[[alternative]]The effect of Prior Knowledge, Text Structure, and Multimedia Presentations on Text Learning

[[abstract]]The purposes of this study were to explore the effects of a reader's prior knowledge, text structure, and multimedia presentation mode on text learning, and to test the predictions generated from the 'cognitive load theory', the 'structure-processing interaction model', and the 'structure-interestingness-engagement model'. Eighth-graders participated in the first and second experiments, which investigated the effects of prior knowledge and text structure on text learning, and the possible mechanism of their interactive effect on text learning. In the third and fourth experiment, eighth-grade students studied the text with picture-illustrations or the animation-illustrated text for investigating the possibility of using multimedia texts to enhance the reader's situational interest and reading engagement, and thereby eliminates the interactive effect of text structure and prior knowledge on text learning. The main findings are abbreviated in the following: First, a reader's prior knowledge and the text structure may influence the process and outcome of text learning. The perceived cognitive load of the higher-prior knowledge students was significantly lower than the lower-prior knowledge students. Yet there were no significant differences in reading time and situational interest between the two student groups. The higher prior knowledge students were able to build a better textbase (recall and recognition of factual knowledge) and situation model (transfer of text content for problem solving) after reading the text. A less coherent text may induce the reader to perceive a higher cognitive load than more coherent text. Yet the same situational interest and reading time was stimulated by the more coherent text. A more coherent text was helpful for both higher and lower prior knowledge students to build a textbase. The more coherent text was also helpful to lower-prior knowledge student for building a situation model. The above findings support the 'split attention hypothesis' of the cognitive load theory and the 'structure enhances processing hypothesis' of the structure-processing interaction model. Second, there exists an interactive effect of prior knowledge and text structure on text learning. Lower prior knowledge students benefit more from coherent text for reducing mental load and building textbase and situation model than higher prior knowledge students. For higher prior knowledge students, a more coherent text may not always be facilitative, and may sometimes be destructive to the building of a situation model. The destructive effect of a coherent text has not been completely verifiable, so the findings of McNamara & Kintsch (1996) and McNamara et al. (1996) have not been completely replicated. Besides, the data of the student's mental load, reading time, and learning performances from different reading tasks support the 'structure-processing trade off hypothesis' of the structure-processing interaction model as the most acceptable mechanism for the possible destructive effect of the coherent text on higher prior knowledge students' text learning. Therefore, in this study we named the above phenomenon the 'structure-processing trade off effect'. Third, multimedia texts may influence the process and outcome of text learning. A picture-illustrated text can reduce a reader's mental load and reading latency in text study, yet it cannot enhance the reader's situational interest. The animation-illustrated text can reduce the reader's mental load, and enhance their situational interest and reading latency in text study. In the mean time, the enhanced situational interest decreases the reader's perceived mental load, which supports the 'interest deteriorates load hypothesis' of the structure-interestingness-engagement model. Both the picture- and animation-illustrated texts are helpful for building a student's textbase and situation model. The animation-illustrated text has a substantial effect in reducing the 'structure-processing trade off effect', though it is unable to eliminate that effect completely. The findings that the multimedia text can enhance a reader's situational interest and studying latency, thereby facilitates their reading achievement, support the 'interest enhances processing hypothesis' of the structure-interestingness-engagement model. Fourth, both approaches using the mental load and reading achievement, or the three components of mental load, situational interest, and reading achievement, as the basis for designing the instructional efficiency indices for instructional material, have their limits. The instruction efficiency index calculated through the mental load and reading achievement is appropriate for the 'high load - high achievement' and 'high load - low achievement' learning situations. Yet this index has low discriminating efficiency for the 'high load - high achievement' and 'low load - low achievement' learning situations. When situational interest is added to the above instruction efficiency index, the discriminating efficiency of the new index is increased. Yet texts which result in 'high load - high achievement' and 'low load - low achievement' still cannot be differentiated by this new index. Fifth, According to the prior knowledge, text structure, and multimedia presentation influences on the cognitive process, interest status, and text learning outcome, the 'structure-interestingness-engagement model' was supported more than the 'cognitive load theory' and the 'structure-processing interact model'. These findings were discussed in detail and suggestions for further studies and applications were also recommended.