医学研究生学位论文统计学方法应用状况的调查分析

Analysis of the Status Quo of Statistical Methods Employed in MS and PhD Graduation Theses of Medical Students

统计方法的应用科学与否对实证性科学研究的质量具有决定性的影响。为了解某校研究生学位论文统计方法应用现状,对该校2006-2008年度医学研究生学位论文进行了统计审核。结果表明,多数学位论文存在2-7类的统计学错误,仅1.8%的学位论文没有统计学错误。医学研究生学位论文的统计学错误主要集中在假设检验结果的表达和解释、统计分析方法和统计软件的应用和表述、以及统计图表的应用等方面。硕士和博士学位论文中居前三位的统计学错误类型相同,包括缺P值或P值不精确、缺少相应的检验统计量、统计软件和统计方法交代不够等。本研究结果显示,应进一步改善该校医学研究生的统计学知识的应用能力。

The scientific employment of statistics is of great significance in the academic quality of quantitative medical research. To understand the status quo of statistical methods used in MS and PhD graduation theses at a medical university, we reviewed the theses written by medical graduate students who graduated in 2006 to 2008. The results showed 2-7 types of statistical errors in the majofity of the completed MS and PhD theses, and only 1.8% of them had no statistical error. The error types in the degree papers included poor expression and interpretation of statistical test results, misuse and wrong presentation of statistical analytical methods and software, and inappropriate use of tables or graphs. MS and PhD theses had the same most common three types of statistical errors ： failure to specify a p-value or giving an imprecise p-value, lack of corresponding statistical test results, and unclear explana- tion of statistical methods and software. The results show that it is necessary to improve the medical graduates＇ ability to apply statistical knowledge in the medical university.