写作模糊综合评价研究——以全国高校俄语专业八级水平测试为例

Research on Fuzzy Comprehensive Evaluation of Writing Test——Take TRM-8 as an Example

写作评价具有较强的主观性和模糊性,如何保证评分结果一致且可靠始终是高风险语言测试亟须解决的难题。本文基于对现行写作评分方法的分析,主张运用模糊综合评价法对写作评分标准进行量化处理,并以全国高校俄语专业八级水平考试为例构建写作质量评估的模糊数学模型,分析该模型的优缺点,以期提供一套更加科学、客观、准确、全面的评分方法。

Since writing assessment is highly subjective and imprecise,ensuring the consistent and reliable assessment standard has been a difficult and urgent problem of high-stake language tests.The evaluation of writing tasks in foreign language tests usually adopts the overall scoring method or the subitem scoring method.The challenge lies in the inability of the rater to use and interpret the grade descriptions consistently,and it is also impossible to know whether they are scoring candidates strictly according to the grade standard.The author believes that the fundamental method to solve the problems above is to adopt a unified and objective scoring standard,which can be constructed through a fuzzy comprehensive evaluation method.The writing part of TRM-8 now uses the overall scoring method.In order to improve its scoring reliability,this paper built a fuzzy comprehensive evaluation model based on the writing scoring method of TRM-8,compiled a new rating scale and described the basic principles and specific steps of fuzzy comprehensive evaluation.This paper found that fuzzy comprehensive evaluation had the following advantages:First,the scoring reliability is high.The fuzzy comprehensive scoring method uses the principle of fuzzy synthesis to greatly reduce the influence of extreme factors,narrow the gap between personal evaluations,and reduce scoring errors.At the same time,it can ensure that the composition score is given according to the scoring standard,avoid subjective assumptions,and it is convenient to score candidates with unbalanced writing ability.Second,the operation is relatively simple.When judging,the scorer only needs to select the appropriate level in the specific indicators based on the student’s writing performance.The statistics and calculation of the score are performed by the computer according to the preset algorithm,and the scorer does not know the final score of the candidates,which greatly reduces the interference of human factors.Third,the results are easy to interpret.The "excellent,good,medium,and poor" scale is easier to grasp,and the scope of evaluation is limited to these four levels,which greatly improves the accuracy and consistency of the scoring.Fourth,the weight is diversified.The fuzzy comprehensive scoring method can not only weight different scoring dimensions,but also weight the sub-factors within each dimension,so that the scoring results are more accurate.Fifth,the compatibility is strong.The fuzzy comprehensive scoring method can combine the overall score with the sub-item score and make the overall and sub-item fuzzy evaluation respectively.The disadvantage of fuzzy comprehensive evaluation is that the determination of the evaluation index and its weighting system contains a certain degree of subjective arbitrariness;the fuzzy comprehensive scoring method itself cannot avoid the confusion of evaluation results caused by correlation,crossover or even duplication between evaluation indexes;fuzzy comprehensive evaluation method excludes the situation with full marks and zero marks.The fuzzy comprehensive scoring method provides an evaluation idea of "objectification of subjective scoring",which can effectively reduce random errors in the scoring process,improve the accuracy and reliability of test results,and reduce the drawbacks of existing scoring methods.It is worthy of wide application.