Intelligent assessment and learner personalisation in adaptive educational systems for STEM education

Abstract

The growth and success of e-learning is largely dictated by the technology and pedagogical approaches which support it. Many early e-learning systems were developed with the ‘one-size-fits-all’ paradigm, reflecting capabilities at the time. Differences such as students’ ability and learning style were disregarded, and the same learning materials were supplied to all students [1]. Adaptive Educational Systems (AES) have contributed towards the delivery of improved e-learning, to some degree, by providing tailored personalised theoretical content which has been repurposed specifically to meet the needs of individual students. However, whilst this is an improvement on the ‘one-size-fits-all’ paradigm, the needs of the student are often determined explicitly or statically through fixed assessments of learning styles or preferences. Research states that it can be counter-productive to lock the student into a fixed learning style and that it is more fruitful to identify the learning styles dynamically [2], [3], [4]. Moreover the content delivered through AES is predominantly theoretical. It is important that complex, practical subjects associated with Science, Technology, and Engineering and Maths (STEM) education are accommodated to meet the increasing demands from industrial professional bodies and government departments of education.

The work presented in this thesis focuses on the design, development, implementation and evaluation of an AES named IPAL. Drawing on fields of education, psychology and computer science IPAL is introduced as a supplementary and complementary course tool in higher education, which facilitates teaching and learning in STEM education. This thesis addresses associated challenges with AES and specifically investigates dynamic user modelling by offering a two step approach in determining learning styles. An initial primary learning style is first established explicitly, addressing the cold start issue which occurs when the information required to determine a student’s learning style is not available [5], [6]. A secondary learning style is then determined implicitly from student engagement, using educational data mining. In addition, the thesis extends the personalisation of theoretical content by supporting complex, practical STEM education, by extending the two step learning style approach within IPAL to an integrated Game Based Learning (GBL) environment. GBL aims to present, complex concepts or theories in new and innovative ways, providing highly engaging learning experiences through, for example 3D and avatar based interactions. This approach explores learning style adaption in GBL for STEM education.

A mixed method research approach was employed whereby quantitative and qualitative data was obtained from over 100 participants using STEM content to determine the effectiveness of IPAL. The results in this thesis demonstrate that IPAL is an effective AES in STEM education which improves the learning outcomes of students.

Date of Award	Aug 2015
Original language	English
Supervisor	Jim Harkin (Supervisor), Shane Wilson (Supervisor) & Michael Callaghan (Supervisor)