Doctoral thesis

Learner models in online personalized educational experiences : an infrastructure and some experim


179 p.

Thèse de doctorat: Università della Svizzera italiana, 2014 (jury note: Magna cum laude)

English Technologies are changing the world around us, and education is not immune from its influence: the field of teaching and learning supported by the use of Information and Communication Technologies (ICTs), also known as Technology Enhanced Learning (TEL), has witnessed a huge expansion in recent years. This wide adoption happened thanks to the massive diffusion of broadband connections and to the pervasive needs for education, highly connected to the evolution in sciences and technologies. Therefore, it has pushed up the usage of online education (distance and blended methodologies for educational experiences) to, even in lately years, unexpected rates. Alongside with the well known potentialities, digital-based educational tools come with a number of downsides, such as possible disengagement on the part of the learner, absence of the social pressures that normally exist in a classroom environment, difficulty or even inability from the learners to self-regulate and, last but not least, depletion of the stimulus to actively participate and cooperate with lectures and peers. These difficulties impact the teaching process and the outcomes of the educational experience (i.e. learning process), being a serious limit and questioning the broader applicability of TEL solutions. To overcome these issues, there is a need of tools to support the learning process. In the literature, one of the known approach to improve the situation is to rely on a user profile, that collects data during the use of the eLearning platforms or tool. The created profile can be used to adapt the behaviour and the contents proposed to the learner. On top of this model, some researches stressed the positive effects stimulated by the disclosure of the model itself for inspection purposes by the learner. This disclosed model is known as Open Learner Model (OLM). The idea of opening learners' profile and eventually integrate them with external on-line resources is not new and it has the ultimate goal of creating global and long-run indicators of the learner's profile. Also the representation aspect of the learner model plays a role, moving from the more traditional approach based on the textual and analytic/extensive representation to the graphical indicators that are able to summarise and to present one or more of the model characteristics in a way that is considered more effective and natural for the user consumption. Relying on the same learner models, and stressing the different aggregation and representation capabilities, it is possible to either support self-reflection of the learner or to foster the tutoring process to allow proper supervision by the tutor/teacher. Both the objectives can be reached through the graphical representation of the relevant information, presented in different ways. Furthermore, with such an open approach for the learner model, the concepts of personalisation and adaptation acquire a central role in the TEL experience, overcoming the previous limits related to the impossibility to observe and explain to the learner the reasons for such an intervention from the tool itself. As a consequence, the introduction of different tools, platforms, widgets and devices in the learning process, together with the adaptation process based on the learner profiles, can create a personal space for a potential fruitful usage of the rich and widespread amount of resources available to the learner. This work aimed at analysing the way a learner model could be represented in visual presentation to the system users, exploring the effects and performances for learners and teachers. Subsequently, it concentrated in investigating how the adoption of adaptive and social visualisations of OLM could affect the student experience within a TEL context. The motivation was twofold. On one side was to show that the approach of mixing data from heterogeneous and not already related data sources could have a meaningful didactic interpretations, whether on the other one was to measure the perceived impact of the introduction on online experiences of the adaptivity (and of social aspects) in the graphical visualisations produced by such a tool. In order to achieve these objectives, the present work analysed and addressed them through an approach that merged user data in learning platforms, implementing a learner profile. This was accomplished by means of the creation of a tool, named GVIS, to elaborate on the collected user actions in platforms enabling remote teaching. A number of test cases were performed and analysed, adopting the developed tool as the provider to extract, to aggregate and to represent the data for the learners' model. The GVIS tool impact was then estimated with self- evaluation questionnaires, with the analysis of log files and with knowledge quiz results. Dimensions such as the perceived usefulness, the impact on motivation and commitment, the cognitive overload generated, and the impact of social data disclosure were taken into account. The main result found by the application of the developed tool in TEL experiences was to have an impact on the behaviour of online learners when used to provide them with indicators around their activities, especially when enhanced with social capabilities. The effects appear to be amplifies in those cases where the widget usage is as simplified as possible. From the learner side, the results suggested that the learners seem to appreciate the tool and recognise its value. For them the introduction as part of the online learning experience could act as a positive pressure factor, enhanced by the peer comparison functionality. This functionality could also be used to reinforce the student engagement and positive commitment to the educational experience, by transmitting a sense of community and stimulating healthy competition between learners. From the teacher/tutor side, they seemed to be better supported by the presentation of compact, intuitive and just-in-time information (i.e. actions that have an educational interpretation or impact) about the monitored user or group. This gave them a clearer picture of how the class is currently performing and enabled them to address performance issues by adapting the resources and the teaching (and learning) approach accordingly. Although a drawback was identified regarding the cognitive overload, the data collected showed that users generally considered this kind of support useful. There is also indications that further analyses can be interesting to explore the effects introduced in the teaching practices by the availability and usage of such a tool.
  • English
Education, teaching
License undefined
Persistent URL

Document views: 114 File downloads:
  • 2014COM004.pdf: 81