Temporal Precedence Screen¶
Causal-plausibility filter — instantiates Lag Structure and Feedback Loop Identification
A check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect.
A Temporal Precedence Screen is a gate, not an estimator. Given a claimed lagged influence — "A at an earlier time drives B later" — it asks the minimal disqualifying question: does the supposed cause actually and reliably come before the effect, once reporting artifacts and shared drivers are stripped out? If precedence fails, the claim is rejected before any effect size is entertained. Its defining move is pass/fail filtering on ordering: it does not measure how strong a lag is or what shape it has; it tests whether the candidate lag even survives the requirement that causes precede effects. It is the discipline that keeps a lag analysis from promoting a coincidence, a reversed arrow, or a reporting-delay illusion into a causal story.
Example¶
A pharmacoepidemiologist reviews a claim that a certain heartburn drug "causes" a later stomach-cancer diagnosis, based on records showing the prescription preceding the diagnosis by months. The screen interrogates the precedence. Cancer is slow and often symptomatic before diagnosis — indigestion is an early symptom — so patients may have been prescribed the drug for the early symptoms of the very cancer later diagnosed. The apparent "cause precedes effect" is a reporting-lag illusion: the disease predated both. This is textbook protopathic bias.[n1] The screen also checks whether a common driver (an underlying condition prompting both the prescription and the workup) generates the ordering. Precedence, properly adjusted, does not hold in the causal direction claimed — so the screen rejects the lagged-influence claim rather than sizing it.
How it works¶
- State the candidate lag. Take the specific claimed influence and its direction and lag from the lag candidate set to be screened.
- Adjust for reporting/measurement lag. Correct for delays between when something happened and when it was recorded, since these can manufacture false precedence (or hide true precedence).
- Rule out shared drivers. Check whether a common upstream cause produces the apparent ordering, separating a genuine cause-before-effect from an endogenous artifact.
- Pass or reject. If the cause reliably precedes the effect after those adjustments, the candidate survives to estimation; if not, it is rejected as implausible — the screen returns a verdict, not a coefficient.
Tuning parameters¶
- Precedence-reliability bar — how consistently the cause must precede the effect to pass; a strict bar rejects more (fewer false positives, some real-but-noisy leads lost), a lax bar admits weak orderings.
- Reporting-lag correction — how aggressively to adjust recorded times back toward event times; over-correcting can invent precedence, under-correcting can hide reversal.
- Common-driver set — which shared causes to test against; broader is safer but can reject a real lead that merely correlates with a confounder.
- Latency plausibility window — the biologically/physically credible lag range; a claimed lag far outside it is grounds for rejection regardless of statistical fit.
When it helps, and when it misleads¶
Its strength is cheap disqualification: before anyone invests in modeling a lag, the screen throws out the reversed arrows, the reporting-delay illusions, and the common-cause orderings that would otherwise become confident false findings. It is the guardrail that enforces temporality — the one causal criterion that is genuinely necessary.[n1]
Its failure mode is the boundary of what precedence can prove: passing the screen establishes only that cause-before-effect is not violated, never that causation holds — a spurious relationship with the right timing sails through. Set the precedence bar too strict and it rejects real but noisily-timed leads; too lax and it waves through coincidences. The classic misuse is inverting the tool — reading "passed the precedence screen" as "confirmed causal," when the screen is purely eliminative. The guarding discipline is to treat a pass as necessary but not sufficient, hand survivors to estimation and confounding control rather than declaring victory, and keep the reporting-lag adjustment honest so the screen isn't fooled by the very artifact it exists to catch.
How it implements the components¶
lag_candidate_set— it operates on the specific candidate lags proposed, screening each for precedence.endogenous_vs_exogenous_driver_separation— ruling out a shared upstream cause of the apparent ordering is exactly this separation, in service of the pass/fail verdict.measurement_lag_adjustment— correcting recorded-vs-event timing is what stops reporting delay from faking or hiding precedence.
This screen renders a pass/fail verdict on ordering; it does not estimate the sized, reciprocal structure of a cross-lag — it does not implement cross_lag_variable_set or dependency_order_estimation, which are Cross-Lagged Dependency Review's. Where that review measures how strongly and in which direction two variables lead each other, this screen only decides whether a claimed lead is admissible at all.
Related¶
- Instantiates: Lag Structure and Feedback Loop Identification — it supplies the eliminative precedence-check gate ahead of estimation.
- Sibling mechanisms: Cross-Lagged Dependency Review · Autoregressive Dependency Map · Distributed Lag Model · Causal Loop Diagram with Delay Marks
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Temporal Precedence Screen is defined in the frozen evidence as: A check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect. Its operative deployed or enacted form is therefore Assessment, Review & Assurance.
Nearest alternative: Analysis, Modeling & Optimization — Analysis, Modeling & Optimization can support this mechanism, but the evidence centers the concrete operation described above rather than the alternative family's defining operation.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Single lineage
Present-day reach: Universal
Rationale: The defining operation is: A check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect. In the statistics_experimental_design lineage, that operation is specifically evidenced by authoritative or primary work that states temporality as indispensable: a proposed causal factor must precede the effect. This makes statistics_experimental_design the best historical origin, while the retained alternates document contributing methods and later applications rather than being mistaken for coequal origins.
Related originating lineages:
- Data Science & Analytics — Data science, analytics, and operational monitoring supplies a parallel or contributing lineage for the mechanism's defining operation: a check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect.
- Mathematics — Mathematical modeling, proof, and abstract-structure practice supplies a parallel or contributing lineage for the mechanism's defining operation: a check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect.
- Organizational & Management Science — organizational_management supplies a historically relevant parallel or contributing practice for the defining operation—A check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect—but the evidence does not make it the best primary lineage.
- Philosophy — Philosophical logic, epistemology, and normative reasoning supplies a parallel or contributing lineage for the mechanism's defining operation: a check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect.
- Psychology — Experimental, clinical, and behavioral psychology supplies a parallel or contributing lineage for the mechanism's defining operation: a check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect.
- Systems Thinking & Cybernetics — Systems science's feedback, stock-flow, boundary, and regulation tradition provides a formative adjacent lineage for the same temporal precedence screen operation.
Review resolution: The blind reviewers disagree on primary lineage (organizational_management versus statistics_experimental_design), so I adjudicated the mechanism rather than inheriting either label. The defining operation is: A check that rejects claimed lagged influence when the proposed cause does not reliably precede the effect. In the statistics_experimental_design lineage, that operation is specifically evidenced by authoritative or primary work that states temporality as indispensable: a proposed causal factor must precede the effect. This makes statistics_experimental_design the best historical origin, while the retained alternates document contributing methods and later applications rather than being mistaken for coequal origins. The cited Hill, The Environment and Disease: Association or Causation? directly supports the mechanism-specific operation and its disciplinary lineage. I retain all independently explained historical alternates without a numeric cap. origin_mode=single_lineage records how the mechanism arose; domain_reach=universal separately records how broadly it can now be applied.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
Notes¶
[n1] Protopathic bias arises when a treatment is prescribed for the early, undiagnosed symptoms of the disease it later appears to cause, manufacturing a false cause-before-effect ordering. It is a concrete violation of the temporality criterion (the requirement, emphasized in the Bradford Hill considerations, that a cause precede its effect) — the exact artifact a temporal precedence screen exists to catch. ↩a ↩b