Скачать или смотреть Top-down and bottom-up text analysis methods

Presented at Marymount University Innovations, 2022.
A great deal about a person’s mental world can be revealed by the language that they use. Perceptions, decision making processes, emotional states, and many more can be accurately measured through text analysis methods. Now new automated methods are making text analysis less subjective, less time consuming, and able to handle more and larger samples. My own research has focused on using quantitative text analysis to extract meaningful variables for emotional states, cognitive processing, and analytic style from employees’ open ended survey responses and archival text samples to better understand work burnout and feelings of fit at work. In this talk I will share two of the most common and simplest text analysis methods: bottom-up analysis (more inductive examination of sentiment or topics in large numbers of text) and top-down analysis (more deductive examination of the psychological characteristics of texts based on term frequencies). This talk will focus on equipping researchers to use these powerful methods to examine their data in new ways and formulate new research questions.