The problem of the quality of published scientific researches is often criticized because of the questionable reliability and utility of the data they contain. Research methodology is still an underestimated and neglected issue and has an impact on principles of good publication practice. This paper describes frequent difficulties in data interpretation and illustrates the basic scheme of statistical analysis of data that should be taken under consideration by researchers. Proper choice of group of patients and statistical methods to control potential confounders are also discussed in this article.