Why Scientific Reproducibility Matters

Scientists often repeat experiments to determine whether an initial result is reproducible. Reproducibility is crucial to science, but not for the reason most people think it is. The mistaken idea that scientific experiments can “verify” theories leads people to think that successive instances of an identical result can, in effect, reverify theories (this idea is often expressed in terms of increasing a theory’s “Bayesian credence”). But nothing, not even empirical data, can verify our theories. Empirical data serves only to criticize our theories. In other words, if a scientific theory predicts that a particular observation will occur under the conditions of a specified experiment, then scientists can perform the experiment and see what they see. If the experiment, or any number of successive repetitions of it, reveals any mismatch between the predicted observation and the actual observation, then the scientists will have detected a potential error in the theory, which they can then work to correct. And crucially, detection and correction of errors in our scientific theories constitutes scientific progress. So, reproducibility matters not because repeatable results serve to verify (and reverify) whatever theory predicted those results. It matters because unrepeatable results alert us to errors in our theories, which we can attempt to correct in order to grow our scientific knowledge.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s