Propensity-score clipping and trimming

Create extreme inverse-probability weights, constrain them, and compare that operation with excluding people outside a chosen propensity-score range.

How are clipping and trimming different?

Clipping moves fitted propensity scores away from zero and one when weights are calculated, retaining every person but modifying the estimator. Trimming excludes observations outside a chosen propensity-score range. In this experiment, trimming changes the target to the retained subset.

Try it in the experiment

  1. Increase treatment selection to create poorer overlap.
  2. Raise the clipping threshold and compare the weights and point estimates.
  3. Open the trimming experiment and inspect who is excluded.

Neither operation repairs a structural positivity violation. Clipping reduces sensitivity to extreme weights but can introduce bias. A defined trimming rule can specify a population target. Report the rule and examine sensitivity rather than presenting either as automatic.