FA-12656 / Causal analysis data contracts / Open access

Global treatment overlap conceals an unsupported target stratum · case 01

Global treatment overlap conceals an unsupported target stratum.

Verified by executionVariant 1 · 6 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

Presence of both arms globally is mistaken for conditional design support.

VERIFIED REPAIR

Check both arms in every requested target stratum.

Unsuccessful approach: Checking only observed strata forgets target strata absent from the study.

Case contract

Rows [stratum,arm] encode observed design support. Return target strata lacking either arm, sorted. Multiplicity is irrelevant.

Why this case matters

A deterministic synthetic study model isolates this data-contract defect; outputs alone establish no real-world causal identification.

1 / The failure

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(rows, target):
    return [] if {r[1] for r in rows} == {0,1} else sorted(target)
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('separated arms', solve([['a',0],['b',1]], ['a','b']), ['a','b'])
check('absent target', solve([['a',0],['a',1]], ['a','b']), ['b'])
check('supported', solve([['a',0],['a',1]], ['a']), [])
check('empty target', solve([['a',0]], []), [])
check('empty study', solve([], ['a']), ['a'])
check('duplicate rows', solve([['a',0]]*N, ['a']), ['a'])
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
separated arms	`[]`	`['a', 'b']`	Failed
absent target	`[]`	`['b']`	Failed
supported	`[]`	`[]`	Passed
empty target	`[]`	`[]`	Passed
empty study	`['a']`	`['a']`	Passed
duplicate rows	`['a']`	`['a']`	Passed

SHA-256 / a7c2d64d48f1fea989a09c6c1cd344735f3384553a361f44769a028996570fc9

2 / The unsuccessful fix

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(rows, target):
    observed = {r[0] for r in rows}
    return sorted(s for s in target if s in observed and {a for t,a in rows if t == s} != {0,1})
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('separated arms', solve([['a',0],['b',1]], ['a','b']), ['a','b'])
check('absent target', solve([['a',0],['a',1]], ['a','b']), ['b'])
check('supported', solve([['a',0],['a',1]], ['a']), [])
check('empty target', solve([['a',0]], []), [])
check('empty study', solve([], ['a']), ['a'])
check('duplicate rows', solve([['a',0]]*N, ['a']), ['a'])
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
separated arms	`['a', 'b']`	`['a', 'b']`	Passed
absent target	`[]`	`['b']`	Failed
supported	`[]`	`[]`	Passed
empty target	`[]`	`[]`	Passed
empty study	`[]`	`['a']`	Failed
duplicate rows	`['a']`	`['a']`	Passed

SHA-256 / ebe187a1a9e7f170dfc67a52dc0194a48e5dd51d035dd2541861e8965d5885e1

3 / The verified repair

Exit 0

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(rows, target):
    return sorted(s for s in target if {a for t,a in rows if t == s} != {0,1})
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('separated arms', solve([['a',0],['b',1]], ['a','b']), ['a','b'])
check('absent target', solve([['a',0],['a',1]], ['a','b']), ['b'])
check('supported', solve([['a',0],['a',1]], ['a']), [])
check('empty target', solve([['a',0]], []), [])
check('empty study', solve([], ['a']), ['a'])
check('duplicate rows', solve([['a',0]]*N, ['a']), ['a'])
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
separated arms	`['a', 'b']`	`['a', 'b']`	Passed
absent target	`['b']`	`['b']`	Passed
supported	`[]`	`[]`	Passed
empty target	`[]`	`[]`	Passed
empty study	`['a']`	`['a']`	Passed
duplicate rows	`['a']`	`['a']`	Passed

SHA-256 / ebe82f6ce4e05472f70779c360f856be76f494b15ca32df2f33b977b18358369

Verification & scope

Finite, fully specified synthetic data only; identification assumptions are supplied by the fixture design, not inferred from observations. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.

Observations recorded using Python 3.12.14 at 2026-09-29T14:38:58.794348+00:00.

Case digest / 9d748fe4e29ef9c073cfab033f82bb4fdb16356ae6dde63f6d92ff4572627400