Select a missing-value treatment based on meaning and risk.
Missingness has types. A blank can mean not applicable, not collected yet, unknown, refused, system error, or lost in transformation. Treating all blanks as the same violates fitness for use. The cleaning move should preserve meaning: label true not-applicable values, recover values from an authoritative source, impute only when the analysis can tolerate it, and exclude only when exclusion bias is understood. The principle works because missing values carry information about the process that produced them. If senior employees are missing start dates because of a legacy migration, dropping them biases tenure downward. If contractors correctly lack employee start dates,…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in