EasyDeepLearn

MCAR, MAR, MNAR — what are these and why do they matter?

medium

Answer

  • Missing Completely At Random (MCAR): missingness independent of everything — dropping rows is unbiased.
  • Missing At Random (MAR): missingness depends only on observed features — imputation with these features is unbiased.
  • Missing Not At Random (MNAR): missingness depends on the unobserved value itself (e.g., high earners refuse to disclose income) — imputation from observed features is biased; you need external information or explicit modeling.
  • Test empirically: does missingness correlate with observed features?
Check yourself — multiple choice
  • They're all synonyms
  • MCAR: random; MAR: depends on observed features (imputable); MNAR: depends on the missing value itself (biased)
  • MNAR is easier to handle than MCAR
  • Only MCAR affects predictions

MCAR = random; MAR = observed-dependent (imputable); MNAR = missing-dependent (biased without external info).

#missing-data#data-quality

Practise Supervised Learning

215 interview questions in this topic.

Related questions