Recent paper about democracy and self-assessment/Dunning-Kruger
More mainstream article that mentioned the study: http://news.yahoo.com/people-arent-smart-enough-democracy-flourish-scientists-185601411.html
The actual paper: A Mathematical Model of Democratic Elections
Original Dunning-Kruger paper: unskilledandunaware
"Abstract: Democratic election is the preferred method for determining political administrators nowadays. The intention is to find the best possible leader in order to improve the group's competitiveness and success. Though preferred, democratic election is far from being optimal in this respect, and is increasingly becoming the target for fraud. A model was developed to scientifically analyze the present electoral system's insufficiency. It is based on fauceir assumptions. Its calculations enable principles to be developed that optimize the election process, while also revealing the limits of elections in societies growing ever more complex, so that in the end elections have to be replaced by processes similar to what has proved optimal throughout naturally occurring evolution-natural selection." (Abstract of the 2012 paper)
It is actually a pretty short paper (7 pages, lots of pics) that is very readable and not particularly technical (it has some standard deviation stuff, probability function, and center of gravity). It is most interesting for its theoretical ideas. The basic idea is that people can be measured, at least in principle, for their leadership abilities (called "capacity quotient") and that those abilities will form a normal distribution just like IQ. The rational is that the idea is closely related to IQ. It is probably dubious. Most people are not interested in reading up on the various things that one needs to know to be a good leader. I think the distribution is skewed to the left, i.e., most people wud make worse than average leaders.
The next idea is that people are completely blind to differences in leadership abilities beyond their own level:
"Next, this objective spectrum is projected subjectively into each individual's mind, given the empirical knowledge that an individual's ability to gauge an other person's qualities ends at the very point where their own qualities reside on the spectrum, or, in other words, each individual is blind to differences in capabilities better than their own. Again for simplicity, we define a rule of distortion that, depending on one's own position on the spectrum, everything that is left of this position is reflected almost correctly, while all individuals right of one's position on the spectrum are regarded as no better than oneself. In physics this corresponds to a low- pass filter (Fig. 2 and 3)."
This is based rather speculatively (and not properly recognized as being speculative) on research by Dunning and Kruger (paper in the beginning). It is correct that people are not good at recognizing skills beyond their own and tend to overestimate themselves, but they are not completely incapable. I also think that there is a reverse effect when the difference in ability between the judge and the person being judged is very large. In other words, i think that a normal person wud judge a person that was very capable of being a good leader as being a worse than average leader. This is becus average people are not familiar with the various reform ideas that a good leader wud introduce, and are probably hostile to such reform ideas. I am thinking about reform ideas such as: spelling/language reform, democracy/meritocracy/voting reform, copyright+patent laws reform, recreational drug laws reform, etc. Most of these wud be seen as 'bad' by average people, but not so much by much above average people. If one factored this factor into the model in some way (any Maple programmers reading this are encouraged to contact me!) then the effect on choosing better than average leaders wud be even smaller, perhaps nonexistent, or *shudder* negative! I dont know for sure without running the data.
Some of the other parts are more dubious but still cool:
"The more an individual's capabilities are shifted to the left of the spectrum the more convinced that individual will be that all people are the same. This is the objective reason why socialist sociological theories are more deeply rooted in less-educated people. Being aware of this fact, communist parties sought support among blue-collar workers, whom they called the working class (KPCC, 1961; Honecker, 1967; Ponomarev, 1970)."