Testing Dependences; Linear Regression; Effect Size; Educational Research; Big Data Set

In accordance with quantitative research, a wide spectrum of techniques can be seen. In the case of cardinal variables, regression analyses are suitable tools for the expression of dependences between observed variables. One of their options, the regression coefficients have been considered. However, the effect sizes analyses have not been so widely seen in research works in general. In this contribution, the big data analysis is being presented focusing on the regression effect size behavior following changing the number of samples. Two-dimensional and three-dimensional computations are applied with utilized mathematical regression models. The stable or stochastic behavior of Cohens f squared is discussed in the particular applied quantitative research of the OECD PISA with 397708 answers from respondents.