Inference and missing data
Donald B. Rubin · Biometrika · 1976
When making sampling distribution inferences about the parameter of the data, θ, it is appropriate to ignore the process that causes missing data if the missing data are ‘missing at random’ and the observed data are ‘observed at random’, but these inferences are generally conditional on the observed pattern of missing data. When making direct-likelihood or Bayesian inferences about θ, it is appropriate to ignore the process that causes missing data if the missing data are missing at random and the parameter of the missing data process is ‘distinct’ from θ. These conditions are the weakest general conditions under which ignoring the process that causes missing data always leads to correct inferences.
Cited by 9,714 works. doi:10.1093/biomet/63.3.581
Authors
References
- Bayesian Inference in Statistical Analysis.
- Moments and Distributions of Estimates of Population Parameters from Fragmentary Samples
- Missing Values in Experiments Analysed on Automatic Computers
- Maximum Likelihood Estimates for a Multivariate Normal Distribution when Some Observations are Missing
- Missing Observations in Multivariate Statistics I. Review of the Literature
- Estimation of Parameters in the Multivariate Normal Distribution with Missing Observations
- Maximum Likelihood Estimation with Incomplete Multivariate Data