| dc.description.abstract |
Learning with visual representations has been the focus of research for the past decades. A robust finding is the multimedia effect (Butcher, 2014), that is, learning with a combination of text and pictures is more beneficial than learning with text alone. Three cognitive processes are assumed to be important for multimedia learning: selecting and organizing information from text and pictures, as well as integrating this information in order to build a coherent mental model of the learning content (Mayer, 2009, 2014a). Although integration is crucial for effective multimedia learning, learners show only few attempts to integrate (Hegarty & Just, 1993; Mason et al., 2013, 2015; Scheiter & Eitel, 2015). One way to foster integration of verbal and pictorial information is to instruct learners to self-generate the visual representations. This strategy is described as learner-generated drawing (van Meter & Firetto, 2013, van Meter & Garner, 2013).
Learner-generated drawing requires learners to construct external pictorial representations that include the key concepts and their relations while learning from verbal instruction (Leutner & Schmeck, 2014). Drawing has been shown to promote learning in terms of higher-order knowledge (for overviews see Ainsworth et al., 2011; van Meter & Firetto, 2013; van Meter & Garner, 2005). Benefits of drawing seem to depend on the availability and type of support during drawing construction (e.g., van Meter, 2001, van Meter et al., 2006, Schwamborn et al., 2010, Schmeck et al., 2014). Moreover, the quality of drawings constructed during drawing (i.e., the number of key concepts correctly incorporated into the drawings) is positively associated with learning outcomes; a finding that can be described as the prognostic drawing effect (Schwamborn et al., 2010). The main goal of the present thesis was to investigate which of three factors – generation, visualization, and externalization – mainly contributes to benefits of drawing. Second, it was examined how several boundary conditions would affect benefits of drawing – namely, the type of posttest, the quality of drawings constructed during learning, test delay as well as type of instructional support. Third, it was of interest how a drawing strategy instruction influences learners’ perceived difficulty of learning.
The first study investigated the influence of generation and visualization. Drawing was compared with a multimedia condition, a summary condition, and a text-only condition. After learning, learning outcomes were assessed with an immediate and delayed posttest, whereby the time of assessment was manipulated within subjects. The results indicate that visualization is the factor contributing most to benefits of drawing. Because the results regarding time of testing likely were confounded with benefits of retrieval practice (Roediger & Karpicke, 2006), time of testing was included as a between-subjects factor in the second study.
The second study examined the influence of generation and externalization on benefits of drawing. To this end, drawing was compared with an observation condition and a mental imagery condition. The results of the first study were replicated regarding the influence of an external pictorial representation being the main contributing factor to benefits of drawing. Moreover, it seems that drawing does not constitute a desirable difficulty (Bjork & Bjork, 2011; Bjork, 1994). As expected, the results showed that the size of the prognostic drawing effects depends on the type of posttest in that it was larger for assessments of higher-order knowledge than for assessments of lower-order knowledge. Additionally, the results indicate that free-hand drawing left learners with less cognitive resources available to engage in meaningful learning.
The third study investigated which type of instructional support is most effective in order to help learners benefit more from constructing drawings. To this end, a no-support drawing condition was contrasted to a low-support condition, a high-support condition, and a text-only control condition. The findings indicate that neither reducing the requirement to reason about irrelevant elements of the drawings alone, nor reducing the need to reason about the visual appearance of any element seemed to influence learning beyond an unsupported drawing effect. Moreover, instructional support increased rather than decreased cognitive demands associated with managing the drawing process. Drawing quality was positively associated with learning outcomes; however, the prognostic drawing effect was not larger for assessments of higher-order knowledge than for assessments of lower-order knowledge.
In conclusion, benefits of drawing seem to stem mainly from externalizing a visualization that drawing requires, rather than the actual generation of the drawing. Accordingly, recent theoretical frameworks of drawing (e.g., the CMDC; van Meter & Firetto, 2013) may overemphasize the role of generation. Thus, the results of previous studies comparing a drawing group to a non-drawing control group might be interpreted differently. Further research is needed to get more insight on boundary conditions of drawing including long-term effects, the influence of the type of posttest on the prognostic drawing effect, and the design of beneficial instructional support, as well as the influence of perceived difficulty for learner-generated drawing. Moreover, boundary conditions of multimedia learning could also affect benefits of learner-generated drawing and should be considered in future studies. |
en |
