There are five ways to measure, and they give five different numbers

Almost every disagreement about this subject, online and off, has nothing to do with anatomy. It is two people using the word length to mean two different procedures and assuming the other is lying.

There are five common procedures. They are all called measuring. They produce systematically different answers, and the gaps between them are larger than most of the differences anyone argues about.

1. Flaccid length

The simplest to take and the least informative. Flaccid length responds to temperature, to anxiety, to how recently you exercised, to how long you have been standing in a cold room. Within a single man across a day it varies by more than the difference between many national averages.

It also correlates only weakly with erect length, which is the finding that most surprises people and the one that undermines every changing-room comparison ever made. The visible flaccid state is close to the least predictive measurement available.

2. Stretched length

The flaccid penis extended to its full length, measured from the pubic bone to the tip. This is the clinical workhorse and the measure this game is played on.

It is used because a clinician can take it reliably in a consulting room without waiting for an erection, because it is reproducible between examiners, and because it tracks erect length closely enough to stand in for it. It is also, crucially, the measure with wide enough coverage across countries to build a comparison from primary studies at all.

3. Erect length

What everyone actually means and what almost nobody measures well.

The difficulty is procedural. Obtaining a reliable erection in a clinical setting requires either pharmacological induction or a degree of privacy that undermines standardisation, and studies that ask men to measure themselves at home reintroduce every problem that clinician measurement was there to solve.

This is the reason so much of the literature reports stretched length. Not because erect length is uninteresting, but because stretched length can be measured consistently and erect length usually cannot.

5. Self-reported

The fifth procedure is a survey wearing the word measurement, and it behaves like a survey.

Self-reported figures run systematically higher than clinician-measured ones. Very little of this is dishonesty. It is the accumulation of small, individually forgivable choices: measuring at the most flattering moment, pressing rather more firmly than a clinician would, measuring along the top rather than the side, rounding up, and remembering the best of several attempts.

Each step is a fraction of a centimetre. Together they are the reason no serious review will pool self-reported data with measured data, and the reason any figure you encounter without a stated protocol should be assumed to be self-reported until proven otherwise.

How they relate

Directions and rough magnitudes from the pooled clinical literature. Individual studies vary.

The girth footnote

Everything above concerns length, because length is what the country data reports and therefore what this game can be played on.

Girth is measured too, is reported in the same pooled reviews, and is arguably the more relevant dimension for several practical questions, including condom fit. It gets a fraction of the attention, which is a reasonable summary of how much of the cultural noise around this subject is about anatomy at all.

How to use this

When you encounter a figure on this subject, the useful question is never whether it is high or low. It is: which of the five is this, and does the source say?

A source that states its protocol is doing the minimum required to be quotable. A source that does not is not necessarily wrong, but it has left out the single piece of information needed to compare its number with any other, and in practice the omission correlates strongly with having no primary source at all.

So each row in this game carries its method, and two countries whose figures differ by less than a centimetre should be read as agreeing rather than as ranked.