Height

Height and length are positively but loosely related: across pooled clinical studies, correlations run from r = 0.2 to 0.6. That is real across a population and close to useless for predicting any one man's length. What the data support is narrower than the folk version by a wide margin.

A weak positive correlation, in large samples

Where studies have tested for a relationship between height and length, the reported finding is a positive correlation - taller men are, on average and across a large enough sample, somewhat more likely to measure longer. Veale et al. (2015), pooling studies where health professionals took the measurements, found height gave the "consistent and strongest significant correlation" with stretched or erect length, ranging from r = 0.2 to 0.6 across studies. Squaring those figures gives the share of variation height could account for: about 4% at r = 0.2, and about a third at r = 0.6, leaving most of it unexplained even at the top of the range. The full walk-through of Veale 2015 covers the rest of what the dataset can and cannot support.

Why "real" and "useless for prediction" are both true at once

This is the part that trips people up about weak correlations generally, not just this one. A correlation can be statistically real - meaning it is very unlikely to be a fluke of the particular sample studied - while still being far too weak to use as a predictor for an individual. With a weak correlation, knowing one man's height narrows the plausible range of his length only slightly, and the overlap between "shorter men's" and "taller men's" length distributions is large enough that height tells you almost nothing useful about any one person you might apply it to. A large sample is what makes a weak effect detectable and statistically significant in the first place; it does not make the effect large enough to bet on.

Why the claim persists anyway

The height-length belief circulates as a tidy piece of folk reasoning - taller people are "bigger" generally, so surely everything scales together - and a genuine, if weak, statistical finding gives it just enough cover to survive being tested. People also tend to remember confirming cases and forget disconfirming ones, which is a general feature of how anecdote survives contact with population data, not something specific to this topic.

What this does not license

It does not license using height as a stand-in for a measurement, does not support a specific ratio or formula linking the two, and does not mean any individual case - taller and smaller, shorter and larger - is an exception to a rule, because there was never a strong enough rule for anyone to be an exception to.

Two related folk correlates get their own separate treatment because each has its own specific literature and its own specific caveats: the shoe-size claim, which Shah and Christopher (2002) tested in 104 men and found no significant correlation, and the finger-ratio literature, which is a small and different kind of research. Height is worth treating on its own rather than lumped in with either.

None of this is a measurement in the sense this site otherwise covers - correlating two body measurements across a population is a different exercise from taking either measurement carefully, and it is a different exercise again from a subjective judgement. A rating platform is not testing for a height correlation, and a photo submitted to an AI-based estimator, scored by a human judge, or rated alongside a gallery of others typically will not include a height figure at all, so the correlation described here has no bearing on what any of those assessments produce.