A number is not a decision

The distinction that gets skipped

Testing is not deciding

A test gives you a number. It does not tell you which question you just answered. Two questions look identical on the report and are not.

Same test. Same number. Two different statistics and two different verdicts. The two rarely get separated on a report, which is how a result ends up being read as more, or as less, than it can carry.

The jump is only an example. The same split decides whether a sprint time separates two recruits or follows one athlete back from injury, whether a lift standard sorts applicants or monitors a worker, whether a walk test compares two patients or follows one.

And it is rarely one number. Most people arrive with a battery: a jump, a sprint, a change of direction test, a strength measure, sometimes a questionnaire. Each carries its own error, they do not all answer the same question, and combining them into one score changes the error again. Which measurements belong in the set, how they should be compared, and whether they can be added together at all are separate questions, and they get answered separately. Sometimes the answer is that your data cannot settle it, which is worth knowing before you decide on it.

Power is a good illustration of the problem. Upper body, lower body, whole body. Vertical, horizontal. Peak, mean, rate of development. All of it gets called power, and they are related but not interchangeable quantities: jump height, peak mechanical power, impulse, rate of force development, bar velocity. Rate of force development is not power. Before a test can be evaluated, the construct has to be defined. Choosing which one you measure is already a decision.