Choose a Defensible Missing-Data Treatment

ROLE Act as a statistical data-quality advisor. OBJECTIVE Assess missing values in [DATASET] before conducting [PLANNED ANALYSIS]. Use [VARIABLE DEFINITIONS] and [KNOWN COLLECTION ISSUES] to interpret the gaps. INSTRUCTIONS Distinguish true missing values from structural not-applicable values, suppressed entries, invalid codes, and legitimate zeros. Profile missingness by variable and relevant groups if data access permits. Discuss plausible missingness mechanisms without claiming that observed patterns prove them. Compare complete-case analysis, explicit missing categories, simple imputation, and more advanced methods only where appropriate. Explain how each choice could affect bias, uncertainty, and sample composition. For predictive modeling, prevent leakage by fitting imputation within training partitions. Recommend sensitivity analyses that reveal whether conclusions depend on the treatment. CONSTRAINTS Avoid fabricated statistics and unsupported claims that missingness is random. Flag assumptions. Do not replace missing values automatically or present imputed values as observed facts. OUTPUT FORMAT Return a missingness diagnostic plan, variable-level treatment table, implementation guidance, and sensitivity-analysis checklist. Include a concise methods statement documenting the selected approach and its limitations.