A psychology study performed in 1977 - The College-Bowl Study (Ross, Amabile, and Steinmetz, 1977) - demonstrated a fascinating piece of human psychology, which was that people would over- or underestimate people's intelligence relative to their own, attributing undemonstrated qualities to them, based upon role or situation. They called this an "attribution error".

In the study subjects were split into two groups: the quizmasters and the contestants. The quizmasters were tasked with coming up with some challenging-but-not-impossible questions, which they then presented to the contestants as pairs. Invariably the contestants did poorly, as the quizmasters naturally relied upon their own unique proficiencies for material: Whether they quizzed about classical music, famous art, the Altair 8800, or chemistry, their knowledge (and thus what they thought was "challenging but not impossible") differed greatly from the contestants, as should be expected among a diverse group of people.

Amazingly, when asked afterwards about the intelligence of each other, the contestants overwhelmingly estimated the intelligence and knowledge of the quizmasters as greater than their own. Similarly, in a follow-up study an observer was added, and again the observer believed the the quizmaster was more intelligent than the contestant.

Of course this was a completely ridiculous conclusion: Not only were they randomly assigned arbitrary roles, but general knowledge tests applied across the group showed no correlation between role and intelligence. Logically it seems probable that there is a tremendous bulk of knowledge that the contestant holds that the questioner does not, but because it wasn't demonstrated it was unaccounted for. Out of "sight", out of mind.

This basic human tendency is pervasive, and it goes both ways: Some of us underestimate ourselves because we don't have the (often artificial, superficial, or temporal) domain knowledge of others ("I just saw an episode of Numbers, and boy do I feel dumb now..."), while others under-estimate people who don't share their particular grab-bag of facts ("Boy that guy is an idiot! The guy didn't even know what OPML is!". Us nerds are particularly guilty of this, discounting the incredible array of knowledge and skills that people in other fields have, instead judging them on their knowledge of Linux distros or esoteric Windows shortcut keys. This is a tremendous vice).

Take advantage of this human trait! The next time you're having a big meeting, bone up on fringe facts and edge questions for your peers. Learn some irrelevant facts about an uncommon area - for instance beetles or metallurgy - and bring it up at every opportunity. I just took advantage of this myself, talking about the fairly obvious observations of a 28-year old study. Aren't I clever.