Quality Checks | Licklider

When to use this section

Use Quality Checks when your question is not primarily "Which method should I run?" but "What could weaken, qualify, or limit how I interpret this result?" This section helps you navigate checks around independence, assumptions, robustness, missing data, multiplicity, covariates, model diagnostics, provenance, and validity signals.

Quick self-check

Read this section if you need to decide whether a result is trustworthy enough to interpret, report, or defend.

This section does not choose the analysis method for you or design the figure itself; it helps you review assumptions, diagnostics, and documentation around an analysis you already plan to use.

Quickest route

If your main question is...	Go to...
Do I need a record of preprocessing, exclusions, outlier handling, or attrition?	Audit Trail and Provenance
Could assumptions, outliers, or reasonable analysis variations change the conclusion?	Assumptions and Robustness
Are observations really independent, or do pairing, repeated measures, or batch effects change the analysis unit?	Design and Independence
Could missing data, multiple comparisons, or researcher degrees of freedom make the result look stronger than it is?	Missingness, Multiplicity, and Degrees of Freedom
Has an adjusted analysis handled confounding variables and covariates clearly enough for interpretation?	Confounding and Covariates
Does the model match the data structure, such as regression residuals, bounded outcomes, compositional data, or survival data?	Model Diagnostics
Do I need a compact summary of statistical validity signals before making a claim?	Validity Score

Start here if you are unsure

Start with Design and Independence if you are not sure what counts as an independent observation.
Start with Assumptions and Robustness if the method is chosen but you are worried the conclusion may be fragile.
Start with Audit Trail and Provenance if the main problem is documenting preprocessing, exclusions, or sample attrition.

If this is not your question

If you still need to choose a statistical test, model, or effect size, go to /docs/methods.
If you already know the analysis and need to choose a chart such as a box plot, scatter plot, Kaplan-Meier plot, or diagnostic plot, go to /docs/figures-and-visualization.

Child pages