Loss of Self-Consciousness and autotelic personalities: a Machine Learning contribution
Abstract
Abstract
Background:
The EduFlow-2 measurement instrument (Heutte et al., 2021) approaches the flow psychological state (Csíkszentmihályi, 1975) in online, distance, education and training contexts via four constituting dimensions, linked to cognitive processes: Cognitive Absorption, Time Transformation, Loss of Self-Consciousness, and Autotelic Experience.
Goals:
Highlight the role of the EduFlow-2 D3 dimension Loss of Self-Consciousness (LoSC) as an indicator of autotelic personality in a MOOC.
Methods:
We employ a Variational Bayesian Gaussian Mixture Model (VBGMM) (Roberts, Husmeier, Rezek, & Penny, 1998) unsupervised Machine Learning (ML) algorithm to automatically discern typologies of MOOC participants (n = 1553) following their individual EduFlow-2 dimensional scoring. The ML model attempts clustering them into up to 22 distinct combinations while scoring a reduced number of typologies higher.
Results:
Among the seemingly infinite possibilities for any individual to score among these four flow dimensions, VBGMM has shown MOOC participants to group around only seven distinct typologies. Furthermore, these results unequivocally point to EduFlow-2 D3 dimension Loss of Self-Consciousness (LoSC) as a major indicator of personality that determines behavior in the persistence to will to learn in a lifelong manner. Results from applying an additional unsupervised k-means clustering further cemented the importance of the LoSC dimension.
Discussion:
A very much needed quantitative approach would include the educational and training environment's role on these typologies. Fueled by a biographical dimension of the individual and its impact on the LoSC flow dimension, such approach would prove itself as a determinant of the construction of autotelic personalities. Moreover, results contribute to establishing a flow threshold in MOOC contexts.