The number of methods available for MRI analysis is growing every year. Each of these methods requires the specification of a bewildering array of parameters, not all of which are amenable to optimisation by consensus. The combination of these factors leads to an almost infinite number of ways in which we may analyse our data, and accordingly an infinite number of results which we may choose to report, or not. In this talk, I will discuss concerns of such analytic flexibility and the draw of questionable research practices to help uncover the “publishable story” in our data. I will present recent developments in our discipline which may assist the identification of robust effects, and what individual actions we may take in order to retain confidence in our own findings.