Modern Missing Data Analysis

Replay Access: Expires 1 year after purchase
Original Livestream: June 1-3, 2022
Instructor: Craig Enders

Student: $635
Professional: $795

Registration Type

Clear

Missing data are a ubiquitous feature of nearly all research applications, arising through participant non-response, attrition, and sometimes even by design. Failure to account appropriately for missing values when conducting statistical analyses can result in badly biased estimates and incorrect inferences about the relationships under study. Modern Missing Data Analysis is a three-day workshop focused on three optimal approaches for addressing missing data that can be applied across a variety of research settings: maximum likelihood, Bayesian estimation, and multiple imputation. Participants will learn details about the classic versions of these methods as well as state-of-the-art extensions that have been developed in the last five years. These procedures are advantageous because they use all available data and make realistic assumptions about the cause of missingness; estimates and significance tests are therefore valid in a broader range of situations than historical methods such as deleting incomplete data records. The purpose of this course is to provide participants with foundational knowledge about maximum likelihood, Bayesian estimation, and multiple imputation. Equally important, participants will learn how to select an appropriate missing data handling method and apply it to complex, real-world data sets. To this end, presentations will include a mix of theoretical information, practical tips, and computer applications.

Instructor

Craig K. Enders, PhD

Craig Enders is Professor and Area Chair of Quantitative Psychology in the Department of Psychology at the University of California, Los Angeles. His primary research focus is on analytic issues related to missing data analyses, and he leads the research team responsible for developing the Blimp statistical software application. Read More

Workshop Details

Syllabus

Chapter 1: Introduction to Missing Data
1.1 Missing Data Mechanisms
1.2 Preparing for a Missing Data Analysis
1.3 Selecting Auxiliary Variables

Chapter 2: Historical Tour (Self-Study)
2.1 Listwise and Pairwise Deletion
2.2 Mean Imputation
2.3 Regression Imputation
2.4 Last Observation Carried Forward

Chapter 3: Preview of Modern Missing Data Strategies
3.1 Probability Distributions and Factoring Distributions
3.2 Missing Data Strategies: Marginalizing Over Missing Data Versus Imputing
3.3 Choosing a Missing Data Method

Chapter 4: Full Information Maximum Likelihood (FIML) Estimation for Multivariate Normal Data
4.1 Classic FIML Estimation in Structural Equation Modeling Framework
4.2 Corrective Procedures for Nonnormal Data
4.3 Incorporating Auxiliary Variables in a FIML Analysis

Chapter 5: Factored Regression Specifications
5.1 Factorizing a Multivariate Distribution into Univariate Distributions
5.2 FIML Estimation with Factored Likelihoods
5.3 FIML with Mixtures of Categorical and Continuous Variables

Chapter 6: Bayesian Estimation
6.1 Bayesian Versus Frequentist Paradigm
6.2 MCMC Estimation
6.3 Missing Data Imputation
6.4 Convergence Diagnostics

Chapter 7: Bayesian Estimation With Categorical Variables
7.1 Latent Response Framework
7.2 Binary, Ordinal, Nominal Variables

Chapter 8: Multiple Imputation
8.1 Agnostic Versus Model-Based Multiple Imputation
8.2 Fully Conditional Specification
8.3 Rubin’s Rules
8.4 Model-Based Multiple Imputation

Chapter 9: Missing Not at Random Processes
9.1 Selection Models
9.2 Pattern Mixture Models
9.3 Sensitivity Analyses

Chapter 10: Special Topics and Applications (Self-Study)
10.1 Scale Scores with Missing Data
10.2 Incomplete Scale Scores and Interaction Effects
10.3 Latent Variable Interactions
10.4 Nonnormal Predictor Variables
10.5 Nonnormal Outcome Variables

Who Should Attend

The workshop is designed for graduate students, post-doctoral fellows, faculty, and research scientists from the behavioral, social, and health sciences. Attendees should have a working knowledge of multiple regression, as statistical aspects of the workshop borrow heavily from these models. Participants who would like a refresher on linear regression prior to attending may wish to watch our (no cost) Linear Regression Playlist on YouTube.

Software Considerations

Software demonstrations will be presented throughout the day. Maximum likelihood estimation will be conducted using Mplus as well as the lavaan and mdmb packages in R. Note that R can be downloaded for free, and all Mplus examples will run in the free demo version of the software, available at www.statmodel.com. Bayesian estimation and multiple imputation will be conducted in Blimp. Blimp can be downloaded for free at www.appliedmissingdata.com/blimp. While it is helpful to have some familiarity with one or more of these programs, prior experience is not necessary. The accompanying software demonstration notes provide detailed instructions in how to use the different packages and data files and script are provided for all worked examples.

The Goals of the Workshop

Our motivating goal is to provide an interactive and engaging instructional experience that provides attendees with foundational knowledge about state-of-the-art missing data methods. Instructional time features an equal balance of core statistical concepts and practical data analysis applications. Statistical concepts are presented at a level that is fully accessible to researchers with a background in multiple regression. The workshop is designed to provide participants with the materials and instruction needed to understand core concepts; choose an appropriate missing data handling procedure; apply that procedure to complex, real-world data sets; and describe the methodology and results. Attendees should leave the course with a toolkit that fully prepares them to utilize cutting-edge missing data methods in their own work.

What is Provided

Availability of Recordings

Sample Content

Tuition Support for Students

We offer reduced-price registrations for undergraduate and graduate students who are actively enrolled in a recognized bachelor's, master's or doctoral training program. No application is necessary to qualify for the student tuition rates; simply choose the student rate when beginning the registration process at the top of the page. Confirmation of student status may be requested at a later time.

Cancellation Policy

Reviews

In an effort to continually improve our instruction we obtain student evaluations with each course offering. Here is a sample of reviews from prior workshops offered by Dr. Enders:

Craig is easily one of the best statistics professors I have ever had. He has taken an incredibly complex topic that I have previously struggled to understand and apply to my own data and made it possible for me to feel confident in using and interpreting the analyses on my own

Dr. Enders did a great job of explaining concepts in multiple ways when prompted to do so by student questions, and creating an atmosphere where students felt comfortable asking questions -- Dr. Enders welcomed both basic questions and those that were beyond the scope of the course but related to a student's own research interests.

I felt the slides were quite easy to understand and appreciated that there were figures which helped to illustrate conceptual understanding underlying each concept.

I benefited so much from this class, both in learning the material, and in the professor's example for effective pedagogy.

Prof Enders is a great lecturer. He puts complicated stats concepts in terms that are easy to understand, knows when to pause to let students process, and knows when to go over something multiple times/in multiple ways to maximize student understanding.

I really appreciated this class because it taught me practical skills of how to do statistics in my own research, but also helped me to concretely understand important concepts.

Quick Navigation

Self-Paced Workshops

View Upcoming Livestream Workshops

Modern Missing Data Analysis

Instructor

Craig K. Enders, PhD

Workshop Details

Reviews

Quick Navigation

Self-Paced Workshops

Free Introduction to Structural Equation Modeling

Mixture Modeling and Latent Class Analysis

Applied Measurement Modeling

Causal Inference

Multilevel Modeling

Applied Qualitative Research

Network Analysis

Applied Research Design Using Mixed Methods

Introduction to Data Visualization in R

Longitudinal Structural Equation Modeling

Introduction to Quantitative Meta-Analysis

Analyzing Intensive Longitudinal Data

Machine Learning: Theory and Applications