医学文献中统计图使用比例及其应用错误的抽样调查.doc

医学文献中统计图使用比例及其应用错误的抽样调查

    

摘  要

目的 通过抽样研究调查中华医学期刊网1998-2012年收录的文献中各种统计图的使用及常见错误出现情况，了解医学期刊论文中各种统计图的使用比例及常见错误的出现比例，并对其做分层分析(按期刊类型、发表年份等分层)，为有针对性调整医学统计学中统计图的教学方案与规范医学论文中统计图的制作提供科学依据。方法 采用两阶段完全随机抽样从总体中(1998～2012年中华医学会期刊网中已收录文献共计408600篇)抽取已发表的医学文献作为样本。一阶段以页(按文献发表时间排序将总体中的文献分为40860页)为抽样单元，二阶段以每页内的文献(每页包含10篇文献)为抽样单元；依据医学统计学中统计图的制作和使用规范确定正确使用统计图的判断标准并制定调查表，对每篇文献中的基本信息，统计图使用情况及常见错误出现情况进行调查, 以期刊类型、发表年份、作者性别、作者单位、作者学历、课题资助类型等因素分层对医学期刊文献中各种统计图的使用比例及其常见应用错误的比例进行统计分析。结果 调查结果显示中华医学期刊网1998-2012年收录文献中统计图的使用率为7.4%；直条图，百分条图，圆图，线图，对数线图，直方图，箱式图，统计地图及散点图使用比例分别为：3.3%，0.0%，0.1%，4.4%，0.1%，0.1%，0.1%，0.0%，0.6%，其中以线图使用比例最高(4.4%)，统计地图最低(0.0%)；统计图使用错误的比例为14.3%，直条图，圆图，线图，对数线图，直方图，箱式图及散点图使用错误比例分别为13.2%， 0.0%，0.0%，16.7%，0.0%， 7.5%，其中以直方图使用错误比例为最高(16.7%)；图序编制不规范、标题缺失、标题位置不正确、横纵坐标设置不规范及图例应用不规范的比例分别为2.2%，0.5%，2.0%,9.4%,3.8%，其中以横纵标目设置不规范的出现比例为最高(9.4%)，标题缺失出现比例最低(0.5%)。分层分析结果显示：不同期刊类型、不同作者性别、不同作者单位类别中各种统计图的使用比例及常见错误比例差异均无统计学意义(P>0.05)；不同文献发表年份中统计图使用错误比例(F=3.441，P=0.032)、直条图使用错误比例(F=7.853，P=0.000)、图序使用不规范比例(F=4.386,P=0.013)、横纵标目使用不规范比例(F=4.116，P=0.016)差异均有统计学意义；不同作者学历水平(F=3.299，P=0.000)和不同课题资助类型(F=8.868，P=0.000)间统计图的使用比例差异有统计学意义。结论 中华医学会期刊1998～2012年收录文献中统计图的使用比例(7.4%)不高，但统计图的使用错误的比例相对较高(14.3%)；文献中统计图的使用比例与文献发表年份、作者学历水平、课题资助有关；文献中统计图的使用错误比例与文献发表年份有关；统计图常见错误中图序不规范、横纵标目不规范与文献发表年份有关。

关键词：统计图；二阶段完全随机抽样；错误比例；分层分析

The Sampling survey of the use ratio and its application error of statistical figure in Medical literature

       

                             Abstract

Objectives through sampling survey the use and common errors of various statistical figure in the literature published from 1998 to 2012 in Chinese medicine journal net, understand the use and common error ratio of various statistical figure in medical journal articles, and make hierarchical analysis to them (according to layered factors, such as the journal types,published year), which will provide the scientific basis for targeted adjustments teaching plan of medical statistics statistical chart and standard the statistical chart in medical papers. Methods using two stage completely random sampling the published medical literature from the population (the Chinese medical association journal net from1998 to 2012 has included a total of 408600 literature) as a sample. The sampling unit is page (literature in general can be divided into 40860 pages when it’s sorted by published time) in the first stage, The second phase regards the document in each page(each page contains 10 articles) as the sampling unit; According to the production and specifications of the statistical figure in medical statistics are used to determine the judgment standard of correct use and formulate questionnaire statistical figure. Layered by basic information of each document, including periodical type, year of publication in journal, author of gender, the author unit, the author qualifications, the types of project financing, and analyzed utilization rate and the common application error ratio of a variety of statistical graphs in medical journal articles. Results Results show that  the use proportion of statistical figure in the literature collected in the Chinese medical association journal from 1998 to 2012 is 7.4%.The use proportion of Straight figure, percentage of diagram, chart, diagram, logarithmic diagram, histogram, box-type graph, statistical map and a scatter diagram using ratio is 3.3%, 0.0%, 0.1%, 4.4%, 0.1%, 0.1%, 0.1%, 0.0%, 0.6% respectively, of which diagram use proportion is highest (4.4%) and the statistical map is lowest (0.0%).The using error ratio of Statistical figure is 14.3%. The using error ratio of straight line diagram, chart, diagram, logarithmic diagram, histogram, box-type graph and a scatter diagram is 13.2%, 0.0%, 0.0%, 16.7%, 0.0%, 7.5% respectively , of which the highest using error ratio is the histogram (16.7%). The non-standard Figure order , lack of the title, wrong location of title and non-standard horizontal ordinate and non-standard application of illustration is 2.2%, 0.5%, 2.0%, 9.4%, 3.8% respectively.Stratified analysis results showed that the differences of use proportion of statistic figure and its usual mistakes had no statistical significance between different types of periodicals, the author of sex and author's unit (P > 0.05). the different of using error ratio of Statistical figure in literatures (F = 3.441, P = 3.441), using error ratio article straight figure (F = 7.853, P = 0.000),  using non-standard ratio of statistic figure(F = 4.386, P =0.013), Horizontal ordinate using non-standard ratio of horizontal ordinate (F = 4.116, P = 4.116)between different published years have statistical significance.the difference of use ratio of statistical figure between different authors degree level (F = 3.299, P = 0.000) and different types of project funding (F = 8.868, P = 8.868) is statistically significant.

Keyword：statistical figure；two stage completely random sampling；error rate；layered analysis 

目  录

摘  要 I

Abstract Ⅱ

前  言 1

1.材料与方法 2

1.1 调查对象 2

1.2 调查方法 3

1.3 调查内容 3

1.4 调查表制定 3

1.5 质量控制 4

1.6 资料统计分析 4

2结果 5

2.1调查医学期刊文献的基本特征 5

2.2中华医学期刊网期刊文献中统计图使用及常见错误出现情况 5

2.2.1 中华医学期刊网期刊文献中各种统计图的使用情况 6

2.2.2 中华医学期刊网期刊文献中各种统计图错误使用的情况 7

2.2.3 中华医学期刊网期刊文献中各种统计图各种常见错误情况 8

2.3中华医学期刊网期刊文献中各统计图使用及常见错误情况的分层分析 8

2.3.1不同期刊文献中各种各种统计图的使用及常见错误情况 8

2.3.2不同发表年份文献中各种统计图的错误使用及常见错误情况 11

2.3.3不同作者性别文献中各种统计图的错误使用及常见错误情况 14

2.3.4不同作者单位文献中各种统计图的使用及常见错误情况 17

2.3.5不同作者学历水平文献中各种统计图的使用及常见错误情况 19

2.3.6不同基金资助文献中各种统计图使用及常见错误情况 21

3讨论 25

4结论 27

参考文献 28

致 谢 29