Why we don’t need to be afraid that preregistration prevents creativity

One argument that often comes up when I talk to skeptics of preregistration is that it stands in the way of creative and exciting research. I couldn’t disagree more. Preregistration and registered reports are among the very best developments that have come out of Psychology’s replication crisis. Both guide a way towards better research. But since the sentiment that preregistration and creativity are not compatible is so prevalent and seems so genuine (as opposed to being an excuse to engage in questionable research practices), I do want to expand on some main reasons why preregistration does not dampen creativity in research.

Reason 1: Eureka precedes preregistration

Hardly any preregistration fiend would urge anyone to open their laptop and create a preregistration document whilst they are sitting leaned against an apple tree and pondering new research ideas, or – even more inconveniently given the often not waterproof nature of digital devices – when inspiration strikes them while singing under the shower. Preregistration happens much later: After the initial Geistesblitz, after reading tons of literature, after conceptualizing the developing idea with others, after trying to think of a good way to test it experimentally in lab meeting, after running some pilots and testing out parameters as long as you trust yourself for them to strictly leave them pilots. Eureka precedes preregistration.

Reason 2: Preregistration is open to postcorrection

Still, I suspect that writing down the study design and analysis plan makes people feel restricted. What if you start testing and find that the timing was off or one stimulus not quite right? Assuming a study was piloted and is well thought through, the likelihood of that happening halfway through should be rather low. In fact, I think preregistration even serves a good purpose by forcing people to think through their design more, and thus reduce the likelihood of oversight. But of course, oversight always happens. You can just go back to your preregistration and add a time-stamped passage explaining why you are now changing the design, and start anew. Similarly, what if you look at the data and see, for instance, that the time-window you preregistered for the analysis of your eye-tracking data was totally off? Or you see a pattern in the data you had not thought of before and that you would want to explore and report? This is also no big deal. You report your planned analyses in your manuscript, and add reasons for why they were off or why you want to explore further. If your initial thought and your adjustments and additions are reasonable, rarely a reviewer should object the changes. But, readers can evaluate the findings as confirmatory or exploratory.

If you want to start preregistering, you can read more about the basics of it in this post by our very own Christina, and a very detailed introduction over at data colada.

And remember: The only creativity preregistration goes against is creativity in data analysis, and in hypothesizing after results are known.