What went into the pool

Veale et al. (2015), BJU International, pooled 15,521 men across 17 studies, and it did so under one governing rule: a study only went into the pool if the measurement was taken by a health professional, not reported by the participant. That rule is smaller than it sounds and it decides more about the resulting means than the anatomy does.

The inclusion rule

The paper's search covered studies of penile length and circumference in adult men, and from that wider set it kept the ones where a clinician, nurse or trained researcher applied the ruler or tape and recorded the figure themselves. Studies also needed to describe their measurement method well enough to classify it as flaccid, stretched, or erect, and along a defined line rather than an undefined "size". A study that reported a number but not how it was obtained did not qualify, however large its sample.

What that rule throws out

The obvious casualty is self-report: surveys where men measured themselves and wrote down a figure, unsupervised and unverified. Self-reported figures run consistently higher than clinician-measured ones, for reasons that are mostly about method rather than dishonesty, and pooling them in with clinician data would have pulled the mean upward for reasons that have nothing to do with the population being different. Excluding them is what makes the pooled figures usable as a reference at all.

It also throws out studies with vague or unverifiable protocols - a paper that says "erect length" without saying whether the ruler pressed to bone, or whether the man was standing or supine, gives a number that cannot be classified into the categories the nomograms use.

The scope beyond the measurement rule

The paper's search was not limited to one country or one language - it pooled studies conducted across multiple countries, each contributing whichever measurements its own protocol reported, provided the clinician-measured rule was met. That breadth is part of the trade the inclusion rule makes: a wider geographic base of clinician-measured studies rather than a narrower base that might otherwise have let self-report in. The paper also required participants to be adults, which excludes any data on younger age groups from the pooled figures entirely - a boundary worth noting for anyone trying to apply the nomograms outside the adult population they were built from.

The population this leaves you with

The men in the pool were not sampled from the general population by lottery. Most came from clinic settings - urology, andrology, or research studies recruiting through hospitals - because that is where a health professional with a ruler and a study protocol is standing. Clinic attenders are not a random draw, and the direction of any resulting skew is a caveat that travels with the pooled mean rather than a flaw specific to this one paper - every measured study inherits the population it happened to recruit from. The paper is explicit that this is the trade it made: narrower, less representative recruitment, in exchange for a measurement method you can trust.

Reading the pooled figures with this in mind

None of this means the numbers are wrong. It means they answer a specific question - what does a careful clinician's ruler find in the population that ends up in these studies - and that question is different from "what is the true average across all men", which no study has actually answered and probably cannot. The full walkthrough of what the paper set out to do and where its authors urge caution covers the rest of the protocol; this is only the one rule that decided who was in the room.

That distinction between a measured figure and a subjective one comes up constantly once you start comparing numbers to opinions. A pooled clinical mean is a statement about a ruler in a specific population; it is not the same kind of claim as a photo-based score from a site like Rate Cock, which is answering a different question entirely and does not draw on this literature at all. It is also a different question from what an AI model does with a photograph - recognising visual pattern rather than reading a landmark - or from what a human judge's impression amounts to, which is confidence rather than measurement. If you want the score-style version of a number rather than a pooled clinical mean, that is a different product built for a different question, and neither substitutes for the other.