Charge as a Construct-Validity Factor in Chinese Legal Case Retrieval: A Cross-Benchmark Audit
For researchers and practitioners evaluating legal case retrieval systems, this work reveals that current benchmarks may overestimate system performance due to charge-based construct validity, undermining claims of legal reasoning ability.
This paper audits Chinese Legal Case Retrieval (LCR) benchmarks and finds that most of the performance gap between BM25 and trained systems is due to a construct-validity confound: ranking by shared primary charge alone recovers 99.2% of the gap on LeCaRDv2, with no detectable difference from the best-trained system. The effect varies across benchmarks, and the authors release a reusable protocol to screen for such confounds.
Chinese Legal Case Retrieval (LCR) benchmarks grade a reference judgment relevant when its legal characterization matches the query, and strong systems now reach NDCG@10 of 0.85-0.88. Most of the BM25-to-best-trained gap is recoverable with no retrieval model: ranking candidates only by shared primary charge, broken by BM25, closes 99.2% of it on LeCaRDv2 -- with no detectable difference from the best-trained system. This reflects benchmark design: LeCaRDv2 defines top relevance via the crime's key constitutive elements, which encode the charge, so same-charge cases are relevant by construction (relevance lift 4.49; charge-to-relevance macro-AUC 0.871). Holding charge fixed, the trained reranker's advantage over BM25 collapses to a small within-charge residual (+0.026 NDCG@10, cluster-bootstrap CI excluding zero, about a quarter), the only non-definitional positive. The effect is not uniform: the same rule recovers 84.3% on LeCaRDv1 and is out of spec on CAIL2022, with the charge-to-relevance signal weakening in step (macro-AUC 0.871/0.759/0.728); a predicted-charge cascade reproduces 76.6% on LeCaRDv2 but does not transfer. The construct is also cashable at first stage: an exploratory zero-training charge-pool channel lifts LeCaRDv2 recall (R@100 +0.025, wrong-charge controls hurt), reported as a positive control for the confound, not a retrieval method or novelty claim. Charge is thus a high-leverage construct-validity factor at the benchmark level -- not auniform explanation of NDCG@10, and not evidence that any system relies on charge. We package established construct-validity and partial-input checks as a reusable charge-controlled protocol (CCE); on all three benchmarks its triggers come back null or descriptive, behaving as designed. We release the scripts, schema, and protocol so future benchmarks can be screened before their NDCG@10 is read as legal-reasoning ability.