Looking down at the ruler

Self-measurement has a viewing problem that a clinician measuring someone else never has. Your eye is above and behind the ruler rather than level with it, and that changes what the scale appears to say.

Why the angle matters

A ruler read square-on, with the eye directly above the mark being read, gives the true position on the scale. Read from an angle - looking down and slightly back, which is the only vantage point available when measuring yourself standing up - and the sight line crosses the ruler's thickness at a slant. The tip's true position on the scale is not where it appears to line up from that angle; the apparent mark shifts along the ruler by a small amount that depends on how steep the viewing angle is and how thick the ruler's edge is.

This is parallax, the same effect that makes a car's analogue speedometer read differently from the driver's seat than from a passenger leaning across to look. On a ruler with millimetre markings, a shallow viewing angle over a ruler of ordinary thickness is worth a couple of millimetres - not large, but not nothing against a claimed precision of half a centimetre. A thicker ruler makes the effect worse, because the sight line has further to travel between the scale and the surface being read, which is one small argument for a thin, bevelled edge over a chunky one. Transparent rulers and parallax covers one instrument-side fix for the same problem, letting you sight straight down through the scale rather than across a thick edge. A ruler held slightly loosely, so it drifts a degree or two off the shaft's axis while you crane to read it, compounds the same error further, since now both the scale and the thing being measured are misaligned with the sight line at once.

Which direction it pushes

Looking down and back at your own ruler typically shifts the apparent tip position toward the far end of the scale from your eye, which reads slightly high. It is not guaranteed to run that way for every body and every ruler, but it is consistent for a given person and a given setup, which is exactly what makes it worth correcting rather than shrugging off as noise - a consistent bias does not average out across repeats the way random wobble does.

Three ways to remove it

Sight square to the ruler. Bend or crouch so your eye is directly above the tip mark, looking straight down the ruler's face rather than across it at an angle. This is the most direct fix and costs nothing.

Use a mirror set square to the ruler. A mirror placed so you view the ruler's reflection at a right angle removes the foreshortening differently - useful if bending far enough to sight square is awkward in the position you are measuring in.

Mark, then step back to read. Hold a fingernail or a small clip at the tip position, remove the ruler without disturbing the mark, and read the scale afterward from whatever angle is comfortable, since the reading no longer depends on sighting square while under tension. This is a smaller version of a fuller separation of marking from reading, covered in more detail elsewhere.

Where this sits next to the other self-measurement problems

Viewing angle is one specific, mechanical piece of the larger gap between a self-measured and a clinician-measured reading - a clinician standing beside someone else never has to solve this problem, because they are never looking at their own body from above. The most direct way around all of it, viewing angle included, is handing the ruler to someone else, which trades this bias for a different, smaller one.

None of this changes what the number is once you have it correctly: a distance in centimetres, not an impression. An impression is what you get from a human judge's read of a photo or from a score built on a set of images, and neither one has a ruler to sight square to in the first place - an image model has the same limitation for a structural reason, since a photograph carries no scale for it to recover a length from. If a subjective judgement is what you actually want rather than a corrected reading, Rate Cock is answering that different question; a square sight line only helps with this one.