Problem Analysis: Cognitive Factors in Chemistry Problem Solving at Secondary School Level

[摘要] This project is about Problem Solving, but is unlike most existing Problem Solving research insofar as it concentrates on the nature of the problem itself, and not on the behaviour of the person solving the problem. The project involved two distinct stages, which can be treated (and indeed performed) quite separately. The first stage was the analysis of the chosen problems themselves, by the method of employing a panel of experts to answer questions (propositions) based on each individual problem in turn. The responses from the experts were analysed, and only a high degree of agreement was accepted as evidence for the presence or absence of the chosen factors, which were identified in the propositions. Since, on the first exercise, there was not sufficient agreement on some propositions, the propositions were revised and the exercise was done again. It was not necessary to use exactly the same panel, although some individuals did both exercises. The experts' responses thus formed the basis of a classification of the problems by chosen factors. The second stage used sets of marks obtained by pupils, selected randomly by the Scottish Examination Board, to compare the classification of each problem with pupils' success or failure in that problem. The initial intention, to try to emphasise particular factors by analysis of the patterns produced by linking individual pupils' marks with problem classification, yielded some interesting patterns, from which tentative suggestions could be made, but they could not be regarded as sufficiently reliable to be authoritative on their own account. The pupils were therefore treated as a group, and tables were produced of problems in order of merit, once more seeking patterns from the tables. Several different methods of displaying the data were tried, and again no useful pattern or conclusion emerged. The grouping of the classifications, within the proposition sets of Process, Concepts, Skills and Language, finally produced the long-sought patterns, and statistical analysis showed some trends towards differences in performances, which could be linked to problem characteristics. Even though these differences were not statistically significant, some useful comments could be made, and pointers recognised. The method of analysis used was designed to be applied to any problem set of a type amenable to expert panel analysis, and the software written for the analysis, on the BBC microcomputer, was made "content free" for such use. The project was as concerned with the design of a research tool as with the specific analysis of Chemistry problems, and I would venture to suggest that it succeeded in both these aims.

[发布机构] University:University of Glasgow

[关键词] Science education, Educational psychology