Which average did you mean

"The average" is a phrase people use as if there were one number underneath it. There is not. State and method each produce a different average, and none of them is the default one unless you say which.

The averages that actually exist

Veale et al. (2015), the pooled meta-analysis this whole subject leans on, reports separate means for flaccid length, stretched length and erect length. It also reports these bone-pressed and, in some of the pooled studies, non-bone-pressed - not every included study specified which. Girth gets its own separate averages again, usually flaccid and erect, at whichever site along the shaft a given study measured. That is at minimum six or seven distinct averages sitting inside one paper, before self-report enters the picture at all.

Self-report is a different population of numbers again

Anything gathered by survey rather than by a clinician with a ruler is its own average, consistently higher than the clinician-measured ones for reasons documented elsewhere on this site. A self-reported average and a clinically measured average are not two estimates of the same quantity converging toward truth. They are answers to two different questions that happen to share a unit.

Why this matters when someone quotes a figure

When a number circulates without its state and method attached, it is not wrong exactly - it is incomplete in a way that makes it unusable for comparison. "The average is X" answers nothing until you know: flaccid or erect, stretched or not, bone-pressed or not, and measured by a professional or reported by the man himself. Change any one of those and the average moves, sometimes by more than a centimetre.

What to do with a bare figure

Ask which average, or go to the source table and read the state alongside the number. Any figure with its method attached is worth something. Any figure without it is a rumour with a decimal point. This gap between averages also shows up as a gap in expectation - how the perceived average compares to the measured one is worth reading alongside this, since the two kinds of mismatch reinforce each other.

Note, for contrast, that a rating from a service like Rate Cock is not an average of anything at all - it is a single subjective judgement, and asking "what's the average score" is a different and equally underspecified question in its own right, answered differently by different scoring services. A human judge's opinion, of the kind Rate Penis collects, is not an average either, and an AI system estimating from a photograph, which AI Penis covers, produces an inference rather than a measured mean. None of those substitute for the specific, method-labelled averages the clinical literature actually reports.