The jaws press

A calliper closes until its jaws meet the object, and on something rigid that is a well-defined stop. On soft tissue there is no such stop - the jaws keep closing, compressing what they touch, for as long as you keep applying force, and the reading you get depends on how hard you closed them rather than on a fixed physical boundary.

Why this is different from bone-pressed length

A bone-pressed length reading has a genuine stop: the ruler meets the pubic bone, and pressing harder past that point does not move the reading, as the standard method describes. A calliper on the shaft has nothing equivalent. Tissue compresses continuously as force increases, with no point at which further pressure stops changing the number - which means "how firmly did you close it" is doing the same job that "did you press to bone" does for length, except with no way to standardise the answer the way pressing to bone standardises pressure there.

Which direction it moves the reading, and how much

More closing force always reads smaller, because compressed tissue occupies less space under the jaws than relaxed tissue does. The size of the effect depends on the tissue and the force, but it is not a small correction at the margins - a firm close versus a light touch on the same spot can move a diameter reading by a meaningful fraction of a millimetre, which then compounds with the shape error covered in Cross-Sections Are Not Circles once that diameter gets converted toward a circumference.

A light-touch technique, if you use a calliper at all

Close the jaws only until they first make contact across the full width being measured, with no additional force past that point - the same "just touching, not compressing" specification a tape uses, described for the tape case in the instrument comparison. Take several readings, releasing and re-closing between each rather than holding one closure, and use the median rather than a single number - the same repeat-and-median principle that governs every measurement on this site, because any one reading can land anywhere in its own error band. If the readings from repeat closures spread by more than a small fraction of a millimetre, the closing force is varying, and that is the thing to fix before trusting any single number from the instrument. Practising the same light closure on something firmer first - a pencil, a coin edge - builds the feel for "just touching" before applying it to tissue that will not tell you when you have overshot.

How to notice it is happening

A practical test: take a reading, then take a second at the same spot pressing noticeably harder, and compare. If the two readings differ, compression is a live source of error for that particular jaw-tissue combination, and the gap between the light and firm readings is a rough, honest estimate of how much uncertainty compression alone is adding - larger than that gap is not a plausible true value, and neither reading on its own should be treated as exact. This is worth doing once with any calliper before trusting a single reading from it, the same way checking a tape for stretch is worth doing once before trusting a tape.

Why a tape mostly avoids this

A tape wrapped snug does compress slightly too, but "snug" is a state you can hold steady by feel in a way that a calliper's continuous, force-dependent closure is not - there is a qualitative difference between tensioning a flexible loop to lie flat and squeezing two rigid jaws together with no stop. That difference is part of why girth is conventionally taken with a tape rather than a calliper at all.

What a compressed reading still is not

However carefully the jaws are closed, the result is a length with an uncertainty, not a judgement of any kind. A photograph scored by something like Rate Cock is producing a different category of number entirely - a rating rather than a distance - sitting on a distribution of photographs rather than a jaw closure, and AI Penis explains why no amount of careful photography recovers an actual measurement the way a properly used calliper, however imperfect, at least attempts to. A human judge's score on Rate Penis is an opinion, not a reading in millimetres, whatever instrument sat on the desk next to it.