Methodology
How we evaluate every college essay
A score you cannot inspect is not feedback. This page publishes the six dimensions, the weights by essay type, the score calculation, the evidence requirement and the limits of the review — so you can decide whether the feedback is worth using.
Version 1.0 · Published November 4, 2026 · Independent rubric, internally developed. No external advisory panel has reviewed it yet.
What the score is — and is not
Scores come from our own writing rubric. They are not official admissions ratings and do not predict admission.
It is writing feedback
Six dimensions, each supported by quotes from your draft, plus a ranked revision plan.
It is not an admissions rating
No university has endorsed this rubric, and the score does not estimate your odds.
It is not an AI detector
We do not claim to determine whether AI wrote a draft. Detectors produce false positives.
It does not rewrite
The report points to the problem and asks a coaching question. It does not generate a replacement essay.
The six dimensions
Every review scores these six dimensions. The weights change by essay type, but the questions do not. Each score must be backed by verbatim evidence from the draft.
Personal Voice & Authenticity
Does the essay sound like a real person with a specific way of seeing things, or like a polished application template?
Strong looks like
- Sentence rhythms and word choices that fit the writer
- Observations that only this student would make
- Honest uncertainty instead of a tidy lesson
Common weakness
Formal, adult-sounding sentences that could belong to any strong applicant
Reflection & Insight
Does the essay show how the writer thinks, what changed in their understanding, and what they now notice that they did not before?
Strong looks like
- A before-and-after in thinking, not just in events
- A specific tension or question the writer is still working through
- Insight tied to a moment in the story
Common weakness
A final paragraph that announces a lesson ("I learned the value of hard work") without showing the change
Storytelling & Structure
Does the essay pull the reader forward, with scenes, pacing and an ending that lands?
Strong looks like
- A first paragraph that starts inside a moment
- Scenes with a purpose, not a chronological diary
- An ending that returns to the opening idea with new meaning
Common weakness
A timeline of events that ends with a summary instead of a scene or a thought
Specificity & Details
Are there concrete details — places, objects, dialogue, numbers, sensory moments — that make the story believable and memorable?
Strong looks like
- Details a reader could picture without explanation
- Small, surprising observations rather than general adjectives
- Evidence that supports the essay’s claims
Common weakness
Broad statements ("I have always been passionate about science") with no scene behind them
Character & Perspective
What does the essay reveal about how the writer treats other people, handles difficulty and sees the world?
Strong looks like
- Choices and actions that reveal values
- Awareness of other people, not only the writer
- A perspective that is still developing, honestly
Common weakness
Claims about being curious, kind or resilient that the story never demonstrates
Writing Quality
Is the prose clear, controlled and free of errors, clichés and sentences that try too hard?
Strong looks like
- Clear sentences with varied length
- Precise verbs and few unnecessary adjectives
- Grammar, punctuation and transitions under control
Common weakness
Thesaurus words, overlong sentences and repeated sentence openings
Weights by essay type
Each row sums to 100. A Why This College essay weights fit and contribution at 30 points; a personal statement weights voice and reflection at 20 each.
| Essay type | Personal voice | Reflection | Storytelling | Specificity | Character | Writing | Total |
|---|---|---|---|---|---|---|---|
| Common App essay | 20 | 20 | 15 | 15 | 15 | 15 | 100 |
| Supplemental essay | 15 | 15 | 10 | 20 | 25 | 15 | 100 |
| Why this college | 15 | 15 | 10 | 15 | 30 | 15 | 100 |
| Why this major | 15 | 20 | 10 | 20 | 20 | 15 | 100 |
| Community essay | 15 | 20 | 10 | 20 | 20 | 15 | 100 |
| Challenge essay | 15 | 20 | 15 | 15 | 20 | 15 | 100 |
| UC PIQ | 15 | 20 | 10 | 20 | 20 | 15 | 100 |
| Short answer | 15 | 15 | 10 | 25 | 20 | 15 | 100 |
How the total is calculated
The model returns six integer dimension scores and the evidence for each one. It does not return the total.
Our server clamps each score to its rubric maximum, adds the six dimensions and stores the total with the report. If a dimension is missing, the review is regenerated rather than scored with a guess.
Prompt alignment is scored separately out of 10. It is shown next to the essay score, not added to the 100 points, because answering the wrong question is a different failure from writing a weak essay.
Free and paid reviews use the same calculation. Payment unlocks depth, not a different number.
How to read a score
A distinctive, specific essay that reads like the writer and holds up against strong applicants.
A competitive essay with a clear story. A few targeted revisions would make it memorable.
The essay has a real story, but the reflection, detail or structure is still doing less than it could.
The essay answers the prompt, but key parts are general or underdeveloped. The revision plan matters more than the score.
The essay is not yet doing the work an admissions reader needs. Start with the highest-priority issues.
Calibration levels
The model returns points, not a 1–5 label. These levels describe the same ranges to a reader, so a score has a consistent meaning across dimensions.
Significant improvement needed
The dimension is largely absent or actively working against the essay. A structural rewrite is usually needed.
Developing
The intention is visible, but the execution is general, inconsistent or unsupported by evidence.
Competent
The essay does what the dimension asks at a basic level. The next step is specificity and control, not a new structure.
Strong
The dimension is a real strength. Remaining issues are local: a sentence, a transition, a missed reflection.
Exceptional
The dimension is distinctive and fully realised. Further polishing is unlikely to improve the essay much.
What the evaluation cannot do
It cannot predict admission. The score is writing feedback. It is not an admissions rating, and no university has endorsed it.
It cannot verify your facts. The review reads the draft; it does not check whether a quotation, name or achievement is accurate.
It cannot judge your topic. It evaluates how the topic is executed, not whether the topic itself is impressive to a particular school.
It is not perfectly stable. Model output varies. The rubric, weights and total are fixed, but a re-run can move a dimension by a point or two.
It does not rewrite your essay. It points to the problem and asks a coaching question. The revision has to be yours.
It uses a third-party AI provider. Reviews are generated through DeepSeek’s API. DeepSeek’s privacy policy states that it may use inputs to improve and train its models, retains data as necessary and processes data in China. We do not use your essay to train our own models, and we do not sell it. Do not submit information you would not want processed by a third-party AI provider. Read the privacy policy before submitting.
Review status
The rubric is internally developed and has not yet been reviewed by an external advisory panel. When that review is complete, we will publish the reviewers’ names and the sections they reviewed. We do not claim former admissions officers, university partnerships or student outcomes.
Sources we checked
We check the official requirements for prompts and limits, and we separate those rules from our own rubric. A link does not imply endorsement.
- Common App essay prompts — the seven prompts and the 650-word limit.
- UC Personal Insight Questions — the eight PIQ prompts and the 350-word limit.
- College Board — general application and writing resources.
Frequently asked questions
Who created this rubric?
The rubric was developed internally for Free College Essay Grader. It draws on public admissions guidance and established writing-evaluation practice, not on a proprietary admissions-office formula. We do not have former admissions officers on staff, and we do not claim an external advisory board. The methodology is published here so you can judge it. When external reviewers complete a formal review, we will publish their names and the sections they reviewed.
Is this the same rubric colleges use?
No. Universities do not publish a single essay-scoring formula, and admissions decisions are not made by adding up writing criteria. This is our own feedback rubric. It is designed to surface the qualities readers consistently describe — specificity, reflection, voice, structure, character and writing quality — not to reproduce an admissions office’s internal process.
Why do the weights change by essay type?
A 650-word personal statement and a 250-word Why This College essay are doing different jobs. In a fit essay, school-specific evidence and contribution carry more weight and narrative structure carries less. Changing the weights is how the score reflects the actual assignment instead of pretending every essay is the same genre.
Does paying for a full review change the score?
No. Free and paid reviews use the same six dimensions, the same weights and the same server-side total. Payment unlocks depth — all line feedback, the full revision plan and the voice analysis — not a different score. If a free review and a paid review of the same draft ever disagreed, that would be a bug, not a pricing feature.
Can the same essay get a different score twice?
Model output has some natural variance. The rubric, the weights and the total calculation are fixed, and every quote is verified against the essay, but a re-run can move a dimension by a point or two. Use the scores to compare revisions of the same essay on the same rubric, not as an absolute ranking across applicants or across tools.
Does the score predict admission?
No. This is writing feedback, not an admissions prediction. Admissions decisions depend on your transcript, context, activities, essays, recommendations and the school’s priorities. A high score does not guarantee admission, and a low score does not mean rejection.
Which AI provider processes my essay?
Reviews are generated through DeepSeek’s OpenAI-compatible API. DeepSeek’s privacy policy states that it may use inputs to improve and train its models, retains data as necessary and processes data in China. We do not use your essay to train our own models, and we do not sell it. Do not submit an essay containing information you would not want processed by a third-party AI provider. The privacy policy explains this in full.
Free review
Ready to see what the reader notices?
Paste one essay. Get the score, the six dimensions and the sentences that need work — before you hit submit.