| To scientifically evaluate the quality of the medical research project is one of the difficulties that the research management hoping to solve. Many experts had carried studies on it, but were limited to analyze the output instead of analyzing the input-output in common. In this paper the data of medical research projects ended study between 1990 and 1999 from three sampling units are analyzed, applying the opinion of the systematic theory, conforming to the scientific procedure of the comprehensive evaluation, to institute a set of evaluation indexes, to expore the influential factors of the output variables, research effectiveness and research benefit by employing various statistical methods. Based on Delphi method, advance an applicable discriminant model of the quality of the research project, To explore the optimum comprehensive evaluation technique by assessing the quality of the ended-study projects between 1990 and 1999 from Fujian Mecial University.1 Research DesignTo sample the medical research projects ended study between 1990 and 1999 from Fujian Medical University, the First Hospital Affiliated to Fujian Medical University, Xiehe Hospital Affiliated to Fujian Medical University as the investigation units. To investigate the basic data, data about paper embodiment of the projects and data on the rank promotion of the research technician participated the research projects. To analyze the following four aspects extensively: to institute a set of the evaluation indexes of the quality of the medical research project, set up two discriminant models of the quality of project .analyze the relationship between the input variables and output variables, comprehensively assess the quality of the project, by employing various statistical techniques, and compare the adaptability of every method.2 Research Results2.1 Institute a set of evaluation indexesUnder the principle of the accessibility of the evaluation index, to analyze the basic set of evaluation indexes by employing Delphi method, Index Cluster method, Coefficient of Variation method commonly, and institute a set of three-cluster evaluation indexes that can be applied to set up the discriminant model, to assess comprehensively the quality of projects The research project quality is the total aim, the first cluster includes research input and output, the corresponding second cluster belonging to research input is composed by research project, research technician, et al. Research output include eleven variable in the second cluster, such as academic article, article embodiment, article communication, et al.2.2 The discriminant model of quality of the research projectThis paper advances the two-step discriminant model regarding quality of the research project. First, to categorize all the research projects as advanced, super, modest, and inferior conforming to certain criteria, then discriminating the quality of the modest projects. The two discriminant models are satisfactorily tested.2.3 The relationship between input and output variables 2.3.1 Quantile regression of the output variableSingle output variable does not conform to normality distribution, so quantile regression can be applied toanalyze the influential variable of the output variable. Only academic article, article embodiment, articlecommunication, research product, product appraisal, and personnel foster are carried on median regression.The influential input variables of the academic article are: project source (XI) , the overall fibre of thepresiding (X4) , project fund (X5) , whether the project finished or not (X6) , the total people year of theresearch technician overall fibre (X8)The input variables that can be absorbed into the article embodiment equation are follows: the overall fibre ofthe presiding, project fund, the total people year of the research technician overall fibre.The variables influencing academic communication are: the gender of the presiding, the overall fibre of thepresiding, project fund, the total overall fibre o...
|
PDF Full Text Request