Effect-Level Validation for Causal Discovery
This addresses the problem of decision-making reliability in feedback-driven systems with self-selection for practitioners using causal discovery on telemetry data, representing an incremental improvement in validation methodology.
The paper tackles the problem of unreliable causal discovery in telemetry data by proposing an effect-centric validation framework that evaluates discovered graphs through identifiability, stability, and falsification rather than graph recovery accuracy alone. The result shows that many plausible discovery outputs fail to admit point-identified causal queries under constraints, but when identification is possible, different algorithms converge to similar effect estimates that survive refutation tests.
Causal discovery is increasingly applied to large-scale telemetry data to estimate the effects of user-facing interventions, yet its reliability for decision-making in feedback-driven systems with strong self-selection remains unclear. In this paper, we propose an effect-centric, admissibility-first framework that treats discovered graphs as structural hypotheses and evaluates them by identifiability, stability, and falsification rather than by graph recovery accuracy alone. Empirically, we study the effect of early exposure to competitive gameplay on short-term retention using real-world game telemetry. We find that many statistically plausible discovery outputs do not admit point-identified causal queries once minimal temporal and semantic constraints are enforced, highlighting identifiability as a critical bottleneck for decision support. When identification is possible, several algorithm families converge to similar, decision-consistent effect estimates despite producing substantially different graph structures, including cases where the direct treatment-outcome edge is absent and the effect is preserved through indirect causal pathways. These converging estimates survive placebo, subsampling, and sensitivity refutation. In contrast, other methods exhibit sporadic admissibility and threshold-sensitive or attenuated effects due to endpoint ambiguity. These results suggest that graph-level metrics alone are inadequate proxies for causal reliability for a given target query. Therefore, trustworthy causal conclusions in telemetry-driven systems require prioritizing admissibility and effect-level validation over causal structural recovery alone.