Gear
Test your instruments on something else first
Measure a jar or a bottle with your tape, a string and a paper strip before measuring yourself; the spread you get is your instrument error, isolated from technique.
Guides on Gear: Every instrument, and what each costs you, Everything you can wrap, and what each reads, Verifying the thing you measure with
A rigid cylinder does not compress under a tape, does not change shape between readings, and does not care how you feel about the number. That makes it a better test object than your own body for one specific question: how much of your measurement error is coming from the instrument, as opposed to everything else.
Why a cylinder isolates the instrument
On a real body, technique and instrument error mix together in a way that is hard to pull apart afterward - a spread between readings could be inconsistent pressure, a wandering site, or the tape itself. A jar, a bottle, or a length of pipe removes the technique variables almost entirely: its circumference does not change, its surface does not compress, and there is no state to hold constant. Whatever spread you get between instruments on the same rigid object is, to a good approximation, the instruments disagreeing with each other - not you disagreeing with yourself.
What to use
Pick a cylindrical object with a circumference somewhere near the range you would measure on yourself - a wide marker pen, a small jar, or a bottle neck all work. Mark one point on it so every wrap sits at the same site, and pick something with a firm rather than a soft surface, since the point of the test is a shape that cannot compress or shift between one instrument and the next.
The test
Wrap a cloth or fibreglass tape around the marked point, snug against the surface with no gap, and read it. Do the same with a length of string, marking the overlap and laying the string flat against a rigid ruler. Do it again with a strip of paper, marked the same way. Take three readings with each instrument, not just one, since a single reading cannot tell you whether that instrument is consistent with itself.
Recording it
| Instrument | Reading 1 | Reading 2 | Reading 3 | Median |
|---|---|---|---|---|
| Cloth or fibreglass tape | ||||
| String + ruler | ||||
| Paper strip + ruler |
Fill in your own numbers. The three medians tell you the spread between instruments on an object that cannot move; the three readings within each row tell you how repeatable that one instrument is on its own.
Reading the result
If the three medians land within a millimetre or two of each other, your instruments agree well enough that the choice between them barely matters, which is broadly what the published error bars on this site's method posts assume. If one instrument is consistently high or low relative to the other two, you now know its direction of error on a fixed object, which is the same kind of systematic offset this site catalogues instrument by instrument - a stretched tape, for instance, will read the jar long the same way it would read anything else long. A wide spread within a single instrument's own three readings, on an object that is not moving or compressing, points at technique in the wrapping rather than the instrument - which is useful to know before blaming the tape for something the tape did not do.
What this test does not replace
This isolates instrument error; it does not remove the technique variables that only show up on a real, compressible, state-dependent body - snug tension, site drift, and skin movement under the tape all still need attending to when you take the actual measurement. It is also not a substitute for checking an instrument against a truly known dimension, which a bank card's fixed size does more precisely than any household jar can. None of it applies to instruments that are not physical at all - an image model returning a figure from a photograph has no wrap or tension to test, because it is inferring rather than measuring in this sense, and a rating tool's score is not something a kitchen-table test could calibrate, since it is a position in a distribution of images rather than a physical reading. A person's stated opinion is even further outside this frame - an honest human impression is not an instrument with a spread to test, it is a judgement stated once - and Rate Cock's rating sits in that same non-physical category, which is worth being clear-eyed about before comparing it to anything a tape measure could confirm.