When to use this
Use this before descriptive, inferential, or measurement analysis.
Why it matters
Missingness can change effective sample size and may introduce bias depending on its cause and handling.
Step-by-step instructions
1. Open Data Quality Lab for the active dataset version.
2. Review variable-level missing counts.
3. Inspect patterns and distinguish planned skips from nonresponse where possible.
4. Document the intended strategy.
5. Preview any transformation before approval.
6. Keep the immutable source version for comparison.
Expected outputs
Missingness diagnostics, a documented decision, and—only if approved—a new working version.
How to interpret the results
Percentages describe observed absence; they do not establish why values are missing. Interpret patterns alongside the survey design.
Common mistakes
Do not automatically delete incomplete cases or impute values without a defensible rationale. Do not overwrite the source version.
Was this article helpful?
Your response helps the methodology team improve this documentation.