Missing Data Methods

About

This intensive summer school course on Missing Data Methods is designed for researchers, data analysts, and statisticians who encounter missing data in their work. Missing data can significantly impact the validity of research findings and statistical analyses. This course will provide a comprehensive overview of the concepts, techniques, and best practices for handling missing data. Participants will learn about different types of missing data, the assumptions underlying various methods, and how to implement these methods using the R statistical software. The course will include a mix of lectures, hands-on exercises, and case studies to guarantee a practical understanding of the topics covered.

** Prerequisites:

Basic knowledge of statistics and familiarity with the R statistical software

Day 1 — Introduction and Basic Concepts (09:00–13:00)

Foundations of Missing Data Theory and Simple Imputation
Topics Covered

** Welcome and course overview

** What is missing data and why it matters

** Types of missing data

*** Missing Completely at Random (MCAR)

*** Missing at Random (MAR)

*** Missing Not at Random (MNAR)

** Mechanisms and implications for analysis

** Introduction to simple imputation methods

*** Mean imputation

*** Hot‑deck imputation

*** Conditional mean imputation

*** Predictive mean imputation

** Strengths and limitations of simple methods

** Hands‑On Session (R)

** Implementing simple imputation techniques in R

** Exploring the impact of imputation on distributions and models

Day 2 — Advanced Imputation Methods (09:00–13:00)
Multiple Imputation, Maximum Likelihood, and Robust Approaches
Topics Covered

Multiple Imputation (MI): concepts, workflow, and assumptions

Maximum Likelihood approaches

EM algorithm for incomplete data

Model‑based imputation

Regression imputation

Stochastic regression imputation

Doubly robust methods

When and why to use advanced imputation

Hands‑On Session (R)

Implementing MI using mice and Amelia

Running EM‑based imputation

Comparing deterministic vs stochastic regression imputation

Day 3 — Practical Issues, MNAR, and Sensitivity Analysis (09:00–13:00)
Real‑World Challenges and Applied Strategies
Topics Covered

Practical challenges in applied missing‑data problems

Strategies for handling MNAR data

Sensitivity analysis frameworks

Pattern‑mixture models

Selection models

Case studies from epidemiology, social sciences, and clinical research

Interpreting results under uncertainty

Hands‑On Session (R)

Conducting sensitivity analyses

Applying MNAR‑oriented methods

Working through real‑world case studies

Day 4 — Prevention, Planning, and Study Design (09:00–13:00)
Design‑Stage Strategies and Data Collection Techniques
Topics Covered

Preventing missing data through study design

Data collection strategies to minimize missingness

Monitoring missing data during data collection

Documentation and reporting standards

Integrating missing‑data planning into research protocols

Hands‑On Session (R)

Designing a missing‑data monitoring workflow

Creating reproducible templates for reporting missingness

Day 5 — Full Workflow Implementation and Wrap‑Up (09:00–13:00)
End‑to‑End Analysis, Best Practices, and Final Integration
Topics Covered

Building a complete missing‑data workflow

Choosing the right imputation strategy for your context

Combining diagnostics, imputation, modeling, and reporting

Best practices and reproducibility

Final Hands‑On Session (R)

Implementing a full missing‑data pipeline from raw data to final model

Comparing multiple imputation strategies

Producing publication‑ready outputs

Closing

Q&A

Course feedback and certificates

Course leader

Dr. Emmanuel Abatih

Target group

Designed for researchers, data analysts, and graduate students

Course aim

Learning Outcomes:

By the end of this course, participants will be able to:

** Identify different types of missing data and understand their implications.

** Apply appropriate methods for handling missing data in different contexts.

** Utilize statistical software to implement missing data techniques.

** Evaluate the impact of missing data methods on the results of their analyses.

** Develop strategies to prevent and address missing data in research designs.

Fee info