# How this site reads merit aid and ranks the ways to apply

Every college row and every outside award on this site was read on the college's or provider's own page, and carries the date it was read. The one ranking, of ways to apply to many scholarships, uses weights committed on November 18, 2026 before any score was opened.

## How pages are read and dated

Each college and award page is opened and read in full, not summarized from a search result. We copy the cutoffs, amounts and rules as the page states them, with a short quote and the address of the page. Every row carries the date it was read. If a page cannot be read by an automated browser, such as studentaid.gov, we do not state a figure from it and instead tell the student where to find it. Counts on the site, like the number of colleges, are computed from the rows. Nothing is rounded up or estimated, so a college with no stated figure is counted as having none stated.

## How the weekly refresh works

Once a week, the source pages behind the Automatic Merit Table and the Outside Merit Ledger are read again. If a cutoff, amount or deadline changed, the row is updated and shows the new read date. If a page has moved or been removed, the row is flagged until a current page is found. The news items on the [news page](https://whatisameritscholarship.com/news/) are refreshed on the same schedule and show their own read dates, with a quote from the source for each.

## Corrections

Errors are fixed on the row and logged with the date, the page and the old and new values. To report one, email editor@whatisameritscholarship.com with the page address on this site and the address of the source page. The [corrections page](https://whatisameritscholarship.com/corrections/) has the log, which starts on 2026-11-18. When a reader reports a change before the next weekly read, the row is checked right away. The rules for sources and wording are on the [editorial policy page](https://whatisameritscholarship.com/editorial-policy/). Rows that fail a weekly read are marked until the page is read again.

## The weights for the main view

- One profile, many applications: 25
- Matching to eligibility: 20
- Vetted listings: 15
- Tracking: 10
- Essay reuse: 10
- Cost to the student: 10
- Trust signals: 10

Committed in 8382afb on November 18, 2026. Source: Main view: RUN-base.md section 3 default table, unchanged (RUN.md section 5.1). Situation weights: written by a fresh-context Opus subagent given only the seven criteria, RUN-base section 3 and the five reader situations, before any score file was opened.

## The weights for each situation

- [situation 1: a student with a 3.5 GPA and no test score](https://whatisameritscholarship.com/situations/a-35-gpa-no-test/): One profile, many applications 25, Matching to eligibility 30, Vetted listings 10, Tracking 10, Essay reuse 10, Cost to the student 10, Trust signals 5. Many merit awards set a test score cutoff, so the most useful platform hides the ones this student cannot get and lets one profile reach every award that runs on GPA alone.
- [situation 2: a student with a high test score and an average GPA](https://whatisameritscholarship.com/situations/high-test-average-gpa/): One profile, many applications 20, Matching to eligibility 30, Vetted listings 10, Tracking 10, Essay reuse 15, Cost to the student 10, Trust signals 5. This student needs a platform that matches on test score as well as GPA, and essay reuse counts for more because an essay is where they can show strengths beyond their grades.
- [situation 3: an out-of-state applicant](https://whatisameritscholarship.com/situations/out-of-state/): One profile, many applications 20, Matching to eligibility 30, Vetted listings 15, Tracking 10, Essay reuse 10, Cost to the student 10, Trust signals 5. Many merit awards are limited to residents, so matching on residency matters most, and vetting matters more than usual because listings that leave out residency rules waste an applicant's time.
- [situation 4: a family that earns too much for need-based aid](https://whatisameritscholarship.com/situations/too-much-for-need/): One profile, many applications 25, Matching to eligibility 25, Vetted listings 15, Tracking 10, Essay reuse 10, Cost to the student 5, Trust signals 10. This family wants only awards that ignore income, so matching and vetting that screen out need-based listings matter most, and a small fee matters less to them than knowing the platform can be trusted.
- [situation 5: a transfer applicant](https://whatisameritscholarship.com/situations/transfer/): One profile, many applications 20, Matching to eligibility 30, Vetted listings 10, Tracking 15, Essay reuse 10, Cost to the student 10, Trust signals 5. Many merit awards are only for first-year students, so matching on transfer eligibility matters most, and tracking counts for more because transfer deadlines rarely line up with the usual calendar.

## How cost is scored

Cost is scored for the plan a platform's other scores depend on: if a feature that earns points elsewhere in the table needs a paid membership, the platform is scored on that membership's price, and if that price is not published before sign-up it gets the score for an unpublished price.

## When the weights were fixed

The main weights are the published default table, unchanged. The five situation weights were written on November 18, 2026 by a reviewer given only the seven criteria, the default table and the five situations, and committed in 8382afb before this site opened any score. The network's canonical score file already existed at that time on its own record; it was opened here only after the weights commit, and nothing in it was changed.

## Where the scores come from

Scores are the network's canonical score set dated November 18, 2026 (file commit be966a1), copied unchanged for the platforms in this site's view. Merged evidence from six branches (data/evidence/), scored once by a blind fresh-context Opus scorer (blind/scores-pass1.json), checked by a blind fresh-context Opus reviewer (pass 2) and, after a curation audit restored lost facts, by a second blind reviewer (pass 3); every change in blind/review-log.jsonl. Neither saw which platform is the client, any weight, any view or the earlier disagreements. If a page read for this site contradicted a fact a score rests on, the fact is logged for review and the score stays.

## Sources

- Score set and rubric: https://whatisameritscholarship.com/data/views.json (read November 18, 2026)

Last reviewed November 18, 2026

Canonical: https://whatisameritscholarship.com/method/
