book

Statistical Analysis with Missing Data., 3rd Edition

by Roderick J. A. Little, Donald B. Rubin

April 2019

Intermediate to advanced

464 pages

13h 23m

English

Wiley

Read now

Unlock full access

1.1 The Problem of Missing Data1.2 Missingness Patterns and Mechanisms1.3 Mechanisms That Lead to Missing Data1.4 A Taxonomy of Missing Data MethodsProblemsNote
2.1 Introduction2.2 The Exact Least Squares Solution with Complete Data2.3 The Correct Least Squares Analysis with Missing Data2.4 Filling in Least Squares Estimates2.5 Bartlett's ANCOVA Method2.6 Least Squares Estimates of Missing Values by ANCOVA Using Only Complete-Data Methods2.7 Correct Least Squares Estimates of Standard Errors and One Degree of Freedom Sums of Squares2.8 Correct Least-Squares Sums of Squares with More Than One Degree of FreedomProblems
3.1 Introduction3.2 Complete-Case Analysis3.3 Weighted Complete-Case Analysis3.4 Available-Case AnalysisProblems
4.1 Introduction4.2 Imputing Means from a Predictive Distribution4.3 Imputing Draws from a Predictive Distribution4.4 ConclusionProblems
5.1 Introduction5.2 Imputation Methods that Provide Valid Standard Errors from a Single Filled-in Data Set5.3 Standard Errors for Imputed Data by Resampling5.4 Introduction to Multiple Imputation5.5 Comparison of Resampling Methods and Multiple ImputationProblems
6.1 Review of Likelihood-Based Estimation for Complete Data6.2 Likelihood-Based Inference with Incomplete Data6.3 A Generally Flawed Alternative to Maximum Likelihood: Maximizing over the Parameters and the Missing Data6.4 Likelihood Theory for Coarsened DataProblemsNotes

7.1 Introduction7.2 Bivariate Normal Data with One Variable Subject to Missingness: ML Estimation7.3 Bivariate Normal Monotone Data: Small-Sample Inference7.4 Monotone Missingness with More Than Two Variables7.5 Factored Likelihoods for Special Nonmonotone PatternsProblems
8.1 Alternative Computational Strategies8.2 Introduction to the EM Algorithm8.3 The E Step and The M Step of EM8.4 Theory of the EM Algorithm8.5 Extensions of EM8.6 Hybrid Maximization MethodsProblems
9.1 Standard Errors Based on The Information Matrix9.2 Standard Errors via Other MethodsProblems
10.1 Bayesian Iterative Simulation Methods10.2 Multiple ImputationProblemsNotes
11.1 Introduction11.2 Inference for a Mean Vector and Covariance Matrix with Missing Data Under Normality11.3 The Normal Model with a Restricted Covariance Matrix11.4 Multiple Linear Regression11.5 A General Repeated-Measures Model with Missing Data11.6 Time Series Models11.7 Measurement Error Formulated as Missing DataProblems
12.1 Introduction12.2 Reducing the Influence of Outliers by Replacing the Normal Distribution by a Longer-Tailed Distribution12.3 Penalized Spline of Propensity PredictionProblemsNotes
13.1 Introduction13.2 Factored Likelihoods for Monotone Multinomial Data13.3 ML and Bayes Estimation for Multinomial Samples with General Patterns of Missingness13.4 Loglinear Models for Partially Classified Contingency TablesProblems
14.1 Introduction14.2 The General Location Model14.3 The General Location Model with Parameter Constraints14.4 Regression Problems Involving Mixtures of Continuous and Categorical Variables14.5 Further Extensions of the General Location ModelProblems
15.1 Introduction15.2 Models with Known MNAR Missingness Mechanisms: Grouped and Rounded Data15.3 Normal Models for MNAR Missing Data15.4 Other Models and Methods for MNAR Missing DataProblems