Wednesday, November 6, 2013

USA! USA!

The perfect rating of "Who's Bigger"on Amazon has just been lost by the following "book review":

"24 out of the 100 most important historical figures are US Americans (and colonial Americans)? That seems somewhat unlikely, even assuming that the US is the greatest thing since sliced bread."

I suspect we will be fielding similar complaints from other readers, so let's consider this issue in a rational way.  What faction of the most historically significant figures should be from the United States?

First, we will confess to some Anglocentric bias in our results, because our analysis is based almost completely on the English language Wikipedia.   Figures from the English-speaking world are much more likely to be recognized here than comparable figures from other cultures.  An example from our book is Chinese painter Li Keran [663492], whose paintings have been sold for prices similar to Cy Twombly [37124].  Running our methodology on the Wikipedias of other languages would yield different results, and is a project in progress.

But back to the question: what fraction of historically significant figures should be from the United States?  One answer might be 5%, the fraction of the world's population that lives in the US.  Because of explosive population growth, most of the people in human history have lived quite recently, making this a reasonable proxy for population over time.

But do the people around the globe have equal chance to lead historically-significant lives? Unfortunately, the answer is no: access to education, resources, and opportunities have been concentrated in wealthier nations.   The United States slice of the world's economic pie has long been disproportionally large.  Indeed, the first reference I found on the web authoritatively states that the United States represented 25.3% of the world's GDP in 2012, essentially identical to the fraction which so offended this reviewer.

The United States has been the world's dominant economic, cultural, political, and military power for the past hundred years.  Roughly half of all the figures in appearing in Wikipedia had lifespans which intersect this period.  The prominence of Americans here seems quite explainable here for reasons beyond simple Anglocentric bias.

Indeed, our analysis unintentionally biases the rankings against the United States in certain ways.  One important factor is a reputation decay model, which awards greater prominence to figures from earlier times.   There are essentially no recorded American historical figures from before Columbus, or past 500 years ago.  Compare this to European, Asian, and Middle-Eastern civilizations which stretch back thousands of years.  The leading figures of these cultures are highly regarded by our algorithms.

What is the right fraction of Americans among the 100 most significant figures in history?  Our algorithms say 24%, and that seems generally plausible to me.  If you really think this is wildly out of range, OK, but tell me what the right answer is and why.   I'd like to know.