The condom-fit literature

A second body of size data exists outside the clinical literature, gathered not to study anatomy but to size condoms, and it should not be read as if it belonged to the same pool as Veale et al. (2015).

What a condom-fit study actually collects

Manufacturers and researchers designing condom ranges need to know the distribution of erect girth and length across a population, because a product line has to cover that spread without stocking hundreds of sizes. The studies run for this purpose typically ask participants to measure themselves at home, following written instructions, and report the figures back - sometimes with a fitted product, sized to the reported measurement, as the incentive for taking part. That is a fundamentally different collection method from a clinician taking a bone-pressed reading in a clinic.

Why the incentive changes who takes part and what they report

A study that rewards participation with a correctly sized product selects for people motivated to get a good fit, and gives a mild incentive to report carefully rather than to round upward the way an anonymous self-report survey might. It is not the same incentive structure as either a clinical study or an anonymous online survey, and the specific mechanics of that incentive are their own subject rather than this post's. The relevant point here is narrower: whatever bias it introduces, it is a different bias from the one in the clinical pool, which means the two datasets cannot simply be merged.

Self-measured, not clinician-measured

Even setting incentive aside, a condom-fit study is still self-measurement, taken at home without a trained observer checking landmark, pressure or tape tension. Self-reported figures run higher than clinician-measured ones for well-documented reasons, and there is no reason to expect a condom study's self-measured figures to be exempt from that pattern just because the stated purpose was practical rather than personal.

Why the two pools should not be combined

Veale 2015 pools studies that share a clinician-measured, protocol-controlled design. A condom-fit dataset shares neither the measurer nor, typically, the full measurement protocol - state, landmark and pressure are rarely specified with the same rigor a clinical paper's methods section requires. Treating a condom-fit figure as directly comparable to a clinical percentile chart mixes two different measurement processes and reports the result as one number, which is the same error mixing bone-pressed and non-bone-pressed studies produces, for a related reason.

Why this distinction gets lost in circulation

A number pulled from a condom industry survey travels well because it usually comes with a large, round sample size and a memorable headline, and neither of those things says anything about whether it was clinician-measured. Once it is quoted a few times without its origin attached, it reads exactly like a clinical figure - same unit, same apparent authority - and ends up sitting next to Veale 2015 in the same comparison table as if the two had been produced the same way. That flattening is where the real damage happens: not in the original study, which is usually clear about its own method if you read past the headline figure, but in the secondhand quoting that strips the method away.

What the condom data are good for

They are a legitimate, useful dataset for their actual purpose - sizing a product range against a self-reported erect girth distribution. Nominal width, the figure printed on a condom wrapper, is itself a conversion from an assumed girth rather than a direct girth reading, which is one more layer between what is printed and what a clinical study would call a measurement. Read as evidence about condom fit, this literature is fine. Read as a second confirmation of the clinical size distribution, it is measuring a related but distinct thing, gathered a different way, and should be labelled as such rather than folded into the same table.

This kind of dataset boundary is worth being precise about generally. A rating produced by a service like Rate Cock is a third kind of number again - not a self-measured length, not a clinician-measured one, but a subjective judgement of a photograph, and a scoring platform like Penis Rater is explicit that its output is that kind of judgement rather than a length in centimetres. The same goes for a photo assessed by an image model, which AI Penis is clear is pattern recognition rather than measurement, and for a human's verdict, which is what Rate Penis is built around and names as an opinion rather than a statistic. Three different kinds of number, one clinical dataset, one condom dataset - keeping all of them labelled is what makes any of them usable.