《中国循环杂志》刊出论文的统计学方法评价

发布时间：2018-04-29 22:23

本文选题：期刊 + 统计学　；参考：《中国循环杂志》2017年03期

【摘要】：目的:评价《中国循环杂志》刊出论文的统计学方法的使用情况及存在问题。方法:对《中国循环杂志》2010年、2011年和2015年发表的论著类文章中统计学方法使用情况进行收集、分析和评价,发现存在的问题,并在不同年份间进行比较。结果:共404篇文章纳入评价,其中366篇使用了统计检验方法,32篇仅进行了描述性统计,另有6篇论著为Meta分析。使用频率较高的统计检验方法依次为:t检验(57.4%)、卡方检验(56.7%)、方差分析(35.2%)和回归分析(23.0%)。方法学部分对统计学方法描述存在的主要问题有:描述的方法与实际使用的方法不匹配、描述存在欠缺或错误。其中描述存在欠缺或错误的比例在2010年和2011年为36.6%(71/194),2015年为26.2%(55/210),差异有统计学意义(P=0.02)。统计检验方法使用错误主要为忽略研究设计错用统计方法、忽略统计方法使用条件等,统计方法使用错误的比例在2010年和2011年为19.1%(33/173),2015年为19.7%(38/193),两年份间无明显差异(P=0.88)。结论:2015年刊出的论文与2010年和2011年相比,统计学方法正确使用情况有所改善,但仍存在不同程度的问题,应进一步加强论文的统计学审稿工作。
[Abstract]:Objective: to evaluate the use and problems of statistical methods in Chinese Journal of Circulation. Methods: the use of statistical methods in the articles published in 2010, 2011 and 2015 were collected, analyzed and evaluated, and the existing problems were found and compared in different years. Results: a total of 404 articles were included in the evaluation, of which 366 were evaluated by statistical test method, 32 were descriptive statistics, and 6 were Meta analysis. The statistical test methods with higher frequency were: 57.4% t test, 56.7% chi-square test, 35.22% ANOVA and 23.0% regression analysis. In the methodology part, the main problems in the description of statistical method are as follows: the method of description does not match with the method used in practice, and the description is deficient or wrong. In 2010 and 2011, it was 36.6 percent, and in 2015 it was 26.255 / 210. The difference was statistically significant. The errors in the use of statistical test methods are mainly due to the neglect of research and design of statistical methods and the neglect of the conditions of statistical methods. The proportion of errors in using statistical methods was 19.133 / 173 in 2010 and 2011, 19.738 / 193 in 2015, and there was no significant difference between the two years (P < 0.88). Conclusion: compared with 2010 and 2011, the correct use of statistical methods has improved, but there are still some problems, so we should further strengthen the statistical review of papers.
【作者单位】：北京协和医学院中国医学科学院国家心血管病中心阜外医院社区防治部;首都医科大学附属北京儿童医院;
【分类号】：+

【相似文献】