Structured evidence review
What the research shows about AVID
Read this knowing who wrote it.
We run a California Partnership Academy. AVID is a programme many districts fund instead of, or alongside, one. We have an interest in the answer, and you should discount accordingly. What we can offer instead of neutrality is the same method we applied to our own programme — where it produced findings against us, and we published those too.
AVID was founded in San Diego in 1980. The question here is narrow and answerable: after forty-five years, what can be shown to be caused by AVID, and how long after school was it measured?
How far each programme has been followed
Every bar is the strongest study of that programme, and its length is how long after high school the outcome was measured. Gold marks a design that can support a causal claim and a follow-up long enough to reach working life.
end of high schooleight years later
That is the finding. Not that AVID fails — that nobody has looked far enough to say. There is no causal evidence on college persistence, on degree completion, or on earnings.
What AVID does have, stated fairly
WWC Intervention Report: AVID
IES / NCEE, September 2010 (ERIC ED511791) · federal evidence review
- null — “AVID was found to have no discernible effects on comprehension for adolescent learners.” Extent of evidence: small.
- descriptive — Of 66 studies reviewed, ONE met WWC evidence standards with reservations and the remaining 65 met neither the standards nor the eligibility screens. None met standards without reservations.
- descriptive — The single qualifying study covered 96 high-school-age youth at four schools in ONE Colorado school district, and was quasi-experimental rather than randomised.
SCOPE, and it is load-bearing: this review ran under the Adolescent Literacy protocol, so many of the 65 were screened out for measuring outcomes outside that protocol — college enrollment, GPA — rather than for weak design. The 66/1/65 tally is a statement about the research base, NOT a verdict on AVID’s college outcomes, and must never be used as one. The single qualifying study (Rorie 2007, a doctoral dissertation) reported too little for the WWC to compute an effect size: mean difference, effect size and improvement index are all recorded as not reported. The report also carries the federal government’s own description of who AVID recruits: it “focuses on under-served, middle-achieving students (defined as students earning B, C, and even D grades)” and supports them “through a daily elective period and ongoing tutorials”. That is selection into the programme described by the reviewer rather than by us.
How far we read it · primary document read, both the IES and the fuller 16-page ERIC version; quoted verbatim
Outcome of the AVID College Preparatory Program on Adolescent Health: A Randomized Trial
Pediatrics 151(1):e2022057183, PMC9830585 · randomized controlled trial · one school year, from the transition into high school to the end of ninth grade
- positive — lower odds of using any substance (OR 0.66, 95% CI 0.48-0.89)
- positive — lower odds of associating with a substance-using peer (OR 0.74, 95% CI 0.45-0.98); higher odds of associating with a peer engaged in school (OR 1.73, 95% CI 1.11-2.70)
- positive — male AVID students had lower stress and higher self-efficacy, grit and school engagement
This is real evidence and it is positive — the first randomised trial of AVID in the United States, forty-three years after the programme was founded. What it does not do, the authors state themselves: “We did not examine whether AVID achieves its primary objective of increasing college enrollment and persistence.” The trial was designed to run longer: “Our initial study aimed to follow students through 11th grade, however the coronavirus disease 2019 pandemic-related school closures disrupted typical AVID delivery.”
How far we read it · abstract and open-access full text read; limitations quoted verbatim
AVID Participation in High School and Post-Secondary Success: An Evaluation and Cost Analysis
Journal of Research on Educational Effectiveness 13(4):679-701, peer reviewed, DOI 10.1080/19345747.2020.1772421 · coarsened exact matching · high school graduation and college enrolment; no outcome beyond enrolment
- positive — “students who complete at least one AVID elective have higher high school graduation and college enrollment rates than comparable non-AVID students”
- descriptive — not measured
- descriptive — not measured
- descriptive — not measured
This is AVID’s strongest college-outcome study and it is peer reviewed, so treat it seriously. Two limits are visible in the abstract alone. Coarsened exact matching balances OBSERVED baseline characteristics — it cannot balance motivation, and motivation is part of what AVID selects on, since student recruitment and selection is the first of its own certification essentials. And the treatment is defined as completing at least one AVID elective, which conditions on something that happens AFTER baseline: students who finish the elective are the ones who did not leave it. Matching on entry characteristics does not fix selection on staying. The outcome ladder stops at enrolment. Separately, and it is worth saying: the paper is paywalled, so a parent cannot check it.
How far we read it · ABSTRACT AND BIBLIOGRAPHIC RECORD READ, NOT THE FULL PAPER — the article is closed access. Effect sizes and the matching implementation have not been inspected.
AVID II course profile (district-published course description)
tp.sduhsd.net · programme documentation
- descriptive — “All levels of AVID earn the ‘g’ credit on the a-g UC/CSU requirements for college entrance.” The elective itself counts toward the college-readiness measure used to judge it.
- descriptive — “Prerequisite: AVID I or Application and teacher/counselor recommendation, minimum 2.5 GPA requirement. Co-requisite: Needs to be enrolled in college prep courses, must meet UC/CSU a - g requirements.”
Two mechanisms, both in a district’s own public course description. Entry requires a minimum grade point average, a teacher or counsellor recommendation, and concurrent enrolment in college-preparatory a-g courses — so students are filtered on prior achievement and on already being on the track before the programme starts. And the elective earns a-g elective credit, so enrolling in AVID advances the very measure used to show AVID works. This is one district’s published description, not proof that every site operates this way; AVID’s own certification essentials make student selection Essential 1.
How far we read it · primary document read; quoted verbatim
AVID’s own published outcome statistics (avid.org/data)
avid.org · self-collected counts, no comparison group
- descriptive — “9 out of 10 AVID seniors in 2025 are college- and career-ready”
- descriptive — “93% completed 4 years of rigorous coursework necessary to be eligible for college enrollment”; “83% reported taking at least one course of rigor”; “88% submitted the FAFSA”
- descriptive — “86% plan to attend either a 2- or 4-year college” — a statement of INTENT, not of enrolment
- descriptive — no figure published
- descriptive — no figure published
Every figure carries one footnote: “AVID. (2025). AVID senior data collection [Electronic Database, N=49,000].” The page states no response rate and no total senior population, so the denominator cannot be checked. It names NO comparison group of any kind — not a national average, not a state average, not non-AVID students — so nothing here is an effect; they are levels. The headline college figure measures what students PLAN to do. And for a programme whose stated purpose is college success, the page publishes no persistence figure and no degree-completion figure at all.
How far we read it · page fetched and transcribed in full
AVID Center IRS Form 990 filings (EIN 33-0522594)
IRS, via ProPublica Nonprofit Explorer · financial disclosure
- descriptive — AVID Center is a 501(c)(3) NON-PROFIT. In 2023 it took $86,941,213 in revenue against $81,800,649 in expenses — a surplus of $5,140,564.
- descriptive — Of that revenue, $76,941,192 (88.5%) is programme service revenue — fees paid by districts for licensing, curriculum and training — against $1,810,896 (2.1%) in contributions. It is funded by the schools it serves, not by donors.
- descriptive — Accumulated net assets grew from $52,383,443 in 2011 to $149,027,888 in 2023 — a rise of $96,644,445, close to a tripling, across thirteen filing years.
Said precisely, because the imprecise version is the easy one to dismiss. AVID Center is a non-profit: there are no shareholders and no profit, and anyone who checks the EIN will confirm it in seconds. What the filings show is a surplus and a large accumulation of net assets, funded almost entirely by fees from public schools. None of that is wrongdoing and it is not offered as any. With 88.5% of its money coming from what districts pay it, AVID Center behaves like a business selling a service to school districts, whatever its tax status — and that is the fair way to read the filings. It is context for one question: this accumulation happened across the same thirteen years in which the programme produced a single one-year randomised trial that did not measure college outcomes. A research agent reported FY2024 figures that do not reconcile with the ProPublica series and are not used. The 2021 filing is unusual — expenses far below revenue, contributions at $8.99M — most likely pandemic relief, and marked unverified because it has not been checked.
How far we read it · ProPublica Nonprofit Explorer API read directly; 13 filing years
The trial is real, and so is its limit
The 2023 trial is the strongest thing in AVID’s record and it found genuine benefits — less substance use, healthier friendships, and for young men lower stress and more self-belief. Those matter, and a fair reading has to say so. It was run by UCLA in five Los Angeles Unified high schools, and it is the first randomised trial of AVID in the country, forty-three years after the programme began.
“We did not examine whether AVID achieves its primary objective of increasing college enrollment and persistence.”— the trial’s own authors
It was designed to run longer. The authors write that they “aimed to follow students through 11th grade” and that pandemic school closures ended it at the end of ninth. That is misfortune, not evasion — but the gap it leaves is still a gap.
The circle the evidence is drawn inside
The measure most often used to show AVID works is a-g completion — whether a student finished the courses a California public university requires. Two sentences from one district’s own published course description explain why that measure cannot settle anything.
“Prerequisite: AVID I or Application and teacher/counselor recommendation, minimum 2.5 GPA requirement. Co-requisite: Needs to be enrolled in college prep courses, must meet UC/CSU a - g requirements.”
“All levels of AVID earn the ‘g’ credit on the a-g UC/CSU requirements for college entrance.”
Read them together. To enrol, a student must already hold a qualifying grade average, already have a teacher’s recommendation, and already be taking the college-preparatory courses that a-g measures. Then the elective itself earns a-g credit. So the programme admits students who are on the track, and then advances them along it by the fact of enrolment. Reporting afterwards that AVID students complete more a-g requirements describes the arithmetic, not an effect.
This is one district’s published description and not proof that every site works this way. It is quoted because it is public, checkable, and written by a school district rather than by us. AVID’s own certification framework makes student selection the first of its essentials.
The number we will not use
A federal review of AVID examined 66 studies and found one that met its evidence standards, with reservations. That is a real statement about the research base and we have kept it. It is also the most misusable figure here, so we are marking it ourselves: that review ran under a reading-and-literacy protocol. Many of the 65 were set aside for measuring the wrong thing for that review — college enrolment, grades — not for being badly designed. Anyone quoting 66-to-1 as proof AVID does not work is overreaching, and a reader who opens the report will see it within a minute.
Not yet read
Named so the gaps are visible. We are not publishing numbers from these until we have opened them, including the ones that would suit us.
- UChicago Consortium analysis of AVID in 14 Chicago high schools — CHASED AND NOT FOUND, which is the finding. The result quoted everywhere — that ninth-grade AVID participants 'gained little advantage' and 'remained off track for graduation and college' — comes from a 2011 conference presentation written up in the education press. The Consortium’s own news item says the full report was 'set for release in the fall' of 2011, and no such report is locatable in its publications. What the Consortium HAS published on AVID is different work: ‘CCSR’s 2007 AVID Reports for Chicago Public Schools’ (Nagaoka, Deutsch, Brake and Roderick, January 2008), which is survey-based and reports what students and teachers said about their experience rather than what happened to their grades or graduation. So this is not a study we can cite. It is a press account of an unpublished analysis, and it is recorded here so that nobody — including us — quotes it as though it were peer-reviewed. If the report exists we would like to read it.
What would change our mind
Does AVID have anything of comparable design strength AND follow-up length to Kemple 2008 (randomised, eight-year earnings), Brunner/Dougherty/Ross 2023 (discontinuity, earnings still elevated past age 23) or Rosen 2023 (lottery, associate degrees at seven years)? On the located and read record: no. Its one randomised trial ran a single year on health outcomes and states it did not test college enrolment or persistence. Its strongest college study is a matched design reaching enrolment. There is no causal evidence on persistence, on degree completion, or on earnings. That is falsifiable: one well-designed study following AVID students that far would overturn it. One well-designed study following AVID students to degree completion or earnings would overturn the conclusion above, and we would publish it here. If you know of one, tell us and we will read it.
Method as on our own review: each study’s design and comparison group stated before any effect; no pooling of estimates that measure different things; and claims with no stated method excluded rather than counted as weak evidence. Sources · the same method applied to us.