Vershinina E A, Safarova G L
I.P.Pavlov Institute of Physiology of PAS, 6 Makarova emb., St. Petersburg 199034, Russian Federation, e mail:
Institute for Regional Economic Studies, RAS, 38 Serpukhovskaya str., St. Petersburg 190013, Russian Federation, e mail:
Adv Gerontol. 2019;32(6):1052-1062.
In the paper a number of issues, including conceptual ones, related to the specific features of making statistical decisions, choosing tests and computing characteristics that will strengthen the evidence base of the conclusions obtained when analyzing data using methods of mathematical statistics, are considered. The paper aims not to describe the methods of mathematical statistics themselves, but to analyze the conditions and the need to apply the most common tests. In particular, the magnitude of the indicator of the statistical significance of the observed effects - p-value - and the sample size to obtain a significant effect are discussed, the effect of multiple comparisons, the application of the Bayesian approach, and others are considered.
