Article Preview
Top1. Introduction
Individual learner differences using Web-Based Instruction (WBI) programs have gained increasing attention in the design and development of hypermedia learning systems. There have been numerous research studies on the effect of hypermedia systems on learners’ performance using WBI programs (Alhajri, Counsell, & Liu, 2013a; Alhajri, Counsell, & Liu, 2013b; Calcaterra, Antonietti, & Underwood, 2005; Chen & Liu, 2008; Chen, Fan, & Macredie, 2006; Farrell & Moore, 2001; Mitchell, Chen, & Macredie, 2005a; Reed & Oughton, 1998; Workman, 2004). When learners use hypermedia systems they may make progress at their own pace and develop their own learning paths (Alhajri, Counsell, & Liu, 2013b; Chen, Fan, & Macredie, 2006). Users should have a flexible interface which accommodates their needs, be easily able to identify relevant content and navigational support, as well as move around freely (Calcaterra, Antonietti, & Underwood, 2005; Mitchell, Chen, & Macredie, 2005a; Alhajri, Counsell, & Liu, 2013c; Alhajri, Counsell, & Liu, 2013d). In a number of recent studies, the focus has been on understanding learners’ performance and behaviour using these systems and investigation of a learner’s performance of these differences individually.
In this article, we investigate the performance of student participants based on their individual differences using a specially designed WBI program (Alhajri, Counsell, & Liu, 2013c) which accommodate their learning needs. This design help learner making progress at their own pace and develop their own learning paths (Alhajri, Counsell, & Liu, 2013b; Chen, Fan, & Macredie, 2006). We investigated the learning performance of learners by combining three individual differences (henceforward called a ‘Multi-ID’ approach) after acquiring knowledge. Additionally, we used three attributes (‘gain’ score which is the difference in participant score obtained post-test minus the score obtained pre-test, number of pages visited and time spent using the system) as a vehicle for measuring learner performance. Data mining techniques, using k-means clustering method, is also used to uncover the relationships between measurement attributes which provides a deeper and better understanding of performance level for the combined individual differences (Multi-ID). The data then was filtered to validate the data mining findings in order to investigate the sensitivity of the results for further exploration of understanding the relationship between such attributes. This deep investigation of the results was done by providing data threshold of g-score greater than zero vs. less than or equal to zero. We then re-run the clustering method in order to evaluate our findings and to understand what could affect the original conclusions of learners’ performance. ANOVA test is then conducted to understand how the performance level was affected after moving the threshold by comparing the two clustering tests.
The remainder of the article is structured as follows. First, we present previous work and an appraisal of studies directly relevant to our findings. We then describe the methodology used to conduct the three investigations. The findings of our investigations are then discussed. Finally, conclusions and possibilities for future work are identified.