Synthetic cohorts reproduce the level and shape of the weekly β0 partition but vary 2.6–4.9 times less across the term than real cohorts at a common working point¶

Claim · All claims
Evidence · 1 study (2 entries) · design r2 · q2
Currency · newest study 2026 · checked against its sources 2026-10-20 (automated, 2 of 2 entries) · not reviewed by a person · recheck by 2029: tool and model accuracy changes quickly

Subclaims¶

q2 i? In every cohort the real weekly β0 curve moves between 2.6 and 4.9 times as much across the term as the synthetic curve, with no overlap in coefficient-of-variation ranges. → Inoue 2026 q2 i? The separation holds at a pre-specified data-independent threshold grid: across sixteen synthetic and sixteen real values no synthetic value reaches any real value. → Inoue 2026 (2)

Evidence¶

Inoue 2026¶

Inoue, H., & Yasutake, K. (2026). What Fidelity Metrics Miss: A Structural Check on Synthetic Educational Data. Proceedings of the 34th International Conference on Computers in Education. Asia-Pacific Society for Computers in Education. https://arxiv.org/abs/2609.27265 doi:10.48550/arxiv.2609.27265

q2 · i? · design · r2

Analysis of four annual cohorts (2022–2025, 117–120 learners each) of lower-secondary study-habit logs from ReLEAF, comparing coefficient of variation of the weekly β0 curve at a common working point ε = 0.68 (Table 2). The article reports "The two ranges do not overlap."

"The two ranges do not overlap. In every cohort the real curve moves between 2.6 and 4.9 times as much across the term as the synthetic curve."

Inoue 2026 (2)¶

Inoue, H., & Yasutake, K. (2026). What Fidelity Metrics Miss: A Structural Check on Synthetic Educational Data. Proceedings of the 34th International Conference on Computers in Education. Asia-Pacific Society for Computers in Education. https://arxiv.org/abs/2609.27265. doi:10.48550/arxiv.2609.27265

q2 · i? · design · r2

Robustness check at the pre-specified reference grid of 0.40, 0.60, 0.80 and 1.00: synthetic coefficients fall between 0.05 and 0.17 and real ones between 0.34 and 0.54, so "no synthetic value reaches any real value."

"Across the four cohorts and the four thresholds we have sixteen synthetic values against sixteen real values, and no synthetic value reaches any real value."

Discussion¶

Related Claims¶

• Real cohorts separate from marginal-preserving surrogates of themselves in all four years, while three of four synthetic cohorts do not — related
• Synthetic cohorts place week-to-week change in the term's examination weeks, which real cohorts do not — related
• Partition descriptors and the component-membership–study-volume association are reproduced by the synthetic cohorts — related
• In a control experiment on data of known construction, the check is calibrated, can be passed, and separability from marginals is not fidelity — related
• The variation gap traces to what the generator was disclosed: no released statistic constrains which band and which week are zero together — related
• The routine rule for selecting the graph threshold makes naive comparisons of β0 between two datasets invalid — related