Preregistration

This is a very important step for ensuring reproducibility of your research findings and to protect against p-hacking. We encourage all students to make a pregistration for their study.

• You can make a preregistration on AsPredicted - they allow you to make a test preregistration, check it out.

What about pilots?

• Always run a test pilot before preregistration of anything (N < 20). Preregister for either your full study or pilot afterwards.

  • This data may not be useful for publications.
  • If the initial pilot has interesting data that you want to include in a study, mention that it is not pre-registered.

Types of Preregistration

• Exploratory pre-registrations: can be useful but you can over constrain yourself. It’s okay to have very minimally constraining ones.
• Directional pre-registrations: These can be more robust preregistrations where you commit to your hypotheses.
• Read Daniel Lakens (2024) for more information on when to deviate from your pre-registration.

Considerations

• You should have thought about the following – or have some idea of them – by the time you make your preregistration:

  • what kind of test you will run
  • effect size
  • number of conditions

💡 Tips

• When preregistering a sample size, consider adding a 10-15% buffer in anticipation of exclusions to meet your sample size.

  • Here’s an example of this: “Anticipating exclusions, we’ll recruit 400 (or 100 participants per condition).”

• Be detailed - for example, when describing your DVs, include the full items and rating scale.
• Don’t forget to set your alpha - we conventionally set this to .05 for Null Hypothesis Significance Testing (NHST).