Statistical Methods in the Articles of the Journal of the Korean Society of Anesthesiologists from 1981 to 1990 / 대한마취과학회지

ABSTRACT

The purpose of this study is to investigate the error of using statistics in the articles of the Journal of the Korean Society of Anesthesiologists from 1981 ta 1990 and to present a program that may be of some help correcting the statistical error. We classified statistically the journals into original articIe, review and case report, and original articles into that of usng inferential statistics, descriptive statistics and no statistics. Then, we analyzed the articles of inferential statistics in the aspect of our criteria for statistical errors. The proportion of using erroneous inferential statistics was 80-100% from 1981 to 1986 and it decreased by 60% afterwards. But the proportion was still 67% in 1990. The representative errors are the no statistical analysis in spite of the necessary cases, the p-value only without the description of statistical method and using t test inappropriately in the comparison of more than 3 groups without Bonferroni correction. The other errors in using statistics were confusion between related data and independent data, inadequate numbers for Chi-square test, inapproate follow-up to variance analysis and inapproate parametric test for data in nominal or ordinal scales.