n≥30 vs. n=all: Büyük Veri, Veri Obezitesi ve Kaybolan Nedensellikler

Economists have long been criticized based on the fact that the assumptions of the theoretical models in economics are not always realistic. Big data analysis can be an opportunity for economists to fix this problem. I disagree. While big data analysts, to a large extent, need all available data (i.e. n=all) for the models to reach useful conclusions, I argue, they are not always interested in explaining causes and causal relations among variables. Instead, big data analysts strive for revealing correlations among events. Big data models that require n=all to produce useful conclusions lead to data obesity in economics where researchers turn into blind empiricists who look for correlations, rather than causations, within massive data sets. As a result of big data analyses, causes and causal effects may soon disappear from the scientific discourse. Big data analyses can cause economics to replace why-questions with what questions.