Again review of research design and statistical methods of Chinese Journal of Cardiology / 中华心血管病杂志

ABSTRACT

<p><b>OBJECTIVE</b>To re-evaluate and compare the research design and the use of statistical methods in Chinese Journal of Cardiology.</p><p><b>METHOD</b>Summary the research design and statistical methods in all of the original papers in Chinese Journal of Cardiology all over the year of 2011, and compared the result with the evaluation of 2008.</p><p><b>RESULTS</b>(1) There is no difference in the distribution of the design of researches of between the two volumes. Compared with the early volume, the use of survival regression and non-parameter test are increased, while decreased in the proportion of articles with no statistical analysis. (2) The proportions of articles in the later volume are significant lower than the former, such as 6(4%) with flaws in designs, 5(3%) with flaws in the expressions, 9(5%) with the incomplete of analysis. (3) The rate of correction of variance analysis has been increased, so as the multi-group comparisons and the test of normality. The error rate of usage has been decreased form 17% to 25% without significance in statistics due to the ignorance of the test of homogeneity of variance.</p><p><b>CONCLUSION</b>Many improvements showed in Chinese Journal of Cardiology such as the regulation of the design and statistics. The homogeneity of variance should be paid more attention in the further application.</p>