This study was carried out to investigate what types of assumption and conditions are needed for the application of various statistical techniques such as descriptive statistics, t-test, analysis of variance, correlation analysis, regression analysis and chi-square test, and to evaluate that they are used correctly in the rescarch process.
For study materials, 166 original papers using those statistical methods were selected from the journal of Nurses Academic Society published from January 1986 to December 1995. Descriptive statistics were used in 90 papers of the total papers. Also t-test, analysis of variance, correlation analysis, regression analysis and chi-square test were used in 113, 91, 78, 23, 43 papers, respectively. They were assessed by the author-made check list consisting of 34 items (3 items for validity assessment of descriptive statistics, 6 items for t-test, 7 items for analysis of variance, correlation analysis and regression analysis, respectively, 4 items for chi-square test).
Results showed that only 23 papers had a valid description about data among 90 papers which used descriptive statistics. The statistical sumptions, which are linearity, normal distribution, homogeneity of variance, independence of sample, were neglected in most papers using t-test, analysis of variance, correlation analysis, and regression analysis. 18 of 43 papers did not consider small sample size and less expected value. They had some mistakes in inference and interoretation of statistical results in 36 of 113 papers using t-test, 54 of 91 papers using analysis of variance, 57 of 78 papers using correlation analysis and 22 of 23 papers using regression analysis
It is suggested that most of the statistical misuse are due to the lack of the elementary statistical knowledge and the indifference about the statistical misuses committed in the papers. It is another reason to use the forms and the expressions of the pre-represented papers habitually without any criticism or consideration. And the editorial staffs have not staffs have not reviewed carefully before accepting a manuscript.
To reduce the statistical misuses, it is recommended to consult with the statisticians or other experts not only from the data processing, but also from the beginning of the study plan. Also the author of an paper should try to pursue accurate knowledge of statistics and to use her knowledge properly and correctly. It is needed that a basic biostatistics training practical and helpful for nursing investigator be developed in the nursing post-graduates.
|