The ratio

A length-to-girth ratio made by dividing the published mean length by the mean girth is arithmetically valid and statistically wrong: the ratio of two averages is not the average man's ratio. The number shows up constantly in forum threads, and the same error appears anywhere two averaged quantities get divided.

The ratio of means is not the mean of the ratio

Take a population of men. Each one has his own length and his own girth, and therefore his own personal ratio of the two. The mean of those individual ratios - average every man's own length divided by his own girth, then average that - is a different number from the ratio of the two population means - divide the average length by the average girth once, using the pooled figures.

These are not the same calculation, and in general they do not produce the same answer. The gap between them grows with how variable the denominator is: because girth varies less between men than length does, and because dividing by a number near its own mean behaves differently than dividing by numbers scattered further from it, the ratio-of-means figure quoted online systematically misrepresents what a typical individual ratio actually looks like. This is a general property of ratios of correlated but non-identical variables, not something specific to this dataset - the same distortion appears in any statistic built by dividing one average by another.

What the literature actually reports

Veale et al. (2015), the meta-analysis this site treats as the standard reference, reports means and spreads for length and girth as separate figures - 13.12 cm (SD 1.66) for erect length across 692 men, and 11.66 cm (SD 1.10) for erect circumference across 381. It does not report a pooled ratio statistic, and there is a reason for that: a meaningful ratio figure would need to be computed from paired measurements on the same men, tracking each individual's own length against his own girth, rather than from two independently pooled averages that, as those different sample sizes show, do not come from the same men. Anywhere you see a single ratio figure attributed to "the research," it has been constructed after the fact from the two published means, not reported directly by the paper, and it carries the ratio-of-means distortion described above.

Why this particular ratio gets quoted so much

A single number that appears to summarise "shape" is attractive, and forums like round comparable figures. But length and girth are only moderately related within any one man - this site covers that relationship on its own terms elsewhere - so a population-level ratio says very little about what any individual's own proportions look like, on top of the statistical problem with how the ratio was built in the first place. Two men can both be close to the population means individually and still have noticeably different personal ratios, because girth and length do not move together tightly enough for one to imply the other, let alone for a single quoted fraction to describe both at once.

What to do if you want your own ratio

If you want your own length-to-girth relationship, take both measurements carefully - the full method for length covers the technique for both - and divide your own two numbers. That gives you a real ratio for one man, which is a legitimate thing to compute. What it is not is comparable to a "typical ratio" pulled from someone dividing two population means online, because that figure was never a real individual's proportion to begin with.

What a ratio is not standing in for

None of this produces a judgement about shape or proportion in any evaluative sense - that is a different kind of question, closer to what a photograph-based platform like Rate Cock or a scored gallery like Penis Rater is actually assessing, and neither of those is doing arithmetic on a ratio at all. An image model attempting the same thing from a photo runs into the earlier problem twice over - it cannot recover either length or girth from a picture, so a ratio built from its output is not measuring anything a tape would recognise. And a human judge's sense of proportion, the kind Rate Penis is built around, is a visual impression rather than a computed fraction, which makes it a different question again from anything this post is describing.

If what you actually want is the underlying girth figures rather than a ratio built from them, the girth data from Veale 2015 are reported on their own, separately from length, exactly because combining them into a single fraction is where the trouble starts.