[Analysis and importance of statistical power and sample size in empirical scientific research]

In order to define completely the mining of significant test it is essential to have the knowledge of test's statistical power. We intuit that when two tests are performed it is the one with a bigger power that is preferable. It means that when two tests have the same sample size and the same significance level a, better is the one which rejects the false null hypothesis more often. When planning the empirical scientific research it is cardinal to know how big the effect must be and what level of power of statistical test has to be achieved to be satisfied with the overall outcome. Four versions of power analysis are being presented. They are accessible in Power Analysis procedure of STATISTICA software. The analysis of statistical power supports one of the most severe problems in clinical research such as exposing objects of the study to the risk of injury, unpleasantness and inconvenience during the experiment. It happens that when scientist try to draw a particular clinical conclusion too little data are taken into account and the risk of receiving false negative results is getting higher. The opposite situation appears when in order to justify the outcome of the experiment too many data are taken which leads to false positive results.