OULAD snapshot evaluation protocol with leakage-excluded presentation-level split

Design · All designs
Evidence · 2 claims (2 for) · 1 study (1 design), q2 · 0 of 1 report an effect size · 2 claims rest on one study

Description

The evaluation protocol constructs two snapshots per eligible student per assessment, at t = d-14 and t = d-7 before the due day, using only information available up to snapshot time; post-t interactions and assessments are excluded. Features include "aggregated click counts by activity type over rolling 7-, 14-, and 28-day windows", availability-alignment features, prior assessment history, time-to-deadline variables, and immutable learner descriptors. Data are split at the module-presentation level (15 train, 2 validation, 5 test presentations) to avoid leakage across closely related offerings, with pass defined as score >= 40.

Design Implications

Context

Requirements

• Snapshots must exclude post-t information to avoid leakage
• Splits at module-presentation level to avoid leakage across related offerings

Constraints

• Early assessments that would induce negative snapshot times are handled during preprocessing

Target Learners

• Open University students in the OULAD dataset

Learning Goals

• next-assessment pass/fail prediction and recourse generation

Claims

• Oulad Calibrated Logistic Regression Strong Prediction [+M]
• Ip Recourse Compact Plans At Scale [+M]

Related Designs

Examples

-

Key Sources

• Ngoc Luyen Le, Marie-Hélène Abel, Bertrand Laforge. (2026). From Student Risk Prediction to SC2R: Semantics-Constrained Counterfactual Recourse for Educational Decision Support. arXiv preprint. https://arxiv.org/abs/2608.17618

Page metadata

Field	Value
type	design
id	oulad-snapshot-recourse-evaluation-protocol
title	OULAD snapshot evaluation protocol with leakage-excluded presentation-level split
description	The evaluation protocol constructs two snapshots per eligible student per assessment, at t = d-14 and t = d-7 before the due day, using only information available up to snapshot time; post-t interactions and assessmen...
status	review
generated	by process:settle-candidates on 2026-10-11
trust	unverified — no one has confirmed this page's content
sources	1 — ngoc-luyen-le-2026

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