研究生学位论文质量评价指标体系的建立

Study on Quality Evaluation Indicators System on Graduate Degree Thesis

目的:构建研究生学位论文质量评价指标体系,改进现有评价标准。方法:邀请12名相关领域专家作为调查对象,应用Delphi法对初步构建的研究生学位论文质量评价指标体系进行3轮咨询,利用层次分析法确定指标权重并建立综合评价模型,同时采用秩和比综合评价法进行实测评价以分析指标体系的科学性。指标体系的信度评价采用Cronbachα系数,效度评价采用探索性因子分析方法。数据分析采用IBM SPSS 20.0统计软件,以P≤0.05为具有统计学意义。结果:本研究构建了由9个一级指标和52个二级指标构成的指标体系。指标体系整体内部一致性信度高(Cronbachα=0.902),探索性因子分析共提取6个因子,分别命名为"文字表述"、"研究方法"、"资料分析"、"写作规范"、"文献综述"、"选题",累计贡献率为50.32%。该指标体系的实测评价结果与秩和比综合评价法结果呈正相关(r=0.942,P<0.05);秩和比综合评价法回归方程为RR=0.100+0.080×Probit,根据Probit值将实测对象分为优秀、良好、及格和不及格四档。结论:本研究建立的指标体系专家认可程度高,整体信度较好,结构效度尚不理想,且与传统的秩和比综合评价法结果一致性高,说明该指标体系整体上科学、合理、有效,具有应用价值。

Objective:to set up quality evaluation indicators system on graduate degree thesis,optimizing the current evaluation criteria.Method:12 experts in relative fields were invited to give their opinions on the indicators system through 3 rounds of Delphi's questionnaires;analytic hierarchy process(AHP)was used to determine the weight of each item.Rank Sum Ratio(RSR)was used for comprehensive evaluation and classification.Cronbachαcoefficient was used to evaluate the reliability of the indicators system.Exploratory factor analysis was used for validity evaluation.Data analyses were used by IBM SPSS 20.0 software.The difference with P<0.05 was statistically significant.Results:the indicators system consisted of 9 domains and 52 indicators.The overall Cronbachαof the indicators system is 0.902;6 factors were extracted by Exploratory Factor Analysis,and respectively named"expression","research methods","data analysis","standardized writing","literature review"and"thesis topic".The cumulative contribution rate is 50.32%.The empirical study result was positive related to the RSR(r=0.942,P<0.05).The regression equation of RSR was R譓R=0.100+0.080×Probit.According to the value of Probit,the measured objects were graded as"excellent","good","passed"and"failed".Conclusion:the evaluation indicators system was confirmed by experts and was scientific and reasonable with a good reliability,but the structure validity is not ideal.The sorted result of the empirical study is highly related with RSR,which indicates the scientificity and feasibility of the model.