening paragraphs, in the appendix. It does not open the page. A headline that introduces a school by its low-income or English-learner students makes a demographic into that school's public identity, which is the failure we criticise rating sites for. A school is introduced by what it does for all of its students, and the subgroup number is the evidence for that claim rather than a substitute for it. Changed 8 August 2026: before that date, a subgroup clause opened roughly three quarters of the profiles that used the tension framing.
• The headline pairs the strongest eligible signal with the most important catch. That is the tension framing on profiles: the thing the school does unusually well, next to the thing worth asking about. Schools without a meaningful catch get a straightforward verdict — and we say when we looked for a catch and didn't find one.
• Low-scoring schools never get tension drama. Profiles in the Needs Support band open with context, a true positive fact where the data shows one, and where to start a conversation — not a takedown.

The guardrails

• Observational language only. Catches state what the data shows (“proficiency falls from grade 3 to grade 5”), never why. Cross-sectional test data cannot establish causes, so our sentences do not claim them.
• Every catch carries a hedge and ends at your feet. Each verdict with a catch includes one of a fixed set of hedges — “the data can't say why,” “worth asking about on a tour,” “a school visit will tell you more than this number” — and a concrete question to ask the principal.
• Small groups are suppressed. Grade-level claims require at least 15 tested students and subgroup claims at least 30, mirroring the state's own privacy suppression. When data is missing, the verdict says so instead of guessing.

If you believe a verdict misreads your school, report it — a person reads every report, and when something is wrong we fix the rule, not just the page.

Contextual Signals (Not Scored)

Some data tells you important things about a school but shouldn't be reduced to a score. We display these as Community Profile context — visible on school profiles but never factored into the Scope Score.

Physical Fitness (FITNESSGRAM)

California mandates the Physical Fitness Test for grades 5, 7, and 9. School-level PFT results were publicly downloadable through 2018-19, but CDE discontinued research-file publication after a Title 5 regulation change. If CDE reinstates downloadable PFT data, we plan to show the percentage of students meeting all six fitness standards (Healthy Fitness Zone) as a whole-child indicator — context only, never scored, because fitness outcomes correlate with school resources and neighborhood wealth.

Teacher Stability

We show average years of teaching experience and the percentage of first-year and second-year teachers at each school. Veteran teachers are associated with better outcomes — but scoring teacher stability would penalize schools in hard-to-staff areas that serve the students who need the most support. Source: CDE Staff Experience data.

College/Career Readiness (High School)

For high schools, we show the number of students qualifying through AP exams and CTE pathway completion from the CDE College/Career Indicator (CCI). These indicate academic rigor and vocational options — but wealthier schools naturally have higher participation, so we show this as context rather than score it. Source: CDE Dashboard CCI data.

Limitations

We believe in honesty about what our scores can and can't tell you.

• Test scores are one lens, not the whole picture. They correlate with many things we care about but don't capture everything that makes a school great.
• We can't measure teacher quality, school culture, or creative programs. The data doesn't exist in any public dataset. A school with an incredible arts program or a transformative principal won't show that in our score.
• Our Scope Score is our best attempt, not objective truth. Reasonable people could weight these signals differently. The research behind each dimension tells us it matters — it doesn't tell us exactly how much. The weights are our editorial judgment, made in the open. See what the research says for the studies, and where they fall short.
• High school Scope Scores lack growth data. California only tests at grade 11, so there's no earlier tested grade to measure progress against. This means the high school formula relies more heavily on outcomes (graduation, college readiness).
• Cohort growth tracks school-level aggregates, not individual students. We follow the same school's cohort across years (e.g., 2023's 3rd graders as 2025's 5th graders), but kids transfer in and out. Demographic shifts or a particularly strong incoming class can still affect results. For the 2% of elementary and 7% of middle schools without prior-year data, we fall back to cross-sectional comparison (this year's 5th graders vs. this year's 3rd graders), which is less precise.
• Feeder patterns are estimated. We use NCES boundary data to estimate which elementary schools feed into which middle and high schools. These are not official district assignments and may not reflect transfers, magnet programs, or recent rezoning.
• Small schools may have volatile scores. Bayesian smoothing reduces this, but schools with very few tested students still carry more uncertainty than large schools. Our smoothing is honest about that uncertainty rather than hiding it.
• We don't adjust for demographics or income. We show you the raw signal. Context (free/reduced lunch %, community demographics) is presented alongside scores, not baked into them.
• Two climate signals still carry a demographic echo. Raw chronic absenteeism and raw ELPAC proficiency both partly measure which students a school enrolls, not just what the school does for them. See what the research says below for the studies. We weight both at the level shown in the table above because of that concession. The fix that would clean them up further — comparing a school only to others serving similar communities — isn't built yet.

What the research says

Every dimension in our formula is here because some study found it connects to things that matter for kids — not because it's easy to measure. Here's the actual research behind each one, including where it falls short.

Growth

A school's average score mostly reflects which kids walk in the door. Growth reflects what the school does with them. Researchers who studied real school lottery outcomes found that ratings built on growth predict a school's actual effect on students far better than ratings built on proficiency levels (Angrist, Hull, Pathak & Walters, American Economic Review: Insights, 2024). Separately, growth rates carry almost no relationship to a district's poverty rate — a school in a low-income neighborhood can post real growth (Reardon, Stanford Education Data Archive; Urban Institute, 2020). That's the research behind why we raised growth's weight in August 2026.

Exceeded vs. met

Reporting one “percent proficient” number throws away real information. A school where 60% of kids exceed the standard is a different school from one where 60% just barely clear it, even if both report “80% met or exceeded.” Researcher Andrew Ho called percent-proficient a “funhouse mirror” that distorts real differences between schools (Educational Researcher, 2008). Splitting the two levels apart recovers what a single number erases.

Suspensions

Kids assigned to a higher-suspension school go on to worse adult outcomes — more arrests, less schooling completed — even after accounting for who attends (Bacher-Hicks, Billings & Deming, American Economic Journal: Policy, 2024). The study covers one metro area, so we hold the exact size of the effect loosely, but the research design isolates a school's discipline practice from who's enrolled there, which is why we treat suspension rate as a real signal.

Graduation rate (high school)

In one large study of Texas charter schools, campuses that raised test scores didn't produce a measurable earnings bump years later— but campuses that raised graduation rates did (Dobbie & Fryer, 2020). That's one study in one state, and we say so. It's still a big part of why graduation carries the largest single weight in our high school formula.

Absenteeism

Attendance is one of the most reliable predictors of graduation we've found in the research, sometimes stronger than test scores (Allensworth & Easton, University of Chicago Consortium on School Research). Here's the honest catch: a school's raw absenteeism rate is shaped mostly by a family's health, transportation, and housing stability (Liu, Fordham Institute, 2023). A high rate can mean a struggling school, or it can mean a school serving a lot of families dealing with things no school can fix. We weight it at the level shown in the table above because of that concession. It's the dimension we most want to improve by comparing schools only to others serving similar communities, and that comparison isn't built yet.

ELPAC proficiency

No study we've found validates a school's raw English Learner proficiency rate, on its own, as a measure of school quality. We weight it anyway because schools demonstrably shape how fast English Learners reach proficiency, through program design and teacher stability (Umansky & Reardon, American Educational Research Journal, 2014; U.S. Government Accountability Office, 2024). But the rate is also heavily shaped by things a school doesn't control — how recently a student arrived, and how much English they started with (REL Northwest, 2016). A school that enrolls a lot of newcomers will score lower here through no fault of its teaching. We're naming that plainly instead of hiding it.

What no public data can see

The single biggest school factor we found in the research isn't in our formula, or in anyone else's public data: the individual teacher in front of your kid. One study tracked more than a million students through tax records and found that a single year with a better teacher raised college attendance and adult earnings years later (Chetty, Friedman & Rockoff, American Economic Review, 2014). No public dataset publishes teacher-level results, and even trained researchers can't reliably judge a teacher by watching one lesson (the MET Project, 23,000+ videotaped lessons). That's why every school profile ends with a question to ask on your visit, instead of a claim about teaching we can't back up.

Reasonable people can weigh this research differently than we did. The studies above tell us these dimensions matter. They don't tell us exactly how much — the weights in the table above are our judgment call, made in the open so you can check it against your own.

Data confidence

Not all numbers on school profiles have the same certainty. We label data by its source so you know what you're looking at.

• DirectOfficial data for this school, published by CDE or NCES. Test scores, absenteeism, suspension rates, graduation rates, enrollment. These are the most reliable inputs.
• DerivedComputed by SchoolScope from official inputs — growth trajectory, Scope Score, percentile rank. We show our work on the methodology page and in school profile score factors.
• DistrictDistrict-level data applied to a school when school-specific data isn't available. Less precise but still informative. Always labeled when shown.

Frequently asked questions

Further reading

• What School Rankings Get Wrong — And What We're Trying to Fix — our manifesto on why we built SchoolScope
• Why the Exceeded vs. Met Split Changes Everything — a deep dive into the most important distinction other ratings miss
• Why we show diversity but don’t score it — our full reasoning on demographics as context, not score input
• What the research says — the named studies behind every scored dimension, including where they fall short
• Explore all California schools — see our methodology in action
• All blog posts — more data-driven analysis of California school performance
• About SchoolScope — who we are, our independence, our data governance standards, and why we built this