Source notes

A strength test can agree with another about who is stronger without giving the same answer about how strong someone is. This matters when replacing a maximum lift with a test against an immovable resistance. The paper challenges the assumption that a strong statistical correlation makes those methods interchangeable. In an illustrative dataset, a correlation of 0.70 still accompanied an average absolute percentage difference of about 17%. That example demonstrates a measurement problem, not a universal error rate for every device. The authors argue that agreement and the size of measurement differences need direct examination.

The tests also place different demands on the body. Holding a position against resistance differs from completing a lift through a range of motion. Joint angle, movement skill and familiarity can affect the result. Both methods may provide useful information while estimating different aspects of strength. A method's convenience or repeatability therefore does not establish that it can replace another for a particular question.

How Does This Show Up in Real Life?

Imagine following a training program with one test before it starts and a different test afterward. An apparent change could partly reflect the switch in method. Even when the same people tend to score highly on both tests, the difference between their results may be large enough to obscure the training change being assessed. The paper recommends selecting tests with the training task and research question in mind. For someone interested in a particular lift, a measurement in a fixed position may add useful information without reproducing performance across the whole movement. The practical lesson is to ask what the test measures and whether its sensitivity matches the change of interest. A familiar label such as maximal strength does not make every assessment equivalent.

The numerical demonstrations are illustrative rather than a pooled estimate of all strength tests. Studies vary in joint angles, movements and familiarization, and few directly assess agreement. Expected training changes also differ by setting, so there is no single acceptable error for every purpose. The paper evaluates measurement interpretation rather than demonstrating long-term benefits from choosing one testing strategy.

Why We Cite It

Alpekin can use this source when explaining the difference between a convenient proxy and the outcome someone actually cares about. It offers a specific way to examine claims that a quick measurement replaces a more involved assessment, including whether the evidence tests replacement directly or only reports an association.

Take the methodological warning seriously without concluding that either type of test is useless. A useful measurement is matched to its purpose. Confidence comes from understanding its agreement, precision and task demands, rather than from a reassuring correlation alone.