Scope Boundary
What a measurement counts and what it leaves out, which is usually the difference between two figures that appear to contradict each other.
Reviewed July 16, 2026Stable
When not to use it
- Where both parties have stated their boundaries and genuinely disagree about the world, which is a substantive dispute rather than a definitional one.
- As a way to avoid taking a position. Identifying the boundary usually reveals which figure answers the question, and saying so is the point.
- Where the boundaries are the same and the measurements differ, which indicates an error somewhere rather than a definitional gap.
Reach for something else instead
- Multi-boundary reporting — publish the figure at each defensible boundary and let the reader select, which removes the strategic element entirely.
- Mandated definitions — a formal scheme, as with greenhouse gas scopes, which is why emissions figures are more comparable than water or energy ones.
- -
Read more on the blog
- 1 to 2% of the chips, about 30% of the tokensExport controls were designed to constrain compute and have. The measure that moved instead was usage, and the two have separated sharply.
- The good numbers all came from obligationsTerritory 8 closes. Across ten subjects the reliability of a figure tracked whether someone was legally required to produce it, and nothing else.
- The 13% traveled. The authors' caveat did not.A careful study found entry-level employment falling in AI-exposed jobs. Its own authors later narrowed when that becomes significant, and a serious alternative explanation predicts the same pattern.
- Three years in: no disruption, and one 20% holeEvery aggregate measure of AI's labour effect shows continuity. One within-firm comparison shows a fifth of a cohort gone. Both are well evidenced.
Further reading
- Raji et al. (2021), AI and the Everything in the Whole Wide World Benchmark — operationalisation determining what a measurement can support.
- Wong et al. (2021), External Validation of a Widely Implemented Proprietary Sepsis Prediction Model — the difference between overall performance and performance on the population that matters.
Primary sources, listed so you can check the claims on this page rather than take them on trust.
Where people go wrong
- Treating a large discrepancy as evidence that one source is dishonest.
- Pooling figures from sources with different boundaries because the methods look similar.
- Publishing a number without its boundary and assuming context will carry it, when the number travels and the context does not.
At a glance
FieldFoundations
Signallarge disagreement between credible sources
Common pairsdirect against total, marginal against aggregate, all cases against incremental
Formal precedentgreenhouse gas scopes 1, 2 and 3
Practical rulestate the boundary in the same sentence as the figure
DifficultyBeginner → Intermediate
Where this sits
A starting point. Nothing needs to come before it.
0Levelstarting point
0Needs firstnothing
0Opens upnothing further
1Areastays here
Computed from the prerequisite graph, not assigned. How this works