Lecture 7

Question 1

Define Preregistration

Answer 1

Declaring your study and hypothesis before collecting or analyzing data.

Question 2

What is Minimal preregistration?

Answer 2

Publicly post your hypothesis prior to data collection (at least before data analysis)
Even better: Publicly post your exact analysis plan, methods, and draft syntax prior to data collection

Question 3

What are the 2 reasons for preregstration? Who and why does it benefit?

Answer 3

Allows to differentiate confirmatory and exploratory results
Increases scientific rigor by improving the diagnosticity of statistical tests
Benefiting both:
* researchers: it certifies that the research is carried out as planned originally
* readers: it gives more confidence about statistical interpretations

Question 4

What are the two main limitations of preregristation?

Answer 4

Possibility that all the parameters are difficult to pre-specify completely. Researchers may still (knowingly or unknowingly) build undisclosed flexibility into the analysis plan
Pre-registration does nothing at all to stop outright dishonesty. Ways to intentionally cheat a pre-registration system are readily imaginable.
* multiple private pre-registrations can be made, each with a different hypothesis, without disclosing this fact
* dates can be misrepresented in order to falsely pre-register a study that was already run

Question 5

If there is a significant correlation between baseline covariates and the group, it indicates that the (?) is not done well.

Answer 5

randomization