Frederick Eberhardt's research focuses on causal and statistical inference with a particular emphasis on methods of discovery using experiments: Under what circumstances can causal relations be identified and what are the optimal discovery strategies to do so? The question has a normative and a descriptive side: On the normative side the aim is to develop inference techniques for causal discovery that are optimal and reliable, and to analyze the underlying assumptions. On the descriptive side, humans (and animals?) appear to learn causal relations all the time. Which strategies are employed, and how does this learning compare to a normative theory?

More broadly, Frederick's interests include the foundations of probability, rationality, optimization, game and decision theory, graphical models, inference techniques in machine learning, methodological issues in the philosophy of science and the work of the philosopher Hans Reichenbach.