“Graduates” here means the students California counts as high school completers that year. That’s regular-diploma graduates plus students who finished another way, such as passing the CHSPE proficiency exam or earning a GED or an adult-education diploma.
In SchoolScope’s matched-pairs analysis of California Department of Education college-going records and grade 11 CAASPP results, 252 of 1,726 eligible high schools formed 126 pairs for the class of 2023, using 2022 tests. The schools in each pair had similar shares of students meeting or exceeding the standard, meal eligibility and ZIP-code income, but Advanced shares at least 5 points apart. The school with the higher Advanced share had a UC enrollment rate 2.18 percentage points higher on average (nominal 95% range 0.71 to 3.60), and the class of 2020 pointed the same way, at 1.55 points. In these main pairs, college-going overall was inconclusive in both class years. When the class-of-2023 pairs also had to match on the share of nearby adults with at least a bachelor’s degree and on student racial makeup, the UC gap fell to −0.16 points across 58 pairs. This is our analysis, not a CDE finding. It compares schools’ rates without following any child, so it doesn’t tell you where any one student will go. For the college question, we’d look at each school’s own college record, on its SchoolScope profile.
What did we compare?
We started with California high schools that had the records we needed. That meant grade 11 English and math results, college destinations for the next year’s graduating class, and a few facts about the school and its ZIP code. Those school and ZIP-code facts come from recent records, not from the test years. For the class of 2023 that was 1,726 schools, using their 2022 tests. For the class of 2020 it was 1,691 schools, using 2019 tests. The link is school to school: a school’s tests go with its next graduating class, and no student is followed from one file to the other.
Advanced is the top of the four levels on California’s grade-level tests. The historical tests in this study call it Standard Exceeded; Advanced is the current name for that same level. Proficient (met) is the next level down, called Standard Met in those files. A school’s Advanced share, in this study, is our average of its English and math percentages at that level. It’s two subject rates averaged together, so a junior who reached Advanced in both subjects counts in both.
Then we paired schools that look alike on paper. The two schools in each pair had nearly the same share of students meeting or exceeding the standard, but Advanced shares at least 5 points apart. Overall results had to sit within 1 point of each other. The share eligible for free or reduced-price meals had to be within 8 points. Median household income in the ZIP code had to be within 12% of the lower-Advanced school’s. Only 252 of the 1,726 schools ended up in pairs in 2023, making 126 pairs. In 2020, 254 of 1,691 did, making 127. That’s about 15% each year, and the paired schools had higher test results and college-going rates than the schools left out.
We report every gap in percentage points. If one school sends 20% of its graduates to UC and its partner sends 18%, the gap is 2 points: about 2 more students in every 100.
Matching on a few measures still leaves plenty we can’t see, such as family plans and money, or which courses each school offers. One difference we could see stayed unbalanced. In the 2023 pairs, the share of adults 25 and older with at least a bachelor’s degree in the higher-Advanced school’s ZIP code averaged 3.55 points higher than around its partner. In 2020 it was 4.44 points.
Did higher-Advanced schools send more students to UC, or to college at all?
In the main comparison, the higher-Advanced school’s UC enrollment was higher on average in both class years, by about 2 points in 2023 and 1.5 in 2020, while college-going overall was inconclusive.
For the class of 2023, the UC gap was 2.18 points, with a nominal 95% range of 0.71 to 3.60. For 2020 it was 1.55 points (0.31 to 2.83). The range is our uncertainty band around each gap. We call it nominal because it rests on our method’s assumptions and comes with no guarantee.
Both class years point the same way. After we allowed for testing four destinations at once, the 2023 UC result cleared the usual statistical threshold. The 2020 result missed it (the values are under How we checked).
College-going overall is the other half of the question, and here the result was inconclusive. The higher-Advanced school’s rate was 1.70 points higher in 2023 and 2.31 in 2020, but both ranges dipped below zero (−1.36 to 4.70, and −0.18 to 4.78). The data can’t separate a small gap from no gap, so it isn’t evidence that the two kinds of school send the same share to college. In two of the 2020 alternative pairings, the overall college range did sit above zero: leaving out charters (+3.24 points, 0.56 to 6.03) and matching on nearby bachelor’s degrees alone (+3.37, 0.34 to 6.34). So we’d treat this as unsettled.
Enrollment in the California Community Colleges system was inconclusive too: −1.60 points in 2023 and +0.80 in 2020, with both ranges crossing zero. Our fourth measure, the mixed group, adds up UC, Cal State, in-state private colleges and out-of-state four-year colleges. The in-state private part includes two-year colleges, so this mix isn’t a count of four-year enrollment. Its gap was 2.86 points in 2023 (−0.57 to 6.25) and 1.62 in 2020 (−1.26 to 4.46), so it was inconclusive in both years.
These four measures overlap. A student recorded at UC counts in the UC rate, in the mixed group and in college-going overall.
126 pairs from 1,726 eligible California high schools
In the main comparison, only UC’s range sat fully above zero
Each bar is the average gap between paired schools for the class of 2023: the school with more students at Advanced minus the one with fewer. The pale band is the likely range. The next chart shows how the UC gap changes when the pairs are matched more closely.
Gap in percentage points in school enrollment rates
Only UC’s range sits fully above zero, and it shrinks when the pairs are matched more closely (next chart). The ranges are rough guides, not exact limits. UC students are also counted in the mixed group, so the rows overlap.
See the numbers as a table
| Outcome | Gap | Likely range (95%) | Verdict | Evidence |
|---|---|---|---|---|
| Mixed group | +2.9 | −0.6 to +6.3 | Main comparison inconclusive | includes in-state private two- and four-year institutions |
| UC | +2.2 | +0.7 to +3.6 | Changes under stricter neighborhood matching | above zero in the main comparison; close to zero when pairs are matched by neighborhood |
| Any college | +1.7 | −1.4 to +4.7 | Main comparison inconclusive | range crosses zero |
| California community colleges | −1.6 | −4.7 to +1.5 | Main comparison inconclusive | range crosses zero |
Does the gap hold up when the schools are more alike?
The main pairs still differed on one thing we could see, as the first section showed. The higher-Advanced school’s ZIP code had more adults with at least a bachelor’s degree, by 3.55 points on average in 2023 and 4.44 in 2020. So we added two more rules. One, the college-graduates rule, kept the share of adults with at least a bachelor’s degree in the two schools’ ZIP codes within 12 points. The other kept the schools’ recorded racial makeup close. Their Hispanic, white, Black and Asian student shares could differ by at most 45 points in total, or 30 for a tighter version. Each version is a new match from all the eligible schools, so it gives a new, smaller set of pairs.
When the pairs had to meet both rules, the UC gap fell toward zero in both years.
It reached about zero under the strictest rules, while the pairs stayed about as far apart on Advanced share, on average.
Does the UC gap last when the paired schools are more alike?
Once the pairs also match on nearby college graduates and racial makeup, the class-of-2023 UC gap is −0.16 points, and its range includes zero (−1.91 to 1.58, 58 pairs). In the main match it was +2.18 (0.71 to 3.60, 126 pairs).
Average UC gap, higher-Advanced school minus partner Nominal 95% range No gap
Add one rule at a time
Each step is a new match with one more rule and fewer pairs. In all four, the paired schools stay 6.3 to 6.7 points apart on Advanced share, on average.
- Main match+2.18 0.71 to 3.60126 pairs · whole range above 0
- + Nearby college graduates+1.32 −0.34 to 2.9791 pairs · range includes 0
- + Racial makeup too−0.16 −1.91 to 1.5858 pairs · range includes 0
- + Tighter racial-makeup limit−0.35 −2.58 to 1.8543 pairs · range includes 0
Other ways to cut the pairs
The first row matches on racial makeup without the college-graduates rule. The other two keep the main rules and limit which schools can pair.
- Racial makeup rule on its own+0.67 −0.87 to 2.1678 pairs · range includes 0
- No charter schools+2.27 0.68 to 3.8370 pairs · whole range above 0
- 200 or more completers+2.53 −0.24 to 5.3132 pairs · range includes 0
A gap between school rates isn’t an effect of a school, and it isn’t a forecast for a child.
Open the full table, both class years
Same-district pairs are listed here and not drawn. There were 5 in the class of 2023 and 12 in the class of 2020, and the 2023 range runs from −4.97 to +10.17, past both ends of the plot’s scale.
| Match | Pairs | UC gap (points) | Nominal 95% range | Range vs zero | Advanced gap (points) | Gap in nearby college graduates (points) |
|---|---|---|---|---|---|---|
| Main match | 126 | +2.18 | 0.71 to 3.60 | Above 0 | 6.52 | 3.55 |
| + Nearby college graduates | 91 | +1.32 | −0.34 to 2.97 | Includes 0 | 6.63 | 0.39 |
| + Racial makeup too | 58 | −0.16 | −1.91 to 1.58 | Includes 0 | 6.67 | 0.42 |
| + Tighter racial-makeup limit | 43 | −0.35 | −2.58 to 1.85 | Includes 0 | 6.28 | 0.87 |
| Racial makeup rule on its own | 78 | +0.67 | −0.87 to 2.16 | Includes 0 | 6.57 | 2.95 |
| No charter schools | 70 | +2.27 | 0.68 to 3.83 | Above 0 | 6.61 | 4.36 |
| 200 or more completers | 32 | +2.53 | −0.24 to 5.31 | Includes 0 | 6.43 | 10.97 |
| Same district only (not drawn) | 5 | +2.91 | −4.97 to 10.17 | Includes 0 | 5.98 | 5.58 |
| Match | Pairs | UC gap (points) | Nominal 95% range | Range vs zero | Advanced gap (points) | Gap in nearby college graduates (points) |
|---|---|---|---|---|---|---|
| Main match | 127 | +1.55 | 0.31 to 2.83 | Above 0 | 7.18 | 4.44 |
| + Nearby college graduates | 94 | +1.53 | 0.19 to 2.95 | Above 0 | 7.31 | 2.16 |
| + Racial makeup too | 58 | +0.84 | −0.65 to 2.29 | Includes 0 | 7.10 | 1.90 |
| + Tighter racial-makeup limit | 37 | +0.12 | −1.46 to 1.71 | Includes 0 | 7.25 | 1.95 |
| Racial makeup rule on its own | 84 | +1.55 | 0.20 to 2.91 | Above 0 | 6.95 | 3.88 |
| No charter schools | 76 | +2.38 | 1.03 to 3.83 | Above 0 | 7.39 | 4.40 |
| 200 or more completers | 39 | +4.34 | 2.46 to 6.36 | Above 0 | 7.49 | 6.69 |
| Same district only (not drawn) | 12 | −0.41 | −5.38 to 4.59 | Includes 0 | 6.36 | 4.61 |
For the class of 2023, the gap went from 2.18 points in the main match to 1.32 with the college-graduates rule, −0.16 with both rules and −0.35 with the tighter limit. The racial-makeup rule on its own gave 0.67. All four of those ranges include zero. In each of those designs, the pairs stayed 6.3 to 6.7 points apart on Advanced share, on average. In the class of 2020, the gap lasted longer. Each rule alone left the gap about where it was, at 1.53 and 1.55 points, and both together brought it to 0.84, or 0.12 with the tighter limit. The ranges mostly widen as the pairs thin out, from 126 pairs to 43 in 2023, so the strictest rows are less precise as well as more alike.
Not every way of cutting the pairs shrinks the gap. Leaving out charter schools gave 2.27 points in 2023 and 2.38 in 2020. Keeping only schools with 200 or more completers gave 2.53 in 2023 and 4.34 in 2020, the biggest gap we found, with a range of 2.46 to 6.36. Those large-school pairs were also furthest apart on nearby bachelor’s degrees, by 10.97 and 6.69 points. We can put those two numbers side by side, but nothing here tells us one caused the other. We also tried pairing schools inside the same district. Only 5 pairs qualified in 2023 and 12 in 2020, and the five 2023 pairs are especially thin evidence: their range runs from −4.97 to +10.17. They’re left off the plot and listed in its table.
We also checked how much of the main gap the leftover degree difference could cover in a simple model. Within the main pairs, we estimated how much the UC gap moves with each point of degree gap, allowing for the Advanced gap and a rough racial-makeup measure. Then we multiplied by the average degree imbalance. For 2023 that’s 126 pairs and 3.55 points, which comes to about 0.50 of the 2.18-point gap. For 2020 it’s 127 pairs and 4.44 points, about 0.34 of the 1.55-point gap. So in this model, the degree imbalance covers only a little over a fifth of the gap. Matching again on degrees and racial makeup moved the result much further, and it also changed which schools were compared. We can’t tell from these data which picture is closer to the truth.
Our read: the main UC gap moves too much with the choice of pairs. We wouldn’t treat it as something a school’s Advanced share tells you about college.
What should this change about how we choose a high school?
We’d use a school’s Advanced share to ask about challenge, and the school’s own college records for the college question.
The Advanced share tells you what share of its juniors with scores reached the top level of a grade-level test, averaged across English and math. For your own child, grade 11 levels have a concrete use at California State University (CSU). Students choose whether to release them to CSU, which uses them only after admission, to place students in first-year English and math courses. Our CAASPP guide explains what’s different in grade 11.
The college record itself is on a high school’s SchoolScope profile, in its “After graduation” section, when we have the data. “Where graduates went” gives the share who enrolled in college within 12 months and how graduates split across the destination types, including those with no recorded enrollment. The University of California block shows how many students applied, were admitted and enrolled, as counts. “Completed UC/CSU courses” is the share of graduates who finished the A–G list, the set of high-school courses UC and Cal State require for admission. Where it’s reported, “College / career Prepared” is California’s own readiness measure. Further down, “What could my child study here?” lists the AP courses the school is authorized to teach, plus IB and course counts where we have them. You can find a school’s profile and compare two of them.
Most of those numbers describe past graduating classes, and an AP authorization doesn’t confirm this year’s schedule. On a visit, we’d ask what comes after the standard math and English sequence, and how teachers decide who’s ready for honors or AP. A related question is what happens when a student needs more challenge than the class in front of them. We’d also want to know how a ninth or tenth grader finds out whether they’re on track for A–G.
Say your child’s school has an Advanced share at least 5 points below a school across town with about the same overall results, meal eligibility and ZIP-code income. In our main class-of-2023 pairs, schools like yours sent about 2 fewer graduates in every 100 to UC, a gap that shrank toward zero once the pairs also matched on nearby bachelor’s degrees and racial makeup. The gap in college-going overall was inconclusive, so the main comparison gives no clear sign that schools like yours send fewer graduates to college. Our explainer on the split shows how two schools with the same met-or-exceeded rate can differ in Advanced share.
For a ninth grader, the A–G line may be the most useful college number on any profile, because A–G completion is something their own course schedule can still change.
How we checked
Open the method, the full tables and the correction note
This is SchoolScope’s analysis, not a CDE published finding. CDE supplies the underlying school test results and college-destination fields. We chose the inclusion rules, pairs and statistical checks.
Who and when. We linked schools, not individual students. The class-of-2023 destination file is paired with 2022 grade-11 tests; class-of-2020 destinations are paired with 2019 tests. The analysis uses ACS 2020–24 neighborhood estimates and school covariates stored in the snapshot. The stored meal-eligibility measure comes from NCES Common Core of Data enrichment; recorded race/ethnicity shares come from CDE enrollment data. Those are contextual measures, not historical family records for the students in either completion year. We checked the score and destination fields against retained CDE files. We have not independently checked the raw neighborhood and current school-demographic sources.
CDE’s College-Going Rate concerns high school completers, a broader category than diploma graduates. Enrollment means enrolling at a college within 12 months of completion. The analysis is not a comparison of a particular autumn’s incoming class.
CDE’s report definitions assign someone with multiple college enrollments by institution type using their longest cumulative enrollment during the reporting window. These are recorded destinations, not counts of everyone who attended each type at any time. Privacy-blocked National Student Clearinghouse records can lower recorded college-going rates; we have not estimated a school-specific correction for that missing enrollment. Here, community college means the California Community College system, not out-of-state or private two-year institutions.
The test measure averages English and math subject percentages without weighting them by subject count. Its minimum of 50 scored student-subject records sums the English and math Students with Scores counts; the same student can appear in both. It is not a minimum of 50 distinct juniors. Schools also need usable destination results and covariates. We exclude ZIP income reported at the top-coded value, which means “$250,000 or more,” rather than an exact income.
What the intervals mean. A nominal 95% bootstrap interval summarizes the uncertainty from resampling these matched pairs under our procedure. These intervals condition on the selected pairs; we do not repeat the school-matching step inside each resample. They do not account for every source of bias or establish a causal effect. We have not established that this procedure gives 95% coverage for these selected schools or modeled dependence from shared district and neighborhood context. An interval crossing zero leaves a zero average gap compatible with this procedure.
Main comparisons, both completion years
All gaps and intervals are percentage points, higher-Advanced school minus lower-Advanced school. UC is also part of the mixed group; the destination categories are not four separate groups of students.
| Outcome | 2023 gap | 2023 nominal 95% interval | 2020 gap | 2020 nominal 95% interval |
|---|---|---|---|---|
| UC | +2.18 | 0.71 to 3.60 | +1.55 | 0.31 to 2.83 |
| Mixed group | +2.86 | −0.57 to 6.25 | +1.62 | −1.26 to 4.46 |
| Any college | +1.70 | −1.36 to 4.70 | +2.31 | −0.18 to 4.78 |
| California community colleges | −1.60 | −4.67 to 1.51 | +0.80 | −1.96 to 3.58 |
We use a paired sign-flip test, which compares the observed average gap with gaps obtained by randomly reversing its direction within pairs. The sign-flip test assumes that reversing directions within pairs is appropriate under its null; matching schools does not establish that assumption. We also adjust for testing four outcomes within each design using the Benjamini–Hochberg procedure. These adjusted q-values do not establish false-discovery control for this study’s outcome dependence. A p-value is not the probability that a finding is due to luck or that no difference exists.
| Outcome | 2023 p | 2023 adjusted q | 2020 p | 2020 adjusted q |
|---|---|---|---|---|
| UC | 0.0036 | 0.0144 | 0.0173 | 0.0692 |
| Mixed group | 0.0982 | 0.1964 | 0.2828 | 0.3770 |
| Any college | 0.2762 | 0.3147 | 0.0692 | 0.1384 |
| California community colleges | 0.3147 | 0.3147 | 0.5685 | 0.5685 |
At an adjusted q threshold of 0.05, the main UC q-value is below the threshold in 2023 but not in 2020. The adjustment is within each four-outcome design; it does not cover the full set of exploratory matching choices. Neither the threshold nor the positive UC intervals make these school comparisons causal.
Which schools made it into pairs?
Only 252 of 1,726 eligible schools matched in 2023, and 254 of 1,691 in 2020: about 15% in each year. They are a selected sample, with higher test results and college-going rates than the schools left unmatched.
| Measure | 2023 matched | 2023 unmatched | 2020 matched | 2020 unmatched |
|---|---|---|---|---|
| Overall proficiency % | 44.0 | 33.5 | 50.5 | 35.4 |
| Advanced share % | 19.4 | 14.3 | 23.2 | 14.9 |
| Free/reduced meal eligibility % | 60.4 | 67.7 | 59.4 | 68.1 |
| ZIP median income | $102,506 | $96,829 | $103,390 | $97,557 |
| College-going rate % | 67.4 | 57.7 | 67.3 | 52.5 |
| Completers per school | 247.1 | 237.3 | 266.6 | 236.8 |
| Charter schools % | 30.6 | 21.0 | 28.0 | 19.6 |
These figures describe the schools selected by this design, not California high schools generally. Some schools can appear in both completion years; this is not two independent samples of schools.
What changes when we change the pairs?
The table gives gaps followed by nominal 95% bootstrap intervals. Neighborhood education matching limits the difference in the share of adults with a bachelor’s degree. Racial-composition matching limits differences in four recorded student race/ethnicity shares. Both are school or area measures; neither tells us an individual child’s background.
| 2023 match | UC gap [interval] | Mixed group gap [interval] | Pairs |
|---|---|---|---|
| Main design | +2.18 [0.71, 3.60] | +2.86 [−0.57, 6.25] | 126 |
| + neighborhood education | +1.32 [−0.34, 2.97] | +1.31 [−2.79, 5.39] | 91 |
| + racial composition | +0.67 [−0.87, 2.16] | +1.52 [−2.69, 5.56] | 78 |
| + both | −0.16 [−1.91, 1.58] | +1.43 [−3.37, 6.10] | 58 |
| + both, tighter | −0.35 [−2.58, 1.85] | +2.42 [−3.00, 7.70] | 43 |
| No charters | +2.27 [0.68, 3.83] | +3.77 [0.46, 6.95] | 70 |
| At least 200 completers | +2.53 [−0.24, 5.31] | +3.07 [−1.44, 7.45] | 32 |
| Same district | +2.91 [−4.97, 10.17] | −2.25 [−21.89, 17.39] | 5 |
| 2020 match | UC gap [interval] | Mixed group gap [interval] | Pairs |
|---|---|---|---|
| Main design | +1.55 [0.31, 2.83] | +1.62 [−1.26, 4.46] | 127 |
| + neighborhood education | +1.53 [0.19, 2.95] | +1.90 [−1.27, 5.06] | 94 |
| + racial composition | +1.55 [0.20, 2.91] | +3.64 [0.18, 7.06] | 84 |
| + both | +0.84 [−0.65, 2.29] | +2.32 [−1.60, 6.06] | 58 |
| + both, tighter | +0.12 [−1.46, 1.71] | +2.12 [−2.46, 7.01] | 37 |
| No charters | +2.38 [1.03, 3.83] | +2.48 [−0.53, 5.54] | 76 |
| At least 200 completers | +4.34 [2.46, 6.36] | +3.91 [0.43, 7.28] | 39 |
| Same district | −0.41 [−5.38, 4.59] | +0.80 [−6.96, 8.01] | 12 |
The 2023 neighborhood-matched UC estimates approach zero. In 2020 the strictest neighborhood match also approaches zero, but excluding charters or choosing larger completion groups produces bigger positive estimates. Different choices select different schools and change the uncertainty. The five same-district pairs in 2023 are especially thin evidence.
A separate exploratory regression within the original pairs tells a less decisive story. Multiplying its neighborhood-education coefficient by the average remaining imbalance gives about 0.50 points in 2023 and 0.34 in 2020. That calculation cannot explain the whole UC gap or separate school effects from family selection. It also cannot resolve why re-matching changes the result.
Why we cannot claim a four-point limit for California community colleges
The California community-college estimate changes from −1.60 points in 2023 to +0.80 in 2020. Both primary nominal bootstrap intervals cross zero. That is uncertainty, not proof of equal outcomes.
For this California system, the script also computes an approximate equivalence envelope: the absolute estimated gap plus 1.645 standard errors. In the primary designs it is ±4.24 points for 2023 and ±3.13 for 2020. The 2023 alternative designs produce envelopes from ±4.28 to ±6.53, apart from the five-pair same-district case at ±11.63. In 2020 the alternatives range from ±3.62 to ±4.79, apart from the same-district case at ±7.56.
These are normal-approximation calculations within selected pairs. They do not guarantee that a larger difference would have been detected, and they do not bound a causal school effect. Our earlier claim that every check ruled out a shift larger than about four students per hundred was too strong.
What changed on this page?
Correction, September 28, 2026: We restored the exclusion of ZIP income reported as “$250,000 or more.” The 2023 pool changed from 1,732 schools and 127 pairs to 1,726 schools and 126 pairs; the 2020 pool changed from 1,697 and 128 to 1,691 and 127. We updated the primary results, alternative matches and sample descriptions for both years.
An outside review also found that our original random-number generator repeated resamples. We replaced it and recomputed every confidence interval, p-value and adjusted q-value for both years; the paired schools and average gaps did not change. These engineering checks do not certify the statistical assumptions. We also corrected the multiple-testing calculation, the mixed destination label, the distinction between scored subject records and distinct students, and the California-only community-college category and the strength of that conclusion. The page now reports only the four destination measures from this analysis. Graduation, AP and A–G results came from separate exploratory work and are no longer presented as additional outcomes from the same matched destination study.
The earlier published page reported an August 15 correction to its stored record-count input. The figures here come from the September 28 run. The analysis uses a fixed random seed; its school census, exact queries, pair lists and output logs are retained with the correction record. Contact us if you want to discuss the method.
Method: SchoolScope’s matched-school analysis of CDE records: 2023 destinations with 2022 grade-11 tests, 1,726 eligible schools and 126 pairs; 2020 destinations with 2019 tests, 1,691 eligible schools and 127 pairs. Neighborhood context is ACS 2020–24. Recomputed September 28, 2026. These comparisons do not identify individual students or causal school effects.
