Audit how disparity metrics move over time and across cities — morie_predpol_temporal

For every (city, period) cell the four disparity metrics are computed; per city the audit then reports the mean of each metric, the count of periods with DIR above 1, and the DIR temporal range (max minus min) — the headline measure of instability.

Usage

morie_predpol_temporal_audit(
  period,
  city,
  y_pred,
  group,
  privileged = NULL,
  favorable = 1
)

Arguments

period: Time-period label for each record (e.g. "2019-03").
city: City label for each record.
y_pred: The decision/assignment for each record.
group: Protected attribute for each record.
privileged: Reference group; inferred globally from the pooled data when NULL so every cell uses the same reference.
favorable: Value of y_pred counted as favourable (default 1).

Value

A named list: value (worst per-city DIR range), worst_dir_range, cross_city_dir_spread, per_city, cells, privileged, warnings, interpretation.

Examples

period <- c(rep("p1", 10), rep("p2", 10))
city <- rep("A", 20)
pred <- rep(c(1, 1, 1, 1, 1, 1, 1, 1, 0, 0), 2)
grp <- rep(c(rep("X", 5), rep("Y", 5)), 2)
res <- morie_predpol_temporal_audit(period, city, pred, grp, privileged = "X")
res$per_city$A$dir_range # 0 — disparity is stable across periods
#> [1] 0