Developing Quantitative Methodologies for the Digital Humanities: A Case Study of 20th Century American Commentary on Russian Literature

View/ Open
Date
2014Author
Cai, Robert
Carr, Matthew Thomas
Elrafei, Adam
Goniprow, Alexander
Hamins-Puertolas, Adrian
Khural, Manpreet
Li, Andrew
Winter, Alexandra
Yanamandra, Soumya
Yang, Dan
Zhang, Kay
Advisor
Mallios, Peter Lancelot
Metadata
Show full item recordAbstract
Using scientific methods in the humanities is at the forefront of objective literary analysis.
However, processing big data is particularly complex when the subject matter is qualitative
rather than numerical. Large volumes of text require specialized tools to produce quantifiable
data from ideas and sentiments. Our team researched the extent to which tools such as Weka and
MALLET can test hypotheses about qualitative information. We examined the claim that literary
commentary exists within political environments and used US periodical articles concerning
Russian literature in the early twentieth century as a case study. These tools generated useful
quantitative data that allowed us to run stepwise binary logistic regressions. These statistical tests
allowed for time series experiments using sea change and emergency models of history, as well
as classification experiments with regard to author characteristics, social issues, and sentiment
expressed. Both types of experiments supported our claim with varying degrees, but more
importantly served as a definitive demonstration that digitally enhanced quantitative forms of
analysis can apply to qualitative data. Our findings set the foundation for further experiments in
the emerging field of digital humanities.