Up at Night

The Averaged American, historical interpretation, and going digital November 29, 2010

The 1947 Jimmy Stewart film, Magic Town, is the story of a pollster who discovers a shortcut to obtaining the perfect statistical representation of the American people. Stewart finds that Grandview is the ideal town, average in all respects. Once his secret is out, thousands of people migrate to Grandview in hopes of becoming part of the average, normal, and typical American community.

For my next history project, I would like to produce a work of digital scholarship that tests the thesis of historian Sarah Igo.  In her 2007 book, The Average American:  Surveys, Citizens, and the Making of a Mass Public, she argues that Robert and Helen Lynd’s Middletown studies (community studies), the work of pollsters like George Gallup and Elmo Roper (public opinion polling), and the publication of the Kinsey Reports (sexual behavior research) created a mass public by providing Americans with a way to look at themselves.  In doing so, these helped create a new sense of national identity (to hear a 2007 NPR interview with Igo, click here).  In her study, Igo consulted letters to pollsters, newspapers, popular magazines, and scholarly monographs to provide support for her historical argument.

How do I plan to do this project?  Instead of proving or disproving her thesis based on readings, I will do this with digital methods.  I would like to create a series of charts, graphs, visualizations, and other text-mining products to create a historical paper.

I decided to start with text mining popular literature.  The TIME Magazine Corpus, designed by Mark Davies of Brigham Young University, is a useful resource for those wanting to analyze the evolution of linguistic terms in twentieth century America.  What is great about this site is the user’s ability to get an immediate sense of the how much a cultural term was used.  Igo argues that between the 1920s and 1950s, Americans became more conscious of what was “representative” and the “majority.”  With this corpus, you can test this.  I can easily search the use of terms such as “public opinion” and “the people” and the see the frequency (by decade) of these terms in TIME’s editorial content.

George Gallup, profiled by TIME in 1948.

Ideally, I would like to search through other corpora to get a better idea of how much particular terms were embraced by American culture; however, it is difficult to find some all text, well-organized corpora of newspapers.  There are many online databases for both newspapers and magazines; however, it appears that most of these databases are not in the desired format.   Some even charge, such as the Washington Post, which has an archive collection that dates back to 1887.  The New York Times, which would have been an excellent source, does have a corpus that is in an organized format.  However, this dates back to only 1987 and is not open access.

I am just now really getting into this project, but I would like to create a visualization of words used during this time period and/or years around the three decades of Igo’s study.  Also, one of Igo’s arguments were that salient consensus of what was “average” had a white middle class bias.  It would be great to be able to create a visual to test this; however, this will require more thinking on my part.

UPDATE: Since my original post (above), I have expanded my text mining to include some of Google Book’s entire runs of magazines.  They have all LIFE magazine issues since was purchased from Time Inc in 1945.  Also, I will include magazines that are not so “white targeted,” such as Ebony, Jet, and Black World (originally called Negro Digest).  While Google Books does not house the first years of the later publications, what they do provide should still be adequate in providing data for a comparison visual that shows whether terms like “average” and “majority” were ingrained in American black culture.