Adaptivity in complex language tasks in game-based learning environments

This PhD is carried out within the framework of a Baekeland mandate in collaboration with Linguineo BVBA. Linguineo develops game-based applications for language learning and discovered users’ needs for adaptivity within complex language tasks in the applications. By 'complex language tasks' we mean task-based language activities. Complex language tasks go beyond just exercises and contain different language tasks. The goal is to use language in a meaningful way (Ellis 2003; East 2017). An example of such a task is going to the store and buying products. In order to carry out this task, you may have to write down a list of products you need, read the product description in the store, talk to the shop assistant, etc. In short, you are using language in various ways (writing, listening, reading, talking), yet always in a meaningful way. The goals is to effectively do some grocery shopping, not to learn some verb forms. Different learners learn in different ways, they have a different starting level and make different mistakes. Consequently, language learners need an adaptive learning environment that caters for their actual needs. Although there is an increasing interest in adaptivity in computer- and game-based learning in the literature (Basu, Biswas & Kinnebrew 2017; Mavroudi, Giannakos & Krogstie 2017; Vanbecelaere et al. 2019; Afini Normadhi et al. 2019), adaptivity in complex language tasks remains underexamined. The aims of this PhD are (1) creating a framework for adaptivity in complex language tasks and (2) incorporating and evaluating adaptivity in Linguineo's language learning environments. In order to investigate this, 4 work packages were drawn up. In work package 1, we will determine the framework for (the effectiveness of) adaptivity. In what respect do we want to make the learning environment adaptive? What are the exact users’ needs? What data do we need to create a learner model? Work package 2 and work package 3 deal with where we want to make the learning environment adaptive. These two work packages will be carried out analogously. In work package 2, we investigate adaptivity between the tasks (task selection and task sequencing). In work package 3, we examine adaptivity within the task. That is, we see three possibilities here, (1) the hints, (2) the complexity of expressions of the chatbot and (3) the way of interaction with the chatbot (writing answer in full, record answer, multiple choice). Based on the output of work package 1, we will select the most relevant option for the study. The implemented adaptivity will also be evaluated in work packages 2 and 3. In work package 4 we will investigate how we can visualize information about the learning process of the learner for the teacher. In doing so, we want to examine how technology and the teacher can be attuned to each other as much as possible in order to support the learner optimally in the learning process. To investigate this, 4 work packages were drawn up. In work package 1, we will determine the framework for (the effectiveness of) adaptivity. In what respect do we want to make the learning environment adaptive? What are the exact users’ needs? What data do we need to create a learner model? Work package 2 and work package 3 deal with where we want to make the learning environment adaptive. These two work packages will be carried out analogously. In work package 2, we investigate adaptivity between the tasks (task selection and task sequencing). In work package 3, we examine adaptivity within the task. That is, we see three possibilities here, (1) the hints, (2) the complexity of expressions of the chatbot and (3) the way of interaction with the chatbot (writing answer in full, record answer, multiple choice). Based on the output of work package 1, we will select the most relevant option for the study. The implemented adaptivity will also be evaluated in work packages 2 and 3. In work package 4 we will investigate how we can visualize information about the learning process of the learner for the teacher. In doing so, we want to examine how technology and the teacher can be attuned to each other as much as possible in order to support the learner optimally in the learning process.