← Every exhibit

Figure Figure 1 From the research bench

The external-validity transfer matrix

A result can be internally credible and still need a new argument before it is transported to a different target.

A six-row transfer worksheet with seven columns for ID, source or target object, treatment and comparator, outcome and time window, context, changed field, and transfer verdict.IDStablerowidentifier.SOURCE ORTARGET OBJECTSource result ornamed target.TREATMENT ANDCOMPARATORWhat was done, andagainst what?OUTCOME ANDWINDOWWhich event orconstruct, andwhen?CONTEXTWhich operatingor institutionalsetting?CHANGED FIELDPopulation,treatment, outcome,context, none;bridge?VERDICTReplicate,transport,limit or not?Six rows are deliberate: one synthetic source result, one replication candidate, and four one-field target changes. Theillustrative plus-eight-point effect is not a benchmark.
Vector SVG Cite Embed

Reference & Evidence

Source: Author's synthetic transfer matrix grounded in Egami and Hartman (2023) and Campbell (1979). All rows, effects, and decisions are illustrative.

Each line is a claim from the register this journal publishes against, resolved from the register at build time.

  • A "we propose a formal framework that encompasses four dimensions of external validity x t y and c validity populations treatments outcomes and contexts" Egami & Hartman. (2023) · EH23-C1
  • A Two goals, not one: "effect generalization generalizing the magnitude of causal effects" and "sign generalization generalizing the direction of causal effects", each with its own estimators and assumptions Egami & Hartman. (2023) · EH23-C2
  • A "effect generalization generalizing the magnitude of causal effects", for which the authors "introduce three estimators of the target population causal effects" Egami & Hartman. (2023) · EH23-C3
  • A "sign generalization generalizing the direction of causal effects", for which they "propose a novel multiple testing procedure under weaker assumptions": a different target needs a different tool Egami & Hartman. (2023) · EH23-C4
  • B External validity is target-relative; a result does not possess universal transfer by itself Author framework · R01-OWN-C1
  • B Changing one transfer field changes the target claim boundary before a verdict is offered Author boundary · R01-OWN-C3
  • B A self-serve small-business experiment does not automatically answer an enterprise implementation target Author synthetic illustration · R01-OWN-C4
  • B Replication, transport, and analogy are different decisions with different evidentiary burdens Author framework · R01-OWN-C7

Grades: A, verified against the printed page of the primary source · B, primary source, text layer only · C, authoritative secondary · D, reported.

Related exhibits