In the 10 years from 1980 to 1989, out of 745 papers in the Journal of Korean N europsychiatric Association, 506(67.9%) presented num erical results. The authors reviewed 506 papers and pointed out problem s in terms of statistical description and statistical application and analysis. T he results are sum m arized as follows : 1) 12.5% (63/506) had no description o f statistical m ethods which were applied, and 4.9% (25/506) described statistical m ethods which were not applied. 2) 65.6% (332/506) contained at least one o f the statistical problem s in the categories which were selected by the authors. 77.4% (154/506) for the Student s t-test, 69.7 % (101/145) for the chi-squared test, 71.6% (48/ 67) for the analysis of variance, 32.4% (23/71) for the Pearson correlation, 18.2% (2/11) for the nonparam etric correlation, 15.0% (3/20) for the nonparam etric m ethod, and 70.0% (14 /20) for the factor analysis contained statistcial problem s. It was suggested th at a basic training in statistical methods, m ore intensive consultation of investigators with psychiatric statisticians and m ore extensive statistical refereeing are needed to im prove the standard of statistical application and analysis, and prelim inary form of the guidelines for the description and application of statistcis is proposed.
서 론
연구대상 및 방법
연 구 결 과
고 찰
결 론
References
ABSTRACT