Text-mining forma mentis networks reconstruct public perception of the STEM gender gap in social media

Area of Science:

• Cognitive Science
• Computational Linguistics
• Network Science

Background:

• Mindset reconstruction involves understanding how individuals perceive knowledge.
• The mental lexicon reflects cognitive processes and language use.
• Textual forma mentis networks (TFMN) offer a novel method for analyzing mindset structures from text.

Purpose of the Study:

• To introduce and validate Textual forma mentis networks (TFMN) for extracting mindset structures from textual data.
• To apply TFMNs to analyze online perceptions of the gender gap in science using social media data.
• To identify prevalent semantic frames and sentiment associated with gender-related terms in online discourse.

Main Methods:

• Developed and validated TFMNs using network science, psycholinguistics, and Big Data approaches.
• Analyzed 10,000 tweets focusing on the beginning of posts to capture initial opinions.
• Employed unsupervised learning to identify relevant concepts and semantic networks within the textual data.

Main Results:

• TFMNs successfully identified key concepts in benchmark texts without supervision.
• Analysis of tweets revealed positive perceptions of "gender" and "gap," associated with celebrating female scientists and hoping for resolution.
• The term "woman" evoked discussions of sexual harassment and stereotype threat, while "man" indicated awareness of male superiority myths.

Conclusions:

• Twitter discourse, particularly at the start of conversations, largely presents a stereotype-free and positive view of gender disparity in science.
• Monitoring for discriminatory language should extend beyond initial posts to include user replies.
• TFMNs provide a valuable tool for monitoring collective online mindsets and informing policy-making.