Przekształcenia normalizacyjne danych w procedowaniu metod statystycznych wykorzystywanych w badaniach nad edukacją

Abstract

Introduction : Sometimes the mapping of the variable values made during the measurement needs to be transformed due to the conditions related to the data analysis. Research Aim : The aim of this article is to present the issue of normalization transformation and the circumstances of their application. Evidence-based Facts : Due to lack of knowledge or experience, transformations are not taken into account in data analysis in education research. And if they do occur, among the most popular methods of data transformation, are chosen those which reduce the skewness of distributions. However, the method of transformation is not always chosen adequately to the properties of the data and the conditions of analysis. Normalization transformations are among the simpler yet effective solutions for preparing data for analysis. Summary : Normalization transformations minimize the risk of artifacts due to differences in orders of magnitude and units of measurement. This is particularly important when conducting analyses using multidimensional scaling and multivariate classification methods.