Conducting reliable and credible evaluation and statistical interpretation of empirical results related to the operation of production systems

in foundries is for most managers complicated and labour-intensive. Additionally, in many cases, statistical evaluation is either ignored

and considered a necessary evil, or is completely useless because of improper selection of methods and subsequent misinterpretation of the

results. In this article, after discussing the key elements necessary for the proper selection of statistical methods, a wide spectrum of these

methods has been presented, including regression analysis, uni- and multivariate correlation, one-way analysis of variance for factorial

designs, and selected forecasting methods. Each statistical method has been illustrated with numerous examples related to the foundry

practice.