Personalized competency-based education through the dynamic selection of learning tasks in electronic learning environments
Printer-friendly versionPDF version
type of project:
PhD-project
project start:
1 Dec 2003
project end:
1 Dec 2007
date finished:
18 Jan 2008
Modern education emphasizes the need to flexibly personalize learning tasks to the needs and interests of individual learners. Rather than one curriculum for all learners, such approaches allow each learner to have her own curriculum. The sequence of learning tasks performed by the learners can be personalized either by a computer program, by the learners themselves, or by a combination of both, that is, the program and the learner may share control over task selection. The aim of the studies presented in this dissertation is to investigate under which conditions learner control over task selection is optimized.
Chapter 2 introduces a personalized task selection model with shared instructional control. Taking the 4C/ID-model (van Merriënboer, 1997) as a starting point, the first component comprises (a) task characteristics – level of complexity (i.e., from easy to difficult), embedded learner support (i.e., from full support to no support), and other task features (e.g., surface features) – and (b) learner characteristics – task performance and invested mental effort – documented in a learner portfolio. The second component refers to the personalization mechanism. Program controlled instruction includes task-selection rules used by an instructional agent to base its decisions on. Learner controlled instruction lets the learner select the learning tasks from a smaller or larger subset of - pre-selected - tasks. The third component includes the learning-task database with tasks with diverse levels of complexity, embedded support, and other task features. The model combines the strong points of program control and learner control into a model with shared control over task selection, which is expected to make learning more effective (i.e., higher transfer test performance), more efficient (i.e., higher transfer test performance combined with less invested mental effort), and more motivating. Chapter 2 also reports the results of a pilot study carried out with a twofold purpose: (a) to test whether adaptive task selection with shared control yields better results than adaptive task selection with full program control, and (b) to test a web application developed according to the model.
Taking the model described in Chapter 2 as a basis, Chapter 3 (domain: dietetics) describes a study in which 55 health science students participated in an experiment with a 2 x 2 factorial design with the factors adaptation (present or absent) and control over task selection (program control or shared control). It was predicted that adapting the selection of tasks to learner variables would lower cognitive load to an acceptable level, which enables the allocation of learners’ freed-up cognitive resources to learning. This hypothesis was confirmed. Specifically, the results show that adapting task difficulty and support to learners’ level of competence and perceived task load leads to more effective and efficient learning. However, training time in the non-adaptive conditions was unexpectedly low. This could be explained by a lack of learners’ willingness to invest more time in training, probably because in comparison with participants in the adaptive conditions they could not perceive the relationship between their performance and the levels of difficulty and support of the tasks they were working on. The second hypothesis stated that shared control over task selection – which provided learners a choice amongst three tasks varying in surface features – would enhance motivation. As expected shared control yielded higher task involvement (i.e., higher learning outcomes combined with more effort directly invested in learning). However, learners’ interest in training was not enhanced. This was partially explained by the relatively small amount of choice provided and the fact that learner control was limited to task selection.
Chapters 4 and 5 (domain: genetics) tested the effects of control over task selection based on surface task features (shared or learner control and program control). The study described in Chapter 4 is based on the assumption that the positive effects of learner control decrease when learners do not perceive the control given to them, make didactically unsound choices, or are overwhelmed by the amount of choice. Ninety-four students participated in a 2 x 2 factorial experiment with the factors control (program, shared) and variability of surface features (low, high). Two interaction effects reveal that learner control over surface features of selected tasks enhances transfer test performance and task involvement, but only when surface tasks features differ from the previous task. When learners were required to make a selection amongst highly similar tasks, transfer test performance and task involvement were hindered. The notion of perception of control was used to account for this effect, although no direct measures of perceived control were included in this study. Again, no differential effects were found on learners’ interest in training or the other motivational scales measured after the training. Measuring motivation during training, instead of after completing the whole training, could have been a more sensitive measure of motivation. Reported self-efficacy was found to be higher in the conditions with program control, which seems to indicate that when the sequence of tasks is extrinsically controlled, learners have more confidence that they will be able to perform the tasks.
Similarly, Chapter 5 describes a 2 x 2 factorial experimentcarried out with 72 participants, to study the effects of control over the selection of learning tasks that differ in surface task features (program control or learner control) and control over the selection of learning tasks that differ in structural task features (program control or learner control) of a series of completion tasks. Whereas in the study reported in Chapter 4 learners could select from a range of tasks with either similar surface features or dissimilar surface features as compared to each prior task, in the study reported in Chapter 5 learners could select from a range of tasks with both similar and dissimilar surface features and/or similar and dissimilar structural features. It was predicted that learner control over the selection of tasks with salient surface features would enable learners to select a varied set of personally relevant tasks which fosters learning and transfer. This hypothesis was supported for efficiency on the far transfer test. It was argued that learners who were given control over surface features probably constructed general cognitive schemas which enabled them to flexibly apply the learned solution procedure to solve unfamiliar inheritance tasks. In addition, learner control over the selection of tasks with non-salient structural features does not enable learners to select a varied set of personally relevant tasks and therefore was not expected to yield beneficial effects on learning. As expected, learner controlled selection of tasks with different structural features did notenhance learning. It was concluded that learners should be explicitly supported in recognizing structural task features.
Consequently, a final study (Chapter 6; domain: genetics) with 118 students investigates whether feedback on the structural features (operationalized as ‘knowledge of correct response’) would support learners to recognize those features. Feedback was found to enhance efficiency on the near transfer test, although the expected pattern found on the far transfer test failed to reach significance. Probably feedback supported learners in acquiring more or less automated cognitive schemas that allowed them to perform the steps of the near transfer test as ‘routines’. However, far transfer does not allow learners to merely apply a routine but deep understanding of the rationale behind the solution steps is crucial. In addition, in agreement with our predictions the provision of feedback made training in general as well as learner controlled selection of tasks with different structural features more motivating and relevant for learners. However, no support was found for the hypothesis stating that in combination with learner control over task selection, feedback would enhance efficiency. No differences amongst participants in the learner controlled conditions were found on efficiency. This seems to support the idea that less experienced learners are not able to make effective task selections on the basis of structural features and must thus be guided in how to achieve learning objectives.
Finally, Chapter 7 presents an overview and a general discussion of the results of the studies presented in Chapters 3 to 6. The general conclusions are: (a) it is advisable to adapt task difficulty and support to learners’ expertise, (b) learners should be provided with control over the selection of tasks that differ in surface features, provided that choices do not include surface features that are very similar to the surface features of the prior task, (c) learners do not benefit from control over the selection of tasks that differ in non-salient structural features unless they are supported in recognizing those features, and (d) feedback on structural features during task practice may supply this support. The provision of feedback made the given choices more motivating for the learners, but additional feedback strategies are needed to support learners in making an optimal selection of tasks with different structural features. In addition, Chapter 7 provides some explanations for unexpected results, followed by a discussion of some methodological issues concerning the efficiency measures and task involvement measures used in the study described in Chapter 3. The final Chapter closes with some considerations for future research and practical implications. A more fine-tuned task selection model might include other variables than mental effort and performance, such as time on task and motivation, for task selection purposes. Such a fine-tuned model should also increase the amount of learner control along with learners’ expertise. Future studies must include direct measures of perceived control and personal relevance, as well as immediate and delayed measures of transfer of learning. Implications concern the further examination of the effects of perceived control in educational settings, the importance of the role of salient surface features, and the potential effects of dynamic task selection especially for elderly people.
Product(s):
PhD Dissertation: Shared Control over Task Selection: Helping Students to Select their own Learning Tasks
On April 25th 2008 Gemma Corbalan defended her PhD Dissertation.
contact name:
Gemma Corbalan Perez
contact email:
gemma.corbalan@ou.nl
Financing:
This project was funded by NWO, Netherlands Organisation for Scientific Research
