Modern Methods in Data Analysis

Course goals

ALL PARTICIPANTS WILL BE PLACED ON THE WAITING LIST UNTIL TWO WEEKS BEFORE THE START OF THE COURSE

GRADUATE STUDENTS:
Please be aware that you can only select a course option that shows the academic year and is offered Face-to-Face (F2F)

POSTGRADUATE STUDENTS:
Please be aware that you can only select a course option that shows the academic year and is offered Face-to-Face (F2F)  or online (depending on your registration)

(FYI: the other options are options for Continuing Education (onderwijs voor professionals))


At the end of this course, the student:

1. Explain the principles of the likelihood theory and maximum likelihood methods;
2. Explain the principles of the following statistical analysis techniques: linear models (ANOVA, (multiple) regression and ANCOVA), Logistic regression analysis, Poisson regression analysis, Analysis of event history data, including the Cox proportional hazards regression model;
3. Explain regression diagnostics;
4. Explain the principles of and differences between resampling methods, such as bootstrap and permutation testing;
5. Describe and apply the basic principles of longitudinal data analysis;
6. Apply the above-mentioned techniques using R statistical software;
7. Name the situations in which these techniques can be applied and the conditions that should be met to obtain reliable results using these techniques;
8. Explain and interpret the results obtained with these techniques, and apply these results to answer a research question;
9. Recognize when classical statistical methods are inappropriate and other approaches need to be considered (e.g. more complex statistical or data science approaches)

Content

Contact details: Educational Office Epidemiology
E-mail: msc-epidemiology@umcutrecht.nl

Registration:
Face-to-Face: Students can register for this course via Osiris Student. More information about the registration procedure can be found here on the Students' site. NOTE Students of the MSc Epidemiology (Post Graduate) that register in time (i.e. at least two weeks before the start of a course) will always be admitted to the course unless it is completely full. Other students will receive information about their application two weeks before the start of the course.
Online: Online courses are only available for Epidemiology Postgraduate students and can register via Osiris Student. More information about the registration procedure can be found here on the Students' site.

Course coordinator:
Dr. C.L.J.J. (Cas) Kruitwagen, UMC Utrecht, Julius Center for Health Sciences and Primary Care, Utrecht, the Netherlands

Course description:
This course provides statistical methods to study the association between (multiple) determinants and the occurrence of an outcome event. The course starts with an introduction to likelihood theory, using simple examples and a minimum of mathematics. Next, the most important regression models used in medical research are introduced. Topics are: maximum-likelihood methods, logistic regression, model validation and regression diagnostics, Poisson regression, and analysis of `event-history´ data, including an extensive discussion of the Cox proportional hazards regression model. Also, the basic principles of longitudinal data analysis are taught.

Literature/study material used:
-
  
Mandatory for students in own Master’s programme:
Epidemiology & Epidemiology Postgraduate
 
Optional for students in other GSLS Master’s programme: Yes
 
Prerequisite knowledge:
Introduction to Statistics
Classical Methods in Data Analysis