广西14种医学类科技期刊统计学应用调查研究

目的 调查广西医学类科技期刊统计学的应用现况,以期找出存在的不足及差距,为同行提高统计学应用水平提供借鉴.方法 以“中国知网数据库”或“万方数据库”为检索平台,检索2010年广西14种医学类科技期刊发表的所有论文,按约10％的比例抽取有统计学内容的论文共 535篇,分别统计不同期刊的统计学应用情况.结果 535篇论文中,所使用的统计学软件包以SPSS为主,占91.59％（490/535）,其次为PENS占6.54％ （35/535）,SAS最少,仅占1.87％（10/535）;使用1种统计学方法共223篇（41.68％）,2种247篇（46.17％）,3种54篇（10.09％）,4种及以上11篇（2.06％）;使用的统计学方法居前三位的是t/u检验356篇（66.54％）,卡方检验为308篇（57.57％）,方差分析为118篇（22.05％）.结论 广西医学类科技期刊统计学使用现况与国内其他期刊相似,大部分期刊以t/u检验、卡方检验和方差分析为主,且使用的种类较少,复杂方法使用率较低,错误的情况仍不少,应引起编辑同行高度重视,以期进一步提高统计学应用水平.

Investigation on statistics application of 14 kinds of medical sci-tech journals in Guangxi

Objective To investigate application status of statistics in medical sci-tech journals in Guangxi,so as to find out insufficiency and gap as well as to provide reference for editorial counterparts in improving application level of statistics. Methods Retrieval platform such as China National Knowledge Infrastructure （CNKI） or Wanfang Database were used to check all thesises which were published in 14 kinds of medical sci-tech journals in Guangxi in 2010.Then,535 thesises with statistics were selected according to the proportion of about 10%, and statistics application in different medical sci-tech journals were counted,respectively. Results SPSS statistical software package was mainly used in 535 thesises mentioned above, accounting for 91.59 percent （490/535）, followed by PEMS, accounting for 6.54 percent （ 35/535）, and SAS was least used, accounting for only 1.87 percent（10/535） .One statistical method was used in 223 thesises （41.68 percent）, two in 247 （46.17 percent）, three in 54 thesises （10.09 percent）, and four and above in 11 thesises （2.06 percent）.Statistical methods that listed on the top three were t/u test in 356 thesises（66.54 percent ） ,chi-square test in 308 thesises （57.57 percent）, and analysis of variance in 118 thesises（22.05 percent）.Conclusion Application status of statistics in medical sci-tech journals in Guangxi is similar to other journals in China, t / u test, chi-square test, and analysis of variance are largely used in most journals.Moreover, only a few statistical methods are used, usage rate of complicated methods is low, and there are still many errors. Thus, editorial counterparts should pay high attention to those above-mentioned problems, so as to improve application level of statistics.