The Tests, and Why They Disagree
Every system deciding these questions has built a test, no two are identical, and the shared factors are weighted differently enough to give other answers.
Reads as a businessReads as employmentCuts both waysDepends where you are
Four factors, and no total. Every bar is the same length: the side is the claim, the size is not. The section that will not give you an answer, because nothing general can.
Each system that has to decide these questions has built a test. No two are identical, and the factors they have in common are weighted differently enough that the same arrangement produces different answers in different places.
The practical issue in “The Tests, and Why They Disagree” becomes clearer when expectations and evidence are visible to both sides. Used for fte meaning, see the product information here can add time and project context, provided the purpose is disclosed, access is limited and every consequential interpretation receives human review.
What follows describes the shape of such tests in general terms. It is not a test, and it cannot be applied to anything.
Teams documenting “The Tests, and Why They Disagree” can also review Loom asynchronous-work resources. Record which ideas fit the engagement and which still need confirmation from the contract owner or local adviser.
The factors that appear nearly everywhere
Across very different systems, a recognisable set recurs:
Who controls how the work is done. Whether somebody else may be sent. Who supplies the equipment. Who carries the financial risk. How integrated the person is. Whether work must be offered and accepted. How long it has run.
The recurrence is why the device at the top of every page in this collection uses them. It is not a claim that any of them decides anything.
Why no two tests agree
Because they were built at different times, for different purposes, by different institutions, against different industries.
A test developed to decide tax questions in the 1960s, a statutory definition written for platform work in the 2020s, and a judicial test refined over decades of litigation are not going to coincide, and nothing requires them to.
Weighting, which is where the real difference is
Most systems agree that control of method matters. They disagree sharply about how much, and about what can outweigh it.
Some treat a genuine right of substitution as close to decisive. Others barely look at it if it has never been exercised. Some weigh economic dependence heavily; others treat it as context. The factors are familiar and the arithmetic is not shared.
Multi-factor means no threshold
The single most important property of these tests is that they do not produce a score. There is no number of factors that settles anything, and no combination that guarantees an outcome.
That is why the device in this collection refuses to total anything. A sum would be the one figure a reader would act on, and it would be an invention.
Substance over form
The principle most systems share: what actually happened outweighs what the parties called it.
This is why so much of this collection is about ordinary operational detail — who sits on which rota, who was in which meeting, who was given which laptop — rather than about drafting. The detail is the substance.
Statutory answers for particular sectors
An increasing amount of this is being decided by legislation rather than by tests, particularly for platform and gig work.
Where that has happened, the general factors may matter much less than whatever the statute says, and the statute may say something quite different from what a court would have decided. It also changes, which is covered in its own article.
The test applied to a moment or to a period
A detail that catches businesses out. Some determinations look at the arrangement as a whole; others look at a specific engagement, a specific period, or even each individual assignment.
An arrangement that was clearly independent in its first year and much less so in its fourth may be assessed either as a whole or period-by-period, and which one applies affects both the answer and the exposure.
Why reading about another country's test is dangerous
Because the vocabulary is shared and the content is not. The same words appear in several systems with different meanings and different weight, and material written for one jurisdiction reads as though it applies generally.
A business reasoning from an article about somewhere else is likely to be confidently wrong, and this is the specific failure this section exists to warn against.
Intermediate categories
A feature of several systems that confuses businesses used to a binary choice. Between employee and independent business there may be a third status — worker, dependent contractor, economically dependent self-employed, under various names — carrying some protections and not others.
Where such a category exists, an arrangement can fail to be employment and still fall inside it, which means a business that has reasoned only about the two ends has not answered the question it thought it had.
What to do with all this
Use the common factors as a checklist for noticing what your own practices are doing. That is what they are good for and it requires no jurisdiction at all.
Then take the actual question — which test applies to us, here, for this purpose — to somebody qualified in that place, and expect the answer to be specific to it. Anybody who gives you a general answer to a question of this kind is telling you something about themselves rather than about your arrangement.