The new literacies of online reading place new demands on student readers as their reading materials have expanded from textbooks to different representations and sources on the Internet. This requires digital literacy, which is not an innate skill. Rather, readers have to learn how to construct their own reading paths and integrate often contradictory information across web pages and presentation formats into a coherent whole. While this can lead to deeper learning, little is known about how these new demands may affect struggling readers. Therefore, a main aim of this thesis was to investigate possible differences between typical and struggling readers on how they manage the integration demands in multimedia learning. In the first paper we describe a quantitative study comparing students with and without dyslexia (N = 44) on a multiple source integration task. In this study we found that participants without dyslexia clearly outperformed participants with dyslexia on the integration task. Further, observed differences with respect to multiple source integration were largely due to working memory differences between the two groups. The second paper, which describes a multiple case study, investigates differences within the dyslexic group (N = 4) on the same tasks as in the first paper. This study found that, in addition to differences in reading speed and comprehension, participants’ different processing patterns could be related to outcomes on post reading knowledge and integration tasks. As the demands of digital literacy are assumed to draw heavily on working memory, the thesis also reviews how working memory has been conceptualized in contemporary research on multimedia learning that uses cognitive load theory as the major theoretical framework, and how subjective measures have been used in this research. The findings of the third paper showed that most of the reviewed studies did not include any conceptualization or clear definition of working memory, used only general subjective measures containing one or very few items, and did not report findings consistent with the hypothesized relationship between cognitive load and multimedia learning.
.