Potencjał współczesnego wnioskowania statystycznego. W setną rocznicę Statistical Methods for Research Workers
The potential of contemporary statistical inference. On the centenary of Statistical Methods for Research Workers

Summary/Abstract: Although a hundred years have passed since the publication of Ronald Fisher’s groundbreaking book titled Statistical Methods for Research Workers, the author remains – alongside Jerzy Neyman and Egon Pearson – a leading architect of the theoretical foundations of statistical inference. Statistical methods have been systematically enhanced throughout the past ten decades, but over the recent dozen or so years, many new challenges have emerged. Larger and more easily accessible data sets in addition to big data and AI analytical tools have been creating both opportunities and challenges for statisticians. The aim of this paper is to identify and characterise the activity that should be taken by statisticians in order to, on the one hand, take a fuller advantage of the potential of statistical inference in the era of big data and AI, and on the other, persuade more efficiently the users of statistical methods that there are permanent elements of uncertainty in statistical inference. The authors demonstrate that elements of uncertainty, of which some users of statistical methods are unaware, are inherent not only in data (that reflects the volatility of the studied phenomena), but also in the inference model itself. They conclude, among other things, that the discussion ongoing in the global scientific community about the dilemmas connected to statistical significance is the consequence of the practical applications of statistical inference taking into account the model assumptions to an insufficient degree. It also results from attributing a decisive significance to a single sample study in the context of universal scientific research objectives. The authors propose the application of the Bayesian approach to testing statistical hypotheses, viewing it as complementary to the classical approach co-created by Fisher.

• Details
• Contents