ABSTRACT
A systematic treatment of the core issue emphasized in the title and content of this contribution requires a short clarification regarding the use of the term “causal” in our discussion. It seems important to differentiate its meaning in the philosophy of science from that typically implied in the social sciences, for instance, psychology and education. Whereas in the philosophy of science the term “causal” refers to deterministic relationships between variables (“whenever X, then Y”), the term “causal” as used in the social sciences refers to a probabilistic relationship between variables, indicating that individual differences in a specific variable X show a statistically significant impact on individual differences in another variable Y, which is typically assessed at some later time. Based on this definition, it is possible to talk about “multicausality”, meaning that several variables significantly predict a given criterion variable, and also about “causal predominance” whenever it appears that one predictor variable has a substantially stronger impact on a specific criterion than other predictor variables. For instance, multivariate statistical tools such as structural equation modeling (causal modeling) represent tests of causality assumptions based on the probabilistic approach.
