That is literally the dumbest fucking article I have read in a long time.
The Dunning-Kruger effect is a fascinating study about how personal expectations regress (just above) the mean, and what conditions allow a person to readjust their self-appraisals in light of peer observations. Interestingly, it does not need to apply to matters of objective fact, merely intersubjective.
It would seem that this would apply to a question like, "Which of these were good presidents?" Unfortunately for the author, there is no intersubjective consensus on these topics in the first place. You can tell, because the author immediately switches to random shit people believe about foreign policy. You can also tell because this is published in politico.
I'm absolutely floored that this author, otherwise a well-respected and well-educated figure, supposes that there is some intersubjective consensus on foreign policy on which we can establish a mean and give rankings to people. Just look at how nations have behaved throughout the world, and throughout history.
For some real science, instead of social science, consider the following.
Mathematically, if we assume that there is some objective fact of the matter (though we don't know it, and don't even know how to discover it, etc), and we assume that the mean probability that of selection over the voters need be just over 50%, then democracy has a better than 50% chance of selecting the "right" option, and the more voters the better the chances. Note: the probability is averaged over voters, it needn't mean that the population itself has a majority of voters which have a probability over 50% of selecting the "right" option (whatever that may be). This is an extension of Condorcet's Jury Theorem. This can be extended to more than one choice, with suitably-redefined probability assignments (plurality percentage, basically.)
(Reference: "Epistemic Democracy, List and Goodin, 2001)
David Dunning should be embarrassed.