The article misses the point.

> The worst test-takers would also overestimate their performance the most because they are simply the furthest from getting a perfect score.

No. Earlier the article correctly states that all result quadrants more or less guess their raw performance with a similar degree of precision, but fail to guess their performance relative to each other with the same level of precision. These are two entirely unrelated measures that the article attempts to correlate in an invented relationship. The actual Dunning-Kruger results also indicate the highest scoring quadrant tend to underestimate their performance relative to their testing peers, but they are off by much less than the lowest score quadrant.

I am not saying Dunning-Kruger is necessarily correct or valid, and it has been challenged repeatedly. Still it is necessary to know what Dunning-Kruger actually concludes before challenging it.

Secondly, regardless of agreement on the conclusions provided by Dunning-Kruger it does identify something more important than relative bias. People are generally better at guessing from a known, or at least determinable, objective measure than they are at guessing from perceptive qualities that includes themselves as a data point. This is the basis of social intelligence. Social intelligence is the ability to perceive the communications of other people and to introspect. Both introspection and perception of others are required for the definition of social intelligence. Social intelligence has a weak correlation to academic intelligence as it is primarily a perceptive skill (conscientiousness) but it also requires the proper motivation (openness). People with low general intelligence are more likely to score lower in social intelligence and people with lower perceptive skills in any regard are less motivated/capable to measure things for any reason, and that motivation discrepancy is dramatically elevated a person places themselves a data point of measure.