Analysis Planning
Use this section to declare analysis-direction, exclusion, and design decisions before running a confirmatory analysis.
Use Analysis Planning when your question is what should be decided and documented before you run a confirmatory analysis.
This section is for planning choices such as analysis family, hypothesis direction, outlier handling, inclusion rules, randomization, blinding, and stopping rules before results are interpreted. It is not the place to define the structure of the dataset, troubleshoot upload, or read detailed method guidance.
What this section helps you decide
Use this section when your immediate question is one of these:
- Which analysis family or comparison set should be treated as one planned family?
- Should the hypothesis be two-sided or one-sided?
- How should outliers or exclusions be handled before looking at the results?
- Which samples qualify for inclusion, and what design safeguards such as randomization, blinding, or stopping rules should be declared?
These are planning decisions made before interpretation. The goal is to state the rules in advance, not to choose them after seeing the result.
Read the page that matches your planning question
| If you need to decide... | Read this page |
|---|---|
| How analyses should be grouped into a planned analysis family | Analysis Family Declaration |
| Whether the hypothesis is directional and when one-sided vs two-sided testing is justified | Hypothesis Direction and Sidedness |
| How outliers, exclusions, winsorization, or manual overrides should be handled in advance | Outlier and Exclusion Policy |
| Which participants, samples, or rows meet the inclusion rules before analysis | Inclusion and Exclusion Criteria |
| How allocation should be randomized before data collection or analysis review | Randomization Plan |
| Who should be blinded to condition or outcome information during the study | Blinding Plan |
| When a study or data collection process should stop according to predeclared rules | Predefined Stopping Rules |
Start here
If your first question is how multiple comparisons or related tests should be grouped before inference, start with Analysis Family Declaration.
If you already know the core comparison and need to decide whether the hypothesis is directional, continue to Hypothesis Direction and Sidedness.
If the main risk is post-hoc removal, retention, or winsorization of extreme observations, go to Outlier and Exclusion Policy.
If your study depends on predefined sample eligibility, masking, or stopping conditions, continue to Inclusion and Exclusion Criteria, Blinding Plan, or Predefined Stopping Rules.
What this section does not cover
- Defining what each row, ID, batch, or timepoint means in the dataset -> see Data Contract
- Checking file shape, required columns, or import errors -> see Data Requirements
- Choosing the detailed statistical method or model -> see Methods
- Reviewing warnings and diagnostics after the setup is declared -> see Quality Checks
Related
- Related links are registered in frontmatter and rendered below this page.