Skip to contents

Reduces the long score table to one row per ANON record, recording where the *true* RAW record sits in that record's ranking and how many other RAW records are indistinguishable from it.

Usage

reid_per_anon(
  scores,
  confidence = c("margin", "tie"),
  tolerance = reid_tie_tolerance()
)

Arguments

scores

a score table (see [score_num()])

confidence

which confidence measure to put in the CONFIDENCE column, `"margin"` (default since Issue #44) or `"tie"`. See [reid_confidence()].

tolerance

relative tie tolerance (Issue #61), see [reid_confidence()]. `N_BETTER`, `TRUE_TIE_SIZE`, `TRUE_RANK` and `BEST_TIE_SIZE` are all counts of "how many candidates are as good as, or better than, this one", so all four depend on it.

Value

a data frame with one row per ANON record and columns ANON_ROW_NUMBER, N_CANDIDATES, BEST_SCORE, BEST_TIE_SIZE, CONFIDENCE, MARGIN, ECCENTRICITY, TRUE_SCORE, N_BETTER, TRUE_TIE_SIZE and TRUE_RANK.