For example (but that's kind of a meta-point): If you select any skill (say driving) and ask a sample of people in which quantile they place themself, you are very unlikely to get a uniform distribution. (Like: "90% of people think themself above median drivers.")
But they're often right. Most people don't think themselves above-median drivers, they think themselves above-average (mean) drivers, i.e. if you take all the car accidents and divide by the number of drivers, they've been involved in fewer than that number. And they'd probably be right: the median number of car accidents is significantly lower than the mean, because a small number of bad drivers account for a large number of car accidents. Not everything is uniformly distributed - in fact, most things aren't.
I know that most things are strangely distributed. Even the normal distribution is far from uniform. So I was carefully phrasing my comments in terms of quantiles. Quantiles are evenly distributed by definition.
People think themself "better than the typical guy" i.e. above median.
