Skip to interview prep
Apexed

10. Your preregistered analysis no longer fits an unexpected data problem. What should you do?

MediumResearch questions & experimental design

Technical
Show model answer

Preserve the original plan and explain the problem, when it was discovered and why a change is justified. Choose a defensible analysis with appropriate statistical input rather than mechanically applying an unsuitable method.

Report the deviation and its consequences, distinguishing planned from exploratory results. Where useful, present sensitivity analyses or the original analysis with limitations; preregistration supports transparency but is not a reason to hide necessary scientific judgement.

What interviewers listen for

  • Preserve the original plan
  • Justify and timestamp changes
  • Label exploratory conclusions
Your preregistered analysis no longer fits an unexpected data problem. What should you do — Science & research interview question — Apexed