the.com/missingdata
the gaps in your dataset that lie by omission, not commission.
means missingdata refers to values absent from a dataset where a value should exist, and how you handle it can quietly rewrite your conclusions.
from statisticians formalized the concept in the 1970s, notably donald rubin's 1976 taxonomy splitting it into mcar, mar, and mnar—shorthand for how random the vanishing act really is.
three typesmcar, mar, mnar sound like alphabet soup
deletion riskdropping rows can silently bias your entire analysis
imputation guessworkfilling gaps means inventing data you didn't collect