V. Summer School

We organized the third summer school on the topic of “Generalized Linear Mixed Models with Applications in Medicine”. (May 26-27, 2014)


Generalized Linear Mixed Models with Applications in Medicine – a 2 day course funded by ERASMUS

Presenters: Professor Dankmar Böhning and Dr Stefanie Biedermann (University of Southampton), Mr. James Gallagher (University of Reading)
Organizators: Professor Mehmet Orman and Dr Timur Köse, Ege University, Izmir, Turkey

Dates of Course: 
Monday 26th and Tuesday 27th May 2014

Summary of Course:
This course will focus on the application of generalized linear and generalized linear mixed models for medical applications with a binary or count outcome. Topics will include simple and more complex hierarchical data structure such as measurements on patients (for example length of stay in hospital) within wards within hospitals, crossed and nested effects, fixed and random effects as well as random coefficient models. The course will give an introduction to the generalized linear model and extend it generalized linear mixed models to cope with potentially nested fixed and random effects simultaneously. All models will be illustrated at hand of study data. The course will include a mixture of lectures and practical workshops using the software STATA.

Course Objectives:
By the end of the course participants should:

  • Have a practical understanding of the ideas and methods of generalized linear and generalized linear mixed modelling
  • Develop an understanding of analysing hierarchically structured data as well as have an understanding of the multilevel designs that lead to these complex data structures
  • Have a detailed understanding of crossed and nested effects and how they are dealt with statistically
  • Gain a working knowledge of the generalized linear mixed model parts of the package STATA
  • Be able to apply these methods to hierarchical data arising from multilevel designs

Course Content:

  • Data with non-normal response: counts and binary outcome; generalized linear model including link-function, linear predictor and error distribution.
  • Simple cluster structure: several observations per unit- the random effects model with one factor.
  • Crossed and nested effects.
  • Data with more complex hierarchical structure -Several observations per unit, each observation consists of several measurements.
  • The generalized linear mixed model -fixed and random effects - model evaluation (AIC, BIC, LRT).
  • The random coefficient model - Useful for modelling response in time - Random intercept and random slope.

The course will have a practical emphasis with computer workshops allowing participants to work through examples using the STATA software.



Organized by Professor Mehmet Orman (MO)  and Dr Timur Köse (TK)  (Ege University)
Lecturers: Professor Dankmar Böhnning (DB) and Assoc. Prof. Stefanie Biedermann (SB) (University of Southampton) and Mr. James Gallagher (JG)  (University of Reading)



  • 9.00 -9.30       Welcome and opening (MO, TK)
  • 9.30-10.30      L1: From linear models to generalized linear models (DB)
  • 10.30-11.00    Morning break – Coffee and biscuits
  • 11.00-12.00    L2: Poisson and logistic regression (DB)
  • 12.00-12.30    P0: Introduction to STATA (JG)
  • 12.30-14.00    Lunch
  • 14.00-15.00    P1: Poisson and logistic regression (DB)
  • 15.00-15.30    Afternoon break – Tea and biscuits
  • 15.30-16.30     L3: Random effects and hierarchical structures (SB)<


  • 9.00-10.30       L4: Linear mixed and generalized linear mixed models (DB)
  • 10.30-11.00    Morning break – Coffee and biscuits
  • 11.00-12.30    P2: Generalized linear mixed models (SB)
  • 12.30-14.00    Lunch
  • 14.00-15.00     L5: Random coefficient models for non-normal data (JG)
  • 15.00-15.30    Afternoon break – Tea and biscuits
  • 15.30-16.30     P3: Random coefficient models for non-normal data (JG) 
  • 16.30-17.00     Closing Ceremony (MO, TK)




UYTES Program is provided by the State.