Aleksandra Turkiewicz (PhD, statistician and epidemiologist at Clinical Epidemiology Unit, Orthpedics, IKVL), firstname.lastname@example.org
Pär-Ola Bendahl (PhD, Associate Professor, statistician at the Department of Oncology and Pathology, IKVL), email@example.com
The course is primarily intended for doctoral students at the Medical Faculty, Lund University, with research projects were missing data exists and needs to be handled. Also post-doctoral and senior researchers are welcomed to apply, but will be charged a fee of 5,000 SEK for the course.
All participants must have a level of knowledge in applied statistics corresponding to the mandatory statistics courses (level I and II) at our faculty. Access to and substantial skills in one of the statistical software packages SPSS, STATA or R is also required. Additionally, participants must have good theoretical and practical knowledge of linear and logistic regression analysis. This means, that they need to be able to perform such analyses and report and interpret the results.
Before you can apply to this course you have to have taken Applied Statistics I and II OR our former course Statistical methods for Medical research.
Five full days: 26th to 30th November 2018, 08:00-16:30.
Part of the course time will be used for self-studies and examination. In total, 3,5 days will be in-class.
Lund - Lecture room will be announced to the students before course start.
The course is held in English and is based on the following three themes:
1. Introduction to missing data
- To identify missing data
- Potential consequences of missing data
- Mechanisms leading to missing data
- Overview of methods for handling of missing data
2. Multiple imputation (MI)
- Overview of the theory behind MI
- Method “chained equations”
- How to build an imputation model
- Analysis of multiply imputed data
- Diagnostics of MI model (model validation)
3. Reporting results from analyses involving MI
- Reporting guidelines
- Limitations of MI
This level III course in applied statistics introduces the problem of missing data and its consequences and gives a short overview of methods for handling of missing data. Focus is on a method called multiple imputation (MI). Ideas and theory behind the method will be briefly covered - the emphasis will be on the practical applications. How to build an imputation model? How to summarize and interpret the results? What model diagnostics is needed? The course is wrapped up with training in reporting of the results after MI. Practical examples include mostly regression models for cohort studies and randomized controlled trials.
Papers and lecture material will be made available on the course web site before the course starts.
This course receives financial support from EpiHealth (Epidemiology for Health) and LUPOP (Lund University Population Research Platform).