Conscious and Correct Use of Biostatistical Methods in Medical Researches: From Planning to Reporting the Results - Part II

In this part of the review, statistical tests utilized to examine hypotheses regarding population parameters on a representative sample, which forms the fundamentals of inferential statistics are discussed. The selection of an appropriate statistical test by verifying its assumptions and interpreting the results objectively is crucial for obtaining accurate conclusions. Understanding the terms related to type I and type II errors, p-value, power of the study, effect size, and confidence interval will contribute to the correct interpretation of both the results obtained from statistical tests in scientific research and the findings of articles read from the literature. In addition to univariate tests, the three most commonly employed multiple regression models are also addressed to control for the effect of potential confounding factors and other independent variables utilized in the study. Statistical computing has become much more accessible in recent times, with researchers having access to freeware packages or web applications to perform basic and advanced statistical analyses. Researchers frequently focus on the calculation of statistical tests used in data analysis, whereas, understanding the rationale behind statistical methods should be the primary goal. Therefore, this review emphasizes the logic of selecting appropriate statistical methods and interpreting the results rather than mathematical calculations. It is essential to recognize that biostatistical principles should be considered not only in the data analysis phase but also in all phases of research, from planning to report writing. It should be note that, no statistical analysis method can correct erroneous data obtained from a poorly designed study.