[I am not a psychologist, but here is my summary of how I think the real effect works...

Let's say we have a skill we can be assessed on a scale of 1-10. 1 is worst, 10 is best.

The popular perception of the Dunning Kruger effect is that people lower on the scale, say 2 or 3, rate themselves highly - say 7 or 8. And people who are higher - say a 7 or 8 - rate themselves more modestly - say a 5 or 6. i.e. people who are less skilled rate themselves as being better performers than people who are better.

The actual effect is a lot weaker, and a kind of reversion to the mean. If you are a 2 or 3 you may rate yourself a 4 or a 5. If you are a 7 or an 8 you may rate yourself as only a 6 or a 7. Less skilled people are well aware that they are less skilled, but may underestimate the degree of the effect.]