One decimal in the table

A published mean reported as, say, 13.1 cm is not claiming the true figure is exactly 13.1 and nothing else. The decimal places are a statement about how precisely the data support the number, and treating them as exact in either direction is a small but real misreading.

How many decimals a mean can carry

The metrology rule is that a figure should be rounded to match its uncertainty: the Guide to the Expression of Uncertainty in Measurement (JCGM 100:2008, §7.2.6) says estimates "should be rounded to be consistent with their uncertainties". One decimal in centimetres is about what a single careful clinical reading can resolve. A pooled mean can carry a little more, because averaging shrinks its uncertainty: Veale et al. (2015) report erect length as 13.12 cm (SD 1.66) across 692 men, and with a standard error of about 0.06 cm that second decimal is roughly the edge of what the data support, not a claim beyond it.

Where it goes wrong downstream

The common misreading is not underestimating a published figure's precision - it is overestimating it. A figure converted from centimetres to inches and reported to two decimal places is the clearest version of this mistake: converting 13.1 cm gives 5.16 inches by arithmetic, but those two decimal places carry no more real precision than the original figure had, because the conversion cannot manufacture information that was not there. A number that looks more precise after conversion is not more precise - it is the same precision, dressed up.

Rounding is not hiding anything sinister

This is not a criticism of the papers. Matching the decimals to the uncertainty is the appropriate choice, and the pooled clinical literature mostly reports means at one or two decimals for that reason. The point is narrower: read a published mean as good to about the stated decimal, not finer, and be suspicious of any figure - your own included - that claims more precision than its measurement process could have produced.

The same logic applies to your own readings

If three readings taken by hand spread by half a centimetre, reporting the result to a hundredth of a centimetre is the same category of error as reporting a converted inch figure to two decimals - the extra digits are decoration rather than information, borrowed from arithmetic rather than earned by measurement.

A published mean is a different kind of number from a rating. Rate Cock and a scoring platform like Penis Rater report a score, not a mean with an implied precision from pooled measurements, so this specific caution about decimal places does not carry over to a photo-based judgement of the kind Rate Penis or AI Penis produce - a score's precision is a different question, worth asking of those sites on their own terms rather than by analogy to a published table.