Language on twitter: women in stem versus non-stem

Karen Truong*, Janet L. Fraser, and Hannah E. Bailey
John Chambers College of Business and Economics, West Virginia University, Morgantown, WV 26506

Presentation No.: 68

Assigned Category (Presentation Format): Engineering (Poster Presentations)

Student’s Major: Management Information Systems

Many language studies have been conducted between different genders; however, there is much to be investigated regarding language studies between different groups of the same gender. One approach to study within-gender differences was thought to be based broadly on career interest. Thus, we evaluated whether there was a difference between the way women of different fields (STEM versus Non-STEM) used language on social media. Because social media is one of today’s primary methods of communication, we extracted text from Twitter pertaining to the aforementioned groups. Approaching our problem, we utilized Natural Language Processing (NLP) to process tweets and make use of the following techniques for analysis: word clouds, word lists, and sentiment analysis. Preliminary results indicate that, while there is no significant variation in sentiments (positive or negative emotions) between the two groups, women in STEM appear to utilize work-related vocabulary more commonly than women in non-STEM fields. With further investigation, we intend to apply emoji and hashtag-based analysis and dig deeper into our data to understand what differences may exist between these groups.

Funding: West Virginia University

Program/mechanism supporting research/creative efforts: WVU's SURE program (Rita Rio & Michelle Richards-Babb)