Getting ontologically serious about the replication crisis in psychology.

Abstract

Discussions of the so-called replication crisis (RC) in psychology have focused on methodological–epistemological considerations. In this article, I follow a speculative approach within analytic philosophy to focus on ontological considerations, as potentially helpful to address RC. There have been some ontological discussions about RC, but they have not been serious for not paying any attention to academic analytic ontology, which has resulted in much confusion, especially the construal of spurious links to epistemology. I begin with a proposal for a novel sense in which psychological phenomena can be objectively real in the ontological sense of the term “objective,” that is to say, ontologically independent of mental subjects. This sense is an application of modal realism to psychological phenomena that preserves the subjectivity of the psychological by allowing for a relativized ontological objective–subjective distinction. I then use this proposal to ground a propensity interpretation of probability as a viable ontology for the probabilistic character of psychological results. On this interpretation, probabilities measure propensities as objectively real features of reality. I conclude that psychological phenomena are objectively probabilistic, which might contribute to RC. The main implication is that even if psychological research were methodologically perfect, there would still be substantial replication failures, just due to the probabilistic nature of psychological phenomena. Replications thus become indispensable to understand psychological phenomena, and replication failures are as much part of their nature as successes are. I discuss some limitations and future directions.