Effects Of Encoding Solution Steps Of Worked Example On Principle Learning And Transfer

The goal of this paper was to examine the effects of encoding modes of solution steps of worked example on principle learning and transfer. Through the historical review of the research on learning from example since 1950's, firstly, the paper generalized the research on learning from worked example as the following three aspects:(l) The basic process mechanisms of learning from worked example developed itself in two ways. One was the cognitive skill acquisition through problem solving, which could be divided into three stages. The first stage, around 1960's, using the protocol analysis initially, concentrated on the fields of unwell-defined knowledge, with the main research topic shifting from the effects on problem solving to the process of solving the specific problem. At the second stage of 1970's, two kinds of relative research field, decision and inference, were developed. The third stage was through 1980's, in which the research focused on how a novice became into an expert. At the beginning of this stage, the research dealt with practical effects, motor skills, and the general factor model of transfer. With the rapid development and wide use of the computer skill in psychology research, the research on cognitive skill became increasingly precise and in-depth. More and more attention were paid to the role of example in cognitive skill acquisition, to the mechanism of negative transfer in cognitive skill acquisition and the cognitive skip. The other way the basic process mechanisms of learning from worked example developed itself was the research on machine learning based on artificial intelligence. The research can be divided into four stages, with the time ranging from 1960's to nowadays. At around 1960's, all sorts of adaptive system of model recognition were constructed based on network of neuron for the development of thecomputer. The main work at the beginning of 1970's was to construct the machine learning system that can simulate human's concept acquision. The third stage was from mid-1970 to the beginning of 1980's, in which machine learning went into its flourishing time. After that until now, the research on machine learning achieved greatly, and was mainly characterized as bobbing up like a cork of algorithmic of connection learning, increasing perfected of symbol learning;(2) The relationship between the encoding subgoal of worked example and learning. Review had been done on some encoding modes that affected the effective learning, which included intra-example features, such as the integration of diagrams and text, integration of aural and visual information and the integration of steps and subgoals, and inter-example features, such as multiple examples, variability of example and example-problem pairs;(3) Research on interaction between learner and the worked example suggested that self-explanations are an important learning activity during the study of worked examples, which were the dual processes of generating inferences and repairing mental models.Then, based on the anatomizing of the focuses of the research on learning from worked example since 1980's, this paper put forward the basic hypotheses, and tests it through three sets of experiments.Research one aimed at probing into the surface content effects on the access and use of principle. Its included two experiments comprising six sub-experiments. The findings showed that it was the encoding subgoal of steps of worked example that overcame the surface content effects on the access and use of principle, which was caused by the modification of the superficial story lines and superficial roles correspondence between examples and problems. In encoding subgoal of steps of worked example, the participants could develop the generalization of the example during the learning, so the modification of the superficial story lines and superficial roles correspondence between examples and problems had little effect on the access and use of principle.In research two, it was the effects of semantic contents of the subgoal labels had bee...