Abstract :
Purpose: It is required to calculate sufficient number of samples while commencing a study. One of ways of calculating number of samples is to perform power analysis. The purpose of this study is to conduct the power analysis of statistical comparisons used in articles published in a nursing journal between 2009-2011 and present the results descriptively. Method: This descriptive study examined 61 research articles. Power analysis of 725 statistical tests used in these articles was performed. Power analyses were examined by Cohen according to low, moderate and high effect size and three different results were given for each test. In order to conduct the power analysis, a “publication description form”, was used to note statistical tests used in the articles, means, standard deviations, number of samples and other information. SPSS 17.0 software programs were used to perform the statistical analysis.Results: Among the articles examined, the statistical power of only one research was calculated and presented in the text. Powers of parametric tests used in articles are higher compared to non-parametric tests. The statistical powers of 46.4% of tests at moderate-effect size and 88.7% of tests at high-effect size are 0.80 and above. Conclusion: According to results of the study, it is not common in nursing studies to determine the number of samples by conducting power analysis. A greater number of samples is required for some of studies examined to reach the intended power. Thus, it is suggested for nurse academicians to determine the sample size by calculating the statistical power before the study.
