Issues

• PreTreatmentPlaceboCheck passes cancelling pre-trends of +100 and -100

  bug
  pymc-labs/CausalPy#1320 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• McCraryDensityTest ignores n_bins and compares all observations

  bug
  pymc-labs/CausalPy#1319 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• generate_multicell_geolift_data does not return the documented Denmark series

  bug
  pymc-labs/CausalPy#1318 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• WeightedProportion broadcasts the loss, so mse and column-outcome weights are wrong

  bug
  pymc-labs/CausalPy#1317 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• generate_staggered_did_data writes a later event-time effect onto earlier periods

  bug
  pymc-labs/CausalPy#1316 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• Staggered DiD effect_summary ignores alpha and always reports a 94% HDI

  bug
  pymc-labs/CausalPy#1315 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• OLS persistence summary reports probability 1 that a reversed effect persisted

  bug
  pymc-labs/CausalPy#1314 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• OLS effect_summary labels a 97.5% interval as 97%

  bug
  pymc-labs/CausalPy#1313 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• InversePropensityWeighting.get_ate pairs propensities with the wrong outcomes

  bug
  pymc-labs/CausalPy#1312 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• Panel dummy fixed effects report a nonzero coefficient for a time-invariant treatment

  bug
  pymc-labs/CausalPy#1311 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• Panel demeaning does not within-transform interactions on a balanced panel

  bug
  pymc-labs/CausalPy#1310 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• Staggered DiD reports an identified ATT for an always-treated unit

  bug
  pymc-labs/CausalPy#1309 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• Multi-unit Synthetic DiD stores the pooled mean under the first treated unit

  bug
  pymc-labs/CausalPy#1308 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• Synthetic DiD splits pre/post by row position, so an unsorted index flips tau

  bug
  pymc-labs/CausalPy#1307 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• PiecewiseITS post window follows formula order, not the earliest interruption

  bug
  pymc-labs/CausalPy#1306 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• InterruptedTimeSeries post-period summary truncates a float treatment_end_time

  bug
  pymc-labs/CausalPy#1305 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• OLS effect_summary cumulative p-value is 1 when the impact has no residual variance

  bug
  pymc-labs/CausalPy#1304 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• OLS effect_summary cumulative relative effect sums period percentages

  bug
  pymc-labs/CausalPy#1303 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• RegressionDiscontinuity accepts float 0/1 treatment and crashes in fit

  bug
  pymc-labs/CausalPy#1302 · Filter by author drbenvincent opened on Oct 24, 2026
  2

  

• step() and ramp() crash on timezone-aware datetimes with a string threshold

  bug
  pymc-labs/CausalPy#1301 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• Quoted bare datetime predictors are encoded as date fixed effects, not elapsed days

  bug
  pymc-labs/CausalPy#1300 · Filter by author drbenvincent opened on Oct 24, 2026
  2

  

• PrePostNEGD reports the intercept when the group column is named t

  bug
  pymc-labs/CausalPy#1299 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• RegressionKink bandwidth filters column x, not the running variable

  bug
  pymc-labs/CausalPy#1298 · Filter by author drbenvincent opened on Oct 24, 2026
  1

  

• PlaceboInTime and pymc-forecast follow-ups after #1261

  pymc-labs/CausalPy#1284 · Filter by author anevolbap opened on Oct 19, 2026

• [meta] Bayesian synthetic control for imperfect pre-intervention fit

  enhancementgeo project
  pymc-labs/CausalPy#1238 · Filter by author drbenvincent opened on Oct 7, 2026
  5