ABSTRACT
Objective:This study aims to explore the feasibility and practicality of ChatGPT in scientific research statistics through examples, in order to better understand and evaluate the application value of ChatGPT in scientific research statistics.Methods:Literature research was used to illustrate the principles of ChatGPT artificial intelligence model, combined with practical examples and tests, to assess the reliability of its results and analyze the advantages and limitations of ChatGPT in scientific research statistics.Results:ChatGPT can be applied to various scenarios in scientific research statistics, such as literature retrieval and organization, experimental design, statistical analysis programming, and statistical graphing. It can provide efficient scientific research statistical analysis capabilities, but it also has limitations.Conclusions:Although there are limitations, ChatGPT has a promising future in scientific research statistics, and it will be applied widely and deeply to greatly facilitate scientific research.
ABSTRACT
Objective Based on the three years' peer review data of a project fund,the results of fund peer review were analyzed and evaluated,to provide references for further improvement of the peer review and management.Methods Based on the fund's peer review results of 2145 projects in three years,descriptive statistics,single-item identification index were adopted,as well as RJ normality test,t-test and other statistical methods,to analyze and assess the overall data of fund,project categories,individual scores,etc.Results The score distribution of the three year peer review project of the fund is almost normal distribution,and the overall consistency of peer review shows a better trend.The analysis shows that the peer review experts of the fund have better consistency in terms of project innovation,rationality and characteristics.While there were differences in the peer review of applicability.Conclusions The three year peer review data of the fund show that three years' evaluation results are reliable,reasonable,and the quality of evaluation is good,showing a better development trend in both project quality and expert consistency.
ABSTRACT
Objective To study the dynamic changes of scientific research efficiency of 51 departments in a top three hospitals in Beijing from 2014 to 2016,and to provide reference information for improving the efficiency of scientific research in hospital departments.Methods The CCC model of DEA,BCC model were used for lateral analysis,and then Malmquist index was used for longitudinal analysis.Results There were 12 (23.53％) departments in 2014,11 (21.57％) departments in 2015,9 (17.65％) departments in 2016 data envelopment analysis was valid,from 2014 to 2016,41 (80.39％) departments of the hospital increased the total factor productivity.Conclusions The total factor productivity of the 51 departments in the hospital increased from 2014 to 2016,and technological progress was the main reason for its growth.
ABSTRACT
Objective Discussion on the methods of retrospectively selecting peer review experts based on the results of review.Methods Based on the review data of peer reviewers,peer review experts are retrospectively selected based on the voting,rating and comprehensive evaluations,as well as their variations and errors during the review.Results Through the retrospective selection of peer reviewers,there is a better convergence of reviews,which assures the quality of future peer reviews.This method is superior to the prospective selection of peer reviewers.Conclusions Based on the results of the review,retrospective selection of peer review experts is feasible and helpful for the selection of high-level peer review experts,which can improve the quality of peer review and standardize peer review.
ABSTRACT
Objective Based on the data of scientific research output in a top three hospital in Beijing from 2015 and 2016,stochastic frontier analysis is conducted to estimate the efficiency of scientific research performance in different departments of this hospital,so as to draw the impetus for the growth of departments.Methods According to the results of the stochastic frontier model,comprehensive analysis and evaluation are conducted for the research performance of various departments.Results The overall average technical efficiency of the hospital is 0.44,and improvement space for the technical efficiency of scientific research performance is still 0.56.There are significant differences in the technical efficiency of each department.Among the four major categories,the D section is higher than the technical efficiency of class A,B and C departments.Conclusions The overall technical efficiency of the hospital is not high,the level of scientific research output of the four types of departments is not balanced,which should be emphasized in the improvement of scientific research performance level.
ABSTRACT
Objective Evaluate the efficiency of scientific research output of the 54 departments in a hospital,to put forward improvement suggestions based on the evaluation results.Methods Select appropriate indicators of scientific input and output,use the Data Envelopment Analysis method to evaluate and analyze the efficiency.Results According to the analysis of DEA,calculate the values of overall efficiency,technical efficiency,scale efficiency and scale income.Then compare and analyze the relative efficiency of different units scientific output,to identify the relatively superior department a mong the various categories.Conclusions According to the evaluation results,to find out the input surplus and insufficient output of each decision units.Then we will put forward suggestions on hospital resource allocation to optimize the scientific input and output,to improve the competitiveness of the hospital,and to activate the potential of each department's scientific research.