SCI论文产出影响因素的最优尺度回归分析

Optimal scaling regression analysis for impact factors of SCI papers output

目的分析某大学附属医院2013-2017年SCI论文产出情况及影响因素,为科研管理者调整SCI论文激励政策提供支持与依据。方法使用动态数列及最优尺度回归法对医院SCI论文产出情况及影响因素进行统计分析。结果医院SCI论文产出年均增长30.3篇;年均环比增长率37.5%,高于国家9.5个百分点;2017年定基增长率257.4%,高于国家同期水平89个百分点。对全部180位第一/通讯作者的516篇SCI论文进行最优尺度回归分析,模型具有统计学意义(F=10.513,P<0.001,调和R^2=0.409)。硕博导师身份(P<0.001)、平均课题经费(P=0.001)、职称级别(P=0.048)与性别(P=0.005)对医院SCI论文产出有统计学关联(P<0.05)。四者重要程度之和为0.961,即上述4个因素约占该模型总解释度的96.1%。量化评分转换图显示博士生导师、高平均课题经费、高级职称及男性的SCI论文产出量化评分更高。结论硕博导师身份。平均课题经费。职称级别及性别可能是SCI论文产出的影响因素。医院可围绕博士生导师的引进及科研平台建设,科研匹配经费的精准支持,与职称评定政策相结合,提升女性科研工作者的论文激励等方面调整SCI论文激励政策,以促进SCI论文产出。

Objective To analyze the SCI papers output and its impact factors of a hospital during 2013-2017,to provide supports and proofs for scientific research administrators to update incentive policies for stimulate growth of SCI Paper publication.Methods Dynamic sequence and Optimal Scaling Regression methods were used to conduct statistical analysis of the output and impact factors of hospital-published SCI papers.Results The average increased number per year of hospital-published SCI papers output was 30.3 with an annual increase rate of 37.5%,which was 9.5% higher than the national average.In 2017,the growth rate of a fixed base of hospital-published SCI papers was increased to 257.4 %,which was 89% higher than the national rate during the same period.The model was statistically significant (F =10.513,P< 0.001,adjusted R2 =0.409) according to the Optimal Scaling Regression analysis of 180 first/corresponding authors out of 516 SCI papers.We found that authors' status as tutor of post-graduate student (P<0.001),average project fundings (P=0.001),professional titles (P=0.048) and genders (P=0.005) have significant impacts on the output of hospital-published SCI papers (P<0.05).The sum of the importance of the four factors above was 0.961,namely account for 96.1% of the whole model.The quantitative score conversion chart shows that doctoral tutors,high average project funding,senior professional titles,and male authors of the SCI papers have higher quantitative scores.Conclusions Tutor status,average project fundings,professional titles and genders could impact the output of SCI papers.Scientific research administrators could adjust incentive policies of SCI papers,based on the science platform construction inclining to doctor tutors,adding matching funds for science and research,bonding professional title appraisals and rising incentives for female researchers on SCI papers,to promote SCI papers output.