Where the numbers come from.

Four numbers, four sources.

These are counts from the live database, not marketing figures. Where a number is rounded, the exact figure is next to it.

College data: three sources, named

Three datasets sit behind every admit rate, score range, cost figure, and aid number in the product. All three are public. You can check any of them without us.

The Common Data Set

The Common Data Set is a standardized questionnaire that colleges complete once a year. Same questions, same definitions, same order at every school. A college fills in its own applicant and admit counts, its SAT and ACT ranges at the 25th and 75th percentile, the GPA distribution of its entering class, tuition and room and board, the average aid package, the share of demonstrated need it meets, and how much weight its admissions office gives each factor it considers.

Two things make it the strongest source available to a family. It comes from the college itself, published on the college’s own site under its own name. And it is standardized, so a figure from one school means the same thing at the next.

A ranking is a different kind of object. Rankings take figures like these, apply weights somebody chose, and print one position. The weights are an opinion. The Common Data Set is the input those opinions are built on, and it is what we work from.

Coverage, including the gaps

IPEDS

IPEDS is the U.S. Department of Education’s reporting system. Any institution that takes part in federal student aid programs is required to report to it, so its coverage is close to complete in a way no voluntary survey can match. Every college in our database is keyed to its IPEDS Unit ID, which is how records get reconciled and how two campuses of the same university stay separate instead of merging into one. IPEDS supplies enrollment figures and the list of programs a school actually offers.

IPEDS lists more institutions than we carry. Our database holds 2,474 of them, so “every college in the country” is not a claim we make. If a school you are looking for is missing, that is why, and telling us is how it gets added.

College Scorecard

College Scorecard is the federal outcomes dataset, built in part from tax and financial-aid records. We use it for post-graduation earnings. It is the reason an earnings figure can sit next to a college at all.

How the SAT questions are made, and how they’re checked

23,000+ active practice questions. Here is exactly what they are.

We write the questions. The College Board does not. They are not retired official questions and we will not describe them as such. Part of the bank is written by AI against the official format. The rest is built from templates that assemble each item from a rule set, so the answer key falls out of how the question was constructed. Practicing on them should feel like practicing on a real section, which is the reason to build them against the real format instead of inventing one.

Checking them is where the work is. We hold 697 official College Board reference questions in the database. Statistics measured from that set become the benchmark. The AI-written questions are scored against those benchmarks item by item as they are generated, and anything that drifts gets flagged. The bank as a whole is also compared back to the reference set: average question length, how often a passage appears, and how often the key lands on A, B, C, or D.

What the automated check looks at

What the scores are, and what they are not

The Merit Score and the Gap Analysis are arithmetic on data you can see. Neither one is a prediction.

The Merit Score reads what you enter: courses and grades, test scores, activities, awards, work, service. It scores eight dimensions against a fixed rubric. There is no AI in the number. The same inputs produce the same score every time, and you can see which dimension moved and why.

The Gap Analysis compares your profile against the reported figures for the schools on your list: score ranges at the 25th and 75th percentile, GPA and class-rank distribution, and, where the college reports it, how much weight that admissions office gives each factor. The benchmark is the college’s own filing, not our opinion of the school.

Where AI is involved, and where it is not

How often this data changes

Institutional data moves on the institution’s schedule, not ours.

The Common Data Set is an annual filing. Each college publishes its own, usually months after the admissions cycle it describes. We import a new academic year’s records once they are published, and we store them by year, so a new filing sits alongside the old one instead of overwriting it. Nothing gets quietly rewritten under a figure you already read.

So the honest answer on freshness: this data updates once per annual cycle, per college, as each college publishes. It does not update weekly, and we are not going to claim it does. IPEDS and College Scorecard follow their own federal release schedules.

The practice-question bank changes when new questions are generated and pass validation. There is no fixed cadence, and we do not publish one.

Check any number

None of this rests on trusting us. Every source is open.