The course will provide an introduction to the issues raised by missing data and it will illustrate the shortcomings of ad-hoc methods like deletion, recoding, complete cases for 'handling' missing data. Methods like multiple imputation, inverse probability weighting and likelihood procedures for statistical analysis with missing data will be discussed and contrasted. Other topics include some accessible methods for exploring the sensitivity of inference to the missing at random assumption. Special emphasis is given to the problem of missing data into specific contexts (longitudinal studies, time-to-event studies).
- What is missing data;
- Consequences and effects of missing data;
- Missing data and inference;
- Illustration of consequences and effects with some examples.
2. Fundamental Concepts
- Missing data patterns;
- Missing data mechanisms;
- Types of missing data mechanisms;
- Illustration with some examples.
3. Introduction to missing data methodology
- Ad-Hoc Methods and why not to use them;
- Imputation-based procedures.
1. Missing Data Methodology (cont’d)
- Visualize missing Data;
- Weighting procedures;
- Model-based procedures -likelihood methods;
- Sensitivity analysis.
2. Case Studies
- Missing data in a longitudinal study;
- Missing data in a survival analysis.
Virtually anyone who does statistical analysis can benefit from this course.
- To take this course, you should have a good working knowledge of the principles and practice of multiple linear regression, as well as elementary statistical inference.
- We will use R/Rstudio packages to demonstrate implementation of certain methods; thus, some familiarity with R is desirable
• Other academics: €150
• Non-profit/social sector: €250
• Private sector: €500