Methodology
Where Every Figure on This Site Comes From
Page updated 28 September 2026
This site is published by Digital Signet Limited. Every quantitative claim on it is either read from a named primary source with its vintage stated, or it is arithmetic you can redo from inputs shown on the page. Where neither is possible, we say that no one publishes the figure rather than estimating one.
The rule
A number on this site has to pass one of two tests. Either it comes from a document we have opened, in which case the page names that document, its edition or cohort, and links to it. Or it is a model: our arithmetic over inputs printed on the same page, so a reader can check the inputs and redo the sum. A figure that is neither is not published.
The most common failure mode in this subject area is the third option: a plausible number with a respectable label attached. "Compiled from NCES and academic research" is the kind of phrase that appears exactly where a citation should be. We have removed those from this site where we found them, and the per-major table on average GPA by major was rebuilt in September 2026 for that reason.
What sits behind each layer
Average GPA by major
UC Berkeley Office of Planning and Analysis, Average GPA by Major
86 majors, 11,399 bachelor's degree recipients, with headcounts and standard deviations. One institution, labelled as such on every page that uses it.
2025-26 graduating class
Last national mean GPA by field
NCES 2003-165, A Descriptive Summary of 1999-2000 Bachelor's Degree Recipients 1 Year Later, Table 5
Mean cumulative GPA and mean GPA within the major, for twelve fields. The only national per-field means NCES has published, and a quarter of a century old. Reproduced with its vintage stated because the figures in circulation (engineering 2.90, education 3.36) match no national source, while these are 3.11 and 3.22.
1999-2000 bachelor's degree recipients, surveyed 2001
National GPA distribution by field
NCES Baccalaureate and Beyond Longitudinal Study (B&B:08/09), Table 2.3
A percentage distribution across four GPA bands, not a mean. This is the most recent national per-major GPA data NCES has published.
2007-08 graduates, surveyed 2009
STEM course grading gap
NCES 2013-152, Table 8 (BPS:04/09 and PETS:09)
2.8 in STEM courses against 3.0 in non-STEM courses. The table is titled "FIRST-YEAR PERFORMANCE IN STEM AND NON-STEM COURSES", so this is a first-year course GPA: not cumulative, and not a GPA by major. Every page using it carries the first-year qualifier.
2003-04 beginning students, first year
National average undergraduate GPA
Rojstaczer and Healy, gradeinflation.com
The widely quoted 3.15, stated on the site as "By 2013, the average college student had about a 3.15 GPA". The national chart behind it rests on 87 four-year institutions, chosen because they had published grade data or sent it to the author, so the sample is self-selected rather than representative; the 400-school figure often attached to it covers the whole database, not this chart. Treat it as approximate and dated, not as this year's number. It is NOT an NCES figure, and NPSAS publishes no national mean GPA table.
2013 figure, site last substantively updated March 2016
Long-run grade inflation
Rojstaczer and Healy, Teachers College Record
Used for the direction and rate of change (0.14 GPA points per decade over fifty years), and for the finding that engineering departments grade above science departments. Pre-1983 figures are prose statements rather than series points: the earliest dated national GPA figure either source supports is 2.9 in 1973, so this site publishes no 1950s number.
2010 and 2012 papers
Undergraduate admissions
Each university's published Common Data Set
Item C11 is the high school GPA distribution and C12 the average. Both describe enrolled students, not admitted applicants, and several institutions leave them blank. Where a school publishes nothing, we say so rather than estimating.
Most recent cycle published by each institution
Law school
ABA Section of Legal Education, Standard 509 disclosures
Medians and 25th to 75th percentile bands as filed by each school.
Fall 2025 entering class
Medical school
AAMC Facts, applicants and matriculants
Overall and science (BCPM) GPA for matriculants, taken from the published tables rather than secondary summaries.
2025-2026 cycle
Employer screening
The share of employers screening by GPA. Industry-level cutoffs quoted from employers' own published graduate criteria where they exist.
Most recent published edition
Federal employment
US Office of Personnel Management
Superior Academic Achievement thresholds from the qualification standards. Pay figures from the published General Schedule tables.
Current regulation
International conversion
WES and ECE published methodologies
Conversions are presented as each evaluator's own output, because they differ from one another and from individual programmes' internal recalculations.
Current published guidance
What we do not publish
A current national average GPA for each major. No US federal agency publishes one. NCES last published mean GPAs by field for the 1999-2000 bachelor's degree recipients, in NCES 2003-165, Table 5 (B&B:2000/01); we reproduce that table with its vintage stated, and it is the only national per-field mean on the record. Since then NCES has published a distribution across GPA bands rather than an average. The NPSAS study does carry an estimated GPA variable and a field-of-study variable, but any cross-tabulation of the two is a user-generated query with no table number and no published figure behind it, built on a variable NCES itself labels "estimated". So we give a named institution's current published averages, the last published national means, and the national distribution, each labelled with what it is, and decline to invent a current national per-major figure.
Admitted-student GPA at institutions that do not report it. The Common Data Set asks about enrolled students. Where a school leaves items C11 and C12 blank, we say it publishes nothing rather than supplying a figure from a secondary summary.
Precise odds of admission at a given GPA. Admissions are decided on whole files. A conversion from one number to a probability would be a guess wearing a decimal point.
Known limits of the figures we do publish
- The national average of about 3.15 is a 2013 figure from a self-selected sample of 87 four-year institutions, not a current, nationally representative measurement, and it is not an NCES figure. The nearest federal mean is 3.16 for 1999-2000 bachelor's degree recipients (NCES 2003-165, Table 5), which covers graduates only and so runs above a whole-population mean. Institutional data published since 2013 runs higher than both. Read 3.15 as a long-standing reference point rather than this year's number. Nobody publishes a 2026 national average.
- UC Berkeley is a highly selective public university. Its major averages are useful for the ordering between fields and for the size of the gaps, not as a level any student elsewhere should expect to match.
- Several Berkeley majors graduate fewer than twenty students a year. Those rows carry their headcount so a reader can see how thin the base is.
- The national field-level distribution describes graduates from 2007-08. It is the most recent published, and it is old. Where it disagrees with recent institutional data we show both.
Dates
Pages carry two kinds of date and they mean different things. A page-updated date records when the content was last reviewed. A data vintage records the cohort, term or edition the figures describe. A page reviewed this month can legitimately carry data from an older cohort, and where that is the case both dates appear.
Independence
The site is not affiliated with any university, testing organisation, scholarship programme, accrediting body or admissions consultancy. It is free to read and supported by display advertising (Journey by Mediavine). The advertising is served by third parties, is not sold or placed by us against particular content, and does not influence which sources are cited or what the data says. See the privacy policy for how advertising cookies are handled.
Corrections
If a figure here is wrong, or a source has moved or been superseded, tell us and it gets fixed. Write to [email protected]. Corrections are made in place: the figure is changed and the page date moves. We do not leave a false statement standing anywhere on the page in order to annotate it.
Published by Digital Signet Limited. Written by Oliver Wakefield-Smith. Educational reference, not admissions advice. See about for the author bio and the wider portfolio.