Responsibility is usually held by a board of directors or trustees that appoints its own successors but that is not universal.
Often the managers of a non-profit will balloon their salaries or give bonuses to the board of directors if there is money left over. This is a form of corruption but remains common.
A 501(c)3 corporation in the USA is a type of non-profit that receives special privileges, the most obvious of which is that contributions can be written off the taxes of the donor. To be a 501(c)3, a non-profit corporation has to be former for one of a specific set of public charitable purposes and meet special reporting requirements. The principal purposes that qualify are churches and schools but also include aid to the poor and others.
IRS publication 557 details the rules for 501(c)3 and other tax-exempt non-profit organizations.
I guess this is why the 2012 ML class is delayed; I'm wondering if it will still happen :(
This article seems to hint that the Stanford administration isn't all that gung ho about the whole thing. http://www.npr.org/blogs/alltechconsidered/2012/01/23/145645...
What would be a far more interesting research topic would be to measure negative reaction to any notion that some people can work better alone and any correlations those sentiments might entail.
Your proposed research topic would be a lot more interesting without the inherent bias you introduce by measuring only the correlations of negative reaction and ignoring any correlations to positive reaction.
I don't know what type of evidence you're looking for, and I don't know what it has to do with co-founders, but it seems pretty obvious to me that one person couldn't have designed every piece of Apple hardware or software or every version of Windows. That's why we invented collaboration - it gives extremely productive people a multiplier.