A model for using learners' online behaviour to inform differentiated instructional design in MOODLE
- Authors: Leppan, Ronald George
- Date: 2019
- Subjects: Cyberspace -- Psychological aspects , Web applications in libraries Moodle Special education -- Computer programs Computer-assisted instruction -- Computer programs Open source software
- Language: English
- Type: Thesis , Doctoral , DPhil
- Identifier: http://hdl.handle.net/10948/40393 , vital:36155
- Description: This thesis proposes a learning analytics-based process model, derived from a web analytics process, which aims to build a learner profile of attributes from Moodle log files that can be used for differentiated instructional design in Moodle. Commercial websites are rife with examples of personalisation based on web analytics, while the personalisation of online learning has not yet gained such widespread adoption. Several Instructional Design Models recommend that, in addition to taking prior knowledge and learning outcomes into account, instruction should also be informed by learner attributes. Learning design choices should be made based on unique learner attributes that influence their learning processes. Learner attributes are generally derived from well-known learning styles and associated learning style questionnaires. However, there are some criticisms of learning style theories and the use of questionnaires to create a learner profile. Attributes that can be inferred from learners’ online behaviour could provide a more dynamic learner profile. Education institutions are increasingly using Learning Management Systems, such as Moodle, to deliver and manage online learning. Moodle is not designed to create a learner profile or provide differentiated instruction. However, the abundant data generated by learners accessing course material presented in Moodle provides an opportunity for educators to build such a dynamic learner profile. Individual learner profiles can be used by educators who desire to tailor instruction to the needs of their learners. The proposed model was developed and evaluated using an iterative design focused approach that incorporates characteristics of a web analytics process, instructional design models, Learning Management Systems, educational data mining and adaptive education technologies. At each iteration, the model was evaluated using a technical risk and efficacy strategy. This strategy proposes a formative evaluation in an artificial setting. Evaluation criteria used include relevance, consistency, practicality and utility. The contributions of this thesis address the lack of prescriptive guidance on how to analyse learner online behaviours in order to differentiate learning design in Moodle. The theoretical contribution is a model for a dynamic data-driven approach to profile building and a phased differentiated learning design in a Learning Management System. The practical contribution is an evaluation of the expected practicality and utility of learner modelling from Moodle log files and the provision of tailored instruction using standard Moodle tools. The proposed model recommends that educators should define goals, develop Key Performance Indicators (KPI) to measure goal attainment, collect and analyse suitable metrics towards KPIs, test optional alternative hypotheses and implement actionable insights. To enable differentiated instruction, two phases are necessary: learner modelling and differentiated learning design. Both phases rely on the selection of suitable attributes which influence learning processes, and which can be dynamically inferred from online behaviours. In differentiated learning design, the selection/creation and sequencing of Learning Objects are influenced by the learner attributes. In learner modelling, the data sources and data analysis techniques should enable the discovery of the learner attributes that was catered for in the learning design. Educators who follow the steps described in the proposed model will be capable of building a learner profile from Moodle log files that can be used for differentiated instruction based on any learning style theory.
- Full Text:
- Date Issued: 2019
The impact of an in-depth code comprehension tool in an introductory programming module
- Authors: Leppan, Ronald George
- Date: 2008
- Subjects: Algorithms , Computer programming , Algorithms -- Study and teaching -- South Africa , Visual perception
- Language: English
- Type: Thesis , Masters , MSc
- Identifier: vital:10471 , http://hdl.handle.net/10948/847 , Algorithms , Computer programming , Algorithms -- Study and teaching -- South Africa , Visual perception
- Description: Reading and understanding algorithms is not an easy task and often neglected by educators in an introductory programming course. One proposed solution to this problem is the incorporation of a technological support tool to aid program comprehension in introductory programming. Many researchers advocate the identification of beacons and the use of chunking as support for code comprehension. Beacon recognition and chunking can also be used as support in the teaching model of introductory programming. Educators use a variety of different support tools to facilitate program comprehension in introductory programming. Review of a variety of support tools fails to deliver an existing tool to support a teaching model that incorporates chunking and the identification of beacons. The experimental support tool in this dissertation (BeReT) is primarily designed to encourage a student to correctly identify beacons within provided program extracts. BeReT can also be used to allow students to group together related statements and to learn about plans implemented in any semantically and syntactically correct algorithm uploaded by an instructor. While these requirements are evident in the design and implementation of BeReT, data is required to measure the effect BeReT has on the indepth comprehension of introductory programming algorithms. A between-groups experiment is described which compares the program comprehension of students that used BeReT to study various introductory algorithms, with students that relied solely on traditional lecturing materials. The use of an eye tracker was incorporated into the empirical study to visualise the results of controlled experiments. The results indicate that a technological support tool like BeReT can have a significantly positive effect on student comprehension of algorithms traditionally taught in introductory programming. This research provides educators with an alternative way for the incorporation of in-depth code comprehension skills in introductory programming.
- Full Text:
- Date Issued: 2008