Analysis Plan

Updated 17 Sep 2026

•
What it is
- •
  a documented mapping from each research question to the data, analysis, and outputs needed to answer it
•
Contents
- •
  data requirements and preprocessing
- •
  treatment of missing values and unusual observations
- •
  comparisons, models, or qualitative analysis procedures
- •
  assumptions and their assessment
- •
  metrics, uncertainty estimates, and visualizations
- •
  criteria for interpreting results
•
Examples
- •
  classification question -> actual/predicted labels -> confusion matrix and suitable performance metrics
- •
  experience question -> interviews -> coding and thematic interpretation
- •
  algorithm comparison -> repeated matched evaluations -> paired analysis where justified
•
Transparency
- •
  distinguish planned analyses from later exploratory analyses
- •
  HARKing presents a hypothesis formed after seeing results as though it was specified beforehand
- •
  p-hacking selectively changes analyses or reporting to obtain a desired significance result
•
Related
- •
  Research Hypothesis
- •
  Unavailable reference

Linked references 7