|
|
MISSING VALUES ANALYSIS & DATA IMPUTATION
Overview
An illustrated tutorial and introduction to missing values analysis and data imputtion using SPSS, SAS, and Stata. Suitable for introductory graduate-level study.
The 2015 edition is a major update to the 2012 edition. Among the new features are these:
- Was 40 pages with 25 figures, now 113 pages with 51 figures
- Now covers Stata and SAS as well as SPSS
- Totally revised throughout
- New coverage comparing MI and EM estimation methods
- New coverage on testing for MI convergence
- New coverage on exploratory testing for missing at random
- New section on when MI should not be used
- Twelve FAQ sections
- Links to all datasets used in the text.
The full content is now available from Statistical Associates Publishers. Click here.
Below is the unformatted table of contents.
MISSING VALUES ANALYSIS AND DATA IMPUTATION Overview 6 SPSS 6 SAS 7 Stata 8 Data examples in this volume 8 Key Concepts and Terms 9 Causes of non-response 9 Item non-response 9 Listwise deletion of cases with missing values 10 Types of Missingness 11 Missing completely at random (MCAR) 11 Missing at random (MAR) 15 Missing not at random (MNAR) 15 Multiple imputation 16 When imputation should not be used 16 Summary 17 Testing for missing at random (MAR) 18 Overview 18 Testing for MAR in SPSS 21 Testing for MAR in SAS 24 Testing for MAR in Stata 27 Single versus multiple imputation 30 MI estimation and monotonicity 31 The imputation model assumption 34 Number of imputations 35 Multiple imputation in SPSS 35 Overview 35 How MI works 36 What MI does 36 Pooled estimates 37 SPSS input 38 Overview 40 The Method tab 41 The Constraints tab 42 The Output tab 44 Checking for convergence 45 SPSS output 46 The "Imputation Models" table 46 The "Descriptive Statistics" tables 47 A logistic regression example 48 Pooling diagnostics 51 Multiple imputation SAS 52 Overview 52 SAS input 52 Checking for convergence 53 SAS output 54 Multiple imputation Stata 56 Overview 56 Stata input 57 Initial assessment of missingness 57 Preparing for data imputation 59 Data imputation 60 Checking for convergence 62 Running statistical procedures on imputed data 63 Stata output 64 Single imputation of missing values 66 Mean imputation 66 Other simple replacement methods 66 SPSS 66 SAS 67 Stata 69 The hot deck method of data imputation 69 SPSS 70 SAS 70 Stata 70 Missing Value Analysis (MVA) in SPSS 70 Overview 70 MVA set-up in SPSS 72 Types of estimation 73 The variables button 75 The patterns button 76 The descriptives button 82 Other MVA output 82 Default output 82 The percent mismatch table 83 Output for t tests 84 Crosstabulation 87 Expectation maximization (EM) estimates 88 Saving EM-imputed data 90 Assumptions 90 Multivariate normality 90 Frequently Asked Questions 91 Why not just delete cases with missing values rather than impute values at all? 91 Is it permitted to impute the dependent variable? 91 Do I need a large sample to do MI? 91 Should I round my MI estimates? 91 MI versus EM or FIML estimation 92 SPSS 95 SAS 95 Stata 95 Can MI be used with hierarchical data? 96 Should I use original data or imputed data when reporting results? 96 In SPSS, which procedures support pooling of MI estimates? 97 Can I use multiple imputation with complex survey data? 101 What is Heckman's correction for sample selection bias? 101 What is approximate Bayesian bootstrapping? 102 How can I identify missing value patterns in SAS? 103 How can I identify missing value patterns in Stata? 105 How can I restrict the bounds of imputed values in Stata? 107 Acknowledgments 107 Bibliography 107 Pagecount: 113