Why theory matters: New arguments for an old conclusion

Theories are what we accept or reject in the game of science; what we use to explain and predict phenomena, observations, and measurements; and what we compare to evaluate progress and stagnation. Yet, in recent years, the motivation for theory has become uncertain and unclear, particularly in fields that rely heavily on observation or experimentation (e.g., psychology) or that are becoming increasingly data-driven (e.g., neuroscience). Traditional arguments for theory may be rationally convincing but fail to change day-to-day research practices in these fields. A new argument for theory is proposed, starting from an analysis of the so-called ‘replication crisis’ and a model of the interplay between theory development and theory testing in psychology. For empirically-minded researchers in these fields, the value of theory development may lie in restricting the space of possible hypotheses and models to those that have greater a priori plausibility, thus reducing the risks (and costs) of pursuing false or even impossible empirical claims. However, the motivation for theory needs to be rediscovered and reconstructed anew as the context of a particular research field shifts, and traditional arguments should not be invoked uncritically if they are to retain their power to influence beliefs and practices.