Cyber Center

Computational Social Science Lecture Series

October 14, 2015

Automatic Statistical Discourse Analysis of Online Message Sequences

With the advent of digital media, new possibilities for data analysis have emerged. Content analysis, which relies on frequency and other types of counts of concepts or themes of discourse, can now be automated.  If valid meaning and language structure extraction mechanisms are used, high reliability results can be obtained. Dr. Ming MingChiu will introduce us to a new method of automatic content analysis applied to highly structured messages.

Online forums (synchronous and asynchronous) offer exciting opportunities to analyze how people influence one another by examining how sequences of messages affect the next message. However, researchers must address several analytic difficulties involving the data (missing values, nested structure [messages within topics], nonsequential messages), outcome variables (discrete, infrequent, multiple, adjacent message similarities), and explanatory variables (sequences, indirect mediation effects, false positives, robustness). I invented a method that addresses these difficulties (Statistical Discourse Analysis or SDA) and illustrate it on 1,330 asynchronous messages written by 17 students in an online forum linked to a high school algebra course for 13 weeks. Both individual characteristics and prior message sequences were linked to online message attributes. Boys gave more explanations but asked for fewer explanations.  Also, opinions were often followed by elaborations, which were often followed by explanations.

Monday, November 2
12:30-1:30 pm
Lawson 1142

Bookmark and Share

More Information