MEASURES AND COVERAGE — RELEASE 004

This is a research snapshot, not a declaration that every school's season is final.
Statistics cutoffs vary by school and may be unrecorded. Preserve those qualifications.
The historical window, age policies and classification rules are retained in
website_metadata; per-record source detail is available through source_records.

U.S. non-transfer minutes: minutes for players with a published U.S. hometown and a
working direct-entry classification, divided by the included working player-minute
total. Hometown is not a citizenship determination. International and transfer
populations overlap and their shares must not be added as exclusive categories.

Current direct entry: retain the established working assumption when no transfer
record is found; named possible prior colleges and documented conflicts remain
qualified. Historical unknown entry is not automatically classified as direct.

Minutes: unusable games are excluded consistently from both numerator and denominator.
Official games played and the reasons for exclusions remain in school/season coverage
records. The six national-review exclusions and their four affected schools are
retained in the reviewed schools, player records and corrections. Missing minutes
are not zero. Explicit reviewed zero-participation assumptions remain identifiable.

Age: published dates, reported ages and conditional estimates remain distinct.
Do not infer age from class year. The website's age reference and upper-estimate
policy remain unchanged; they are not automatically updated to the download date.

Academic outcomes: distinguish report year, entering cohort, sport scope and measure.
FGR and GSR are different measures. Ivy League FGR remains non-comparable/N/A under
the retained athletics-aid cohort policy; it is not replaced by GSR or an institution
or athletics-department rate. See academics.record_json and original source records.

Historical gaps remain documented. Consistency review does not mean every missing
source was recovered. The source databases and original archives are preserved.
Reviewed tables determine the public calculations; source_records describe the
original observations, which can intentionally differ from reviewed working values.

Stable school IDs join the data. A player record's compound key is source_volume,
source_table,source_row_id. Names are not global person identifiers. Do not merge
same-name players across schools without identity evidence. Conference in schools
is the current grouping; historical source conference fields retain their own scope.
