E-Learning Courses Evaluation on the Basis of Trainees' Feedback on Open Questions Text Analysis

Abstract

Life-long learning is a necessity associated with the requirements of the fourth industrial revolution. Although distance online education played a major role in the evolution of the modern education system, this share grew dramatically because of the COVID-19 pandemic outbreak and the social distancing measures that were imposed. However, the quick and extensive adoption of online learning tools also highlighted the multidimensional weaknesses of online education and the needs that arise when considering such practices. To this end, the ease of collecting digital data, as well as the overall evolution of data analytics, enables researchers, and by extension educators, to systematically evaluate the pros and cons of such systems. For instance, advanced data mining methods can be used to find potential areas of concern or to confirm elements of excellence. In this work, we used text analysis methods on data that have emerged from participants' feedback in online lifelong learning programmes for professional development. We analysed 1890 Greek text-based answers of participants to open evaluation questions using standard text analysis processes. We finally produced 7-gram tokens from the words in the texts, from which we constructed meaningful sentences and characterized them as positive or negative. We introduced a new metric, called acceptance grade, to quantitatively evaluate them as far as their positive or negative content for the online courses is concerned. We finally based our evaluation on the top 10 sentences of each category (positive, negative). Validation of the results via two external experts and data triangulation showed an accuracy of 80%.