← Every exhibit
How should a team review a transfer claim?
| Pattern | First question | Smallest honest verdict |
|---|---|---|
| Same source objects and setting, new sample | Was the source design repeated without changing the estimand? | Replication candidate |
| New population with measured source-target overlap | Are the variables that modify the effect observed in both populations? | Transport may be analyzable under stated assumptions |
| New treatment implementation | Is the target intervention equivalent in the causal sense, not only in its label? | New treatment evidence required |
| New outcome or horizon | Does the target event measure the same construct at the same time boundary? | Do not infer the target outcome from the source result |
| New context or institution | Could the mechanism change with regulation, channel, language, or time? | Context bridge required |
| Similar label with no comparison record | Which fields are actually matched, and which are assumed? | Analogy only |
| Target data or overlap missing | Can the target claim be evaluated with the available design and data? | Not identified |
Swipe or scroll horizontally if the table is wider than your screen.
Reference & Evidence
Source: Table from this essay. Sources and interpretation are given in the article.
Related exhibits
-
Which four dimensions change the transfer question?
From the essay What is external validity? A result needs a transfer boundary
-
The external-validity transfer matrix
From the essay What is external validity? A result needs a transfer boundary
-
What does a transfer matrix look like?
From the essay What is external validity? A result needs a transfer boundary