Skip to content

Consistency–equality fairness metric for educational recommenders

Research Method · All research methods
Evidence · no claims cited

Description

The paper formalizes consistency as the similarity between the targeted degree p_u(m) each platform sets for each principle m and the degree q(m) achieved in a learner's recommended list, measured as the complement of a weighted Manhattan distance. It then defines equality as "the complement of the Gini index over the consistencies across learners", where a value of 1 means the recommender is perfectly equal across learners in terms of consistency. Unlike prior individual-fairness work, learners count as positively treated when they achieve high consistency regardless of their past interactions, and no sensitive demographic attributes are used.

Accounts

  • Fairness metric combining consistency (complement of weighted Manhattan distance) and equality (complement of Gini index) across learners: The paper formalizes consistency as the similarity between the targeted degree p_u(m) each platform sets for each principle m and the degree q(m) achieved in a learner's recommended list, measured as the complement of a weighted Manhattan distance. It then defines equality as "the complement of the Gini index over the consistencies across learners", where a value of 1 means the recommender is perfectly equal across learners in terms of consistency. Unlike prior individual-fairness work, learners count as positively treated when they achieve high consistency regardless of their past interactions, and no sensitive demographic attributes are used. (Mirko Marras et al. (2022))

Claims

-

Key Sources

  • Mirko Marras, Ludovico Boratto, Guilherme Ramos, Gianni Fenu. (2022). Equality of Learning Opportunity via Individual Fairness in Personalized Recommendations. International Journal of Artificial Intelligence in Education. https://doi.org/10.1007/s40593-021-00271-1
Page metadata
Field Value
type research-method
id consistencyequality-fairness-metric-for-educational-recommenders
title Consistency–equality fairness metric for educational recommenders
description The paper formalizes consistency as the similarity between the targeted degree p_u(m) each platform sets for each principle m and the degree q(m) achieved in a learner's recommended list, measured as the complement of a weighted Manhattan distance.
status review
generated by process:settle-candidates on 2026-10-28
trust unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.