Missing data : a gentle introduction / Patrick E. McKnight ... [and others].

Format:

Book

Contributor:

McKnight, Patrick E.

Series:

Methodology in the social sciences

Language:

English

Subjects (All):

Social sciences--Research--Methodology.

Social sciences.

Missing observations (Statistics).

Physical Description:

xv, 251 pages : illustrations ; 23 cm.

Place of Publication:

New York : Guilford Press, [2007]

Summary:

While most books on missing data focus on applying sophisticated statistical techniques to deal with the problem after it has occurred, this volume provides a methodology for the control and prevention of missing data. In clear, nontechnical language, the authors help the reader understand the different types of missing data and their implications for the reliability, validity, and generalizability of a study' s conclusions. They provide practical recommendations for designing studies that decrease the likelihood of missing data, and for addressing this important issue when reporting study results. When statistical remedies are needed--such as deletion procedures, augmentation methods, and single imputation and multiple imputation procedures--the book also explains how to make sound decisions about their use.

Contents:

1 A Gentle Introduction to Missing Data 1

The Concept of Missing Data 2

The Prevalence of Missing Data 2

Why Data Might Be Missing 5

The Impact of Missing Data 6

A Cost-Benefit Approach to Missing Data 10

Missing Data-Not Just for Statisticians Anymore 13

Purpose of This Book 14

2 Consequences of Missing Data 17

Three General Consequences of Missing Data 18

Consequences of Missing Data on Construct Validity 20

Consequences of Missing Data on Internal Validity 25

Consequences for Causal Generalization 35

3 Classifying Missing Data 40

"The Silence That Betokens" 41

The Current Classification System: Mechanisms of Missing Data 42

Expanding the Classification System 52

4 Preventing Missing Data by Design 65

Overall Study Design 66

Characteristics of the Target Population and the Sample 72

Data Collection Methods 75

Treatment Implementation 84

Data Entry Process 85

5 Diagnostic Procedures 88

Traditional Diagnostics 89

Dummy Coding Missing Data 90

Numerical Diagnostic Procedures 92

Diagnostic Procedures Using Graphing 110

6 The Selection of Data Analytic Procedures 116

Preliminary Steps 117

Decision Making 127

7 Data Deletion Methods for Handling Missing Data 140

Data Sets 140

Complete Case Method 141

Available Case Method 147

Available Item Method 148

Individual Growth Curve Analysis 152

Multisample Analyses 156

8 Data Augmentation Procedures 160

Model-Based Procedures 161

Markov Chain Monte Carlo 166

Adjustment Methods 168

9 Single Imputation Procedures 173

Constant Replacement Methods 175

Random Value Imputation 181

Nonrandom Value Imputation: Single Condition 185

Nonrandom Value Imputation: Multiple Conditions 190

10 Multiple Imputation 196

The MI Process 199

11 Reporting Missing Data and Results 213

APA Task Force Recommendations 213

Missing Data and Study Stages 214

TFSI Recommendations and Missing Data 218

Reporting Format 221.

Notes:

Includes bibliographical references (pages 229-236) and indexes.

ISBN:

9781593853938

1593853939

9781593853945

1593853947

OCLC:

80460259

Online:

Contributor biographical information