Gardner M J, Machin D, Campbell M J
Br Med J (Clin Res Ed). 1986 Mar 22;292(6523):810-2. doi: 10.1136/bmj.292.6523.810.
Two check lists are used routinely in the statistical assessment of manuscripts submitted to the "BMJ." One is for papers of a general nature and the other specifically for reports on clinical trials. Each check list includes questions on the design, conduct, analysis, and presentation of studies, and answers to these contribute to the overall statistical evaluation. Only a small proportion of submitted papers are assessed statistically, and these are selected at the refereeing or editorial stage. Examination of the use of the check lists showed that most papers contained statistical failings, many of which could easily be remedied. It is recommended that the check lists should be used by statistical referees, editorial staff, and authors and also during the design stage of studies.
在对提交给《英国医学杂志》的稿件进行统计学评估时，通常会使用两份检查清单。一份用于一般性论文，另一份专门用于临床试验报告。每份检查清单都包含有关研究设计、实施、分析和呈现方面的问题，对这些问题的回答有助于进行全面的统计学评估。只有一小部分提交的论文会进行统计学评估，这些论文是在审稿或编辑阶段挑选出来的。对检查清单使用情况的审查表明，大多数论文存在统计学缺陷，其中许多缺陷很容易得到纠正。建议统计审稿人、编辑人员和作者在研究设计阶段以及整个过程中都使用这些检查清单。
