Proposal view
| Proposal Type: | Individual Paper |
|---|---|
| Domain: | Learning and Instructional Technology |
| SIG: | Comprehension of Text and Graphics |
| Type | Submitted Paper |
| Equipment |
Slide projector |
| Paper Details |
|---|
| Title | The effects of realistic detail in learning from dynamic visualizations |
|---|---|
| Abstract | The study investigated the role of realistic and schematic dynamic visualizations for knowledge acquisition. Seventy-nine university students with little prior knowledge studied two dynamic visualizations on the process of mitosis before answering different types of questions that assessed learning outcomes. Both visualizations illustrated the same process and were accompanied by the same verbal explanations. There were four experimental conditions: In a realistic-realistic condition learners studied the same high fidelity dynamic visualization twice (i.e., movies recorded via a microscope). In the schematic-schematic condition mitosis was depicted by a dynamic series of simple line drawings, which were shown twice to students. In the schematic-realistic condition, students first saw the schematic dynamic visualization, which was followed by the realistic one. Finally, in the realistic-schematic condition this presentation order was reversed. The results showed that learners in the realistic-realistic condition answered fewer multiple-choice questions correctly, identified fewer errors in manipulated static pictures, and completed fewer partial static pictures correctly than learners in all the other conditions. This pattern of results was, however, no longer observable in the delayed posttest that was taken two weeks later. The results for the evaluation and cognitive load measures furthermore indicated that especially those students who had seen both types of visualizations rated the schematic visualizations as more helpful and reported lower task demands. Results of the study are discussed against the background of findings by Dwyer, who found that schematic static pictures were superior to realistic pictures only for specific tasks and under specific instructional conditions. |
| Summary | An important question when designing dynamic visualizations pertains to the amount or realistic detail that should be depicted in the visual display. On the one hand, it can be argued that highly realistic visualizations will be more authentic and thus will facilitate the recognition of the phenomenon in the real world. For instance, learning about the process of cell reproduction (i.e., mitosis) with realistic visualizations might enable learners to identify cells in their different stages of cell reproduction under the microscope. On the other hand, realistic visualizations entail more irrelevant details and therefore may direct learners’ attention away from the important aspects (Dwyer, 1976; Lowe, 1998). Schematic visualizations that are kept as simple as possible and contain only relevant information may allow highlighting the important aspects of a process and thereby reduce the need for extensive search in complex displays. However, learners studying schematic visualizations only might have difficulties when being confronted with the real-world phenomenon (e.g., cells under a microscope). Thus, there may be an interaction between the type of dynamic visualization and the type of task that is used to measure knowledge acquisition (Betrancourt & Tversky, 2000). This interaction has been shown to exist for learning from static pictures already in several studies conducted by Francis Dwyer; however, it is not clear yet whether these results will hold for dynamic visualizations. There is evidence that learning from dynamic visualizations is a more demanding process than learning from static pictures due to the rapid changes in the display (Lowe, 1998). Thus, it might well be that realistic dynamic visualizations – contrary to realistic pictures – prove unsuitable for learning under any circumstances and not only for specific tasks. The reported study investigated this role of realistic and schematic dynamic visualizations for knowledge acquisition in the domain of cell biology. Participants. Participants were 79 university students with little prior knowledge in the domain as assessed by a multiple choice pretest. Materials and Design. All students studied two dynamic visualizations that illustrated the process of mitosis. Both visualizations illustrated the same process and were accompanied by the same verbal explanations. The visualizations were system-controlled, that is, learners could not stop or replay them. There were four experimental conditions: In a realistic-realistic condition learners studied the same realistic dynamic visualization twice (i.e., movies recorded via a microscope). In the schematic-schematic condition the process of mitosis was depicted by a dynamic series of simple line drawings, which were shown twice to students. Additionally, two combined visualization conditions were investigated. In the schematic-realistic condition, students first watched the schematic dynamic visualization, and then they saw the same process displayed in the realistic visualization. Finally, in the realistic-schematic condition this presentation order was reversed. The latter two conditions were implemented to test whether a combined presentation of both formats is superior to a single format if both representational formats have complementary roles for instruction (Ainsworth, 1999). Moreover, it was tested which sequence of representational formats would be best for learning. It was assumed that seeing a schematic dynamic visualization first might constrain the interpretation of the subsequent realistic visualization to the relevant aspects and thus might yield better outcomes than the reversed sequence. Dependent Measures. Students were asked to rate the helpfulness of the instructional materials for learning and report on their experienced cognitive load after having seen the first dynamic visualization and after the second one, respectively. Learning outcomes were assessed by 22 verbal multiple-choice questions, which were presented immediately after learning with the visualizations and again two weeks after the experiment had taken place. In an additional pictorial test, learners had to identify errors in manipulated static schematic displays, complete partial drawings for the relevant parts of the cell during the different stages of mitosis, and arrange multiple static realistic pictures according to the order in that they would appear during mitosis. Results and Discussion. Learners in the realistic-realistic condition answered significantly fewer verbal multiple-choice questions correctly, identified fewer errors in manipulated static pictures, and completed fewer partial static pictures correctly than learners in all the other conditions. The latter three conditions that all contained schematic visualizations did not differ from each other. There were no differences for the ability to identify the correct sequence of realistic static pictures among conditions. Interestingly, the inferiority of highly realistic dynamic visualizations with regard to most of the dependent variables was observable only immediately after the learning phase, whereas there were no differences among conditions in the delayed posttest that was taken two weeks later. The results for the evaluation and cognitive load measures furthermore indicated that especially those students who had seen both types of visualizations rated the schematic visualizations as more helpful and reported lower task demands. Taken together, these results do not support the findings of Dwyer, according to which schematic static pictures were superior to realistic pictures only for specific tasks. In the current experiment, realistic dynamic visualizations proved inferior for a variety of tasks. Moreover, learners who had studied only schematic dynamic visualizations were able to sort realistic pictures, although they had never seen them before during the experiment. Therefore, these findings might indicate that for dynamic visualizations the complexity and thus detailedness of the visuals needs to be kept at a minimum irrespective of the learning objective. One issue that has yet to be explored, however, is that Dwyer found a superiority of realistic pictures only when there was sufficient learning time available (e.g., with self-paced materials) and for learners with higher levels of prior knowledge. Therefore, follow-up experiments need to investigate whether the current results can be replicated for interactive visualizations and whether the level of prior knowledge moderates the found effects. Finally, in order to control for possible motivational and training effects, which might occur due to watching two visualizations of the same content, the effects of studying single visualizations should be assessed in future experiments. |
| Keywords | Comprehension Computer-assisted instruction Multimedia and hypermedia |
| Appendices | |
| Authors | ||||||
|---|---|---|---|---|---|---|
| Name | Surname | Institution | Country | EARLI Number | Presenting | |
| Katharina | Scheiter | University of Tuebingen | Germany | k.scheiter@iwm-kmrc.de | * | |
| Peter | Gerjets | Knowledge Media Research Center | Germany | p.gerjets@iwm-kmrc.de | ||
| Thomas | Huk | Westermann Verlag | Germany | t.huk@tu-bs.de | ||
