The shape of the distribution affects the value of the introvert/extrovert distinction pretty strongly, though. If it's strongly bimodal (so a random extrovert

The shape of the distribution affects the value of the introvert/extrovert distinction pretty strongly, though. If it's strongly bimodal (so a random extrovert is very different from a random introvert), then it says a lot, while if it's unimodal with a small variance (so a random extrovert is only a little more extroverted than a random introvert) it doesn't say much.