Multi-Dimensional Exploratory Factor Analysis in TED Talks

Artikel i vetenskaplig tidskrift, 2022

This article conducts Exploratory Factor Analysis (EFA) on a corpus of TED talks (2463 talks, across 427 topic tags) to create a new Multi-Dimensional model. The resultant model contained seven dimensions: i. ‘Spontaneous involved versus edited informational discourse’, ii. ‘Abstract informational versus narrative discourse’, iii. ‘Human-world oriented versus object-oriented discourse’, iv. ‘Subjective perspectives’, v. ‘Persuasive stance’, vi. ‘Expert elaboration’, and vii. ‘Change and inspiration’. When the model was compared to prior research, similarity with MD models based in academic texts was observed. However, some dimensions were found to be indicative of the unique nature of TED talks, such as expert elaboration and change and inspiration. When the EFA model was mapped onto the TED corpus’s subcorpora (defined by topic tags), individual disciplines were characterised in terms of the dimensions and some traditional academic groups were observed.