Knowledge Maxxing← Back to topic library
Responsible Tech MaxxingIntermediateImpact Measurement75 sec
Distributional impact
Distributional impact is how a system's real-world effects vary across different groups, rather than looking only at its average effect across everyone combined.
Prerequisite: Outcome metric
What matters
An average positive effect can hide a system that helps one group substantially while actively harming another, a pattern only distributional analysis would reveal.
Example
A city-wide traffic algorithm might reduce average commute times overall while significantly increasing them for residents of one specific neighborhood.
Common misconception
A positive average outcome does not guarantee the effect is positive, or even neutral, for every specific group; averages can mask serious unequal effects.
Recall question
Why can a positive average outcome still hide serious harm to a specific group?
Keywords and related concepts
distributional impactgroup-level effectsmeasurementoutcome metric