The impact of an automated learning component against a traditional lecturing environment

• Published in: Journal of computer assisted learning 2017-12, Vol.33 (6), p.597-605
• Main Authors: Maycock, K.W., Keating, J.G.
• Format: Article
• Language: English
• Subjects: Adaptive control; Algorithms; Automation; CAI; Cognitive Ability; Comparative Analysis; Computer Assisted Instruction; Content Analysis; Control Groups; Course Content; Course Descriptions; Discovery Learning; Distance learning; Effect Size; Environment; Experimental Groups; Genetic algorithms; Information transfer; Instructional Effectiveness; Learning; Learning Management Systems; Lecture Method; Management Systems; Metadata; MOOCs; Personal communication; Profiles; Scoring; Statistical Analysis; Statistical significance; Teaching Methods; Undergraduate Students
• ISSN: 0266-4909; 1365-2729
• DOI: 10.1111/jcal.12203

This experimental study investigates the effect on the examination performance of a cohort of first‐year undergraduate learners undertaking a Unified Modelling Language (UML) course using an adaptive learning system against a control group of learners undertaking the same UML course through a traditional lecturing environment. The adaptive learning system uses two components for the creation of suitable content for individual learners: a content analyser that automatically generates metadata describing cognitive resources within instructional content and a selection model that utilizes a genetic algorithm to select and construct a course suited to the cognitive ability and pedagogic preference of an individual learner, defined by a digital profile. Using the Kruskal–Wallis H test, it was determined that there was a statistically significant difference between the control group of learners and the learners that participated in the UML course using the adaptive learning system following an examination once the UML course concluded, with p = 0.005, scoring on average 15.71% higher using the adaptive system. However, this observed statistically significant difference observed a small effect size of 20%. Lay Description Current environment Digital‐driven pedagogy required leverages emerging technology tools. Learning management systems are reinforcing the information transfer didactic style of delivering content. Automated systems are domain specific and hindered by the creators' personal profile. The completion rate for 90% of massively open online courses is less than 14%. How this paper adds to the field It describes a suitable digital personal profile focused on digital environmental characteristics. It presents an automated approach that is independent of domain content. The automated system is reprogrammable to generate content to suit different traits. A content analyser is discussed that automatically creates metadata targeting the digital profile. Implications of study findings for practitioners Students scored significantly higher when using the adaptive system. This approach should be used to assist individual exploratory learning. Large amounts of educational content are required for the evolutionary process. A degrading factor should be included within the evolutionary process to ensure a course is created.