King's College London

Research portal

Dr Daniel Stahl

Research interests

I am a Reader in Biostatistics and Head of the Statistical Learning Group. My interest is applying statistical and machine learning methods to identify predictors, mediators and moderators of treatment success and using model-based cluster analysis methods to identify subgroups among psychiatric patients. As a Lead Trial Statistician,  I am responsible for overseeing the statistical aspects of several clinical trials within the IoPPN.  I am further interested in model selection problems and - a blast from the past - in the evolution of social system in primates. 


As Education Lead I lead the educational activities of our department which involves bringing statistics to life for a variety of students and researchers.I am the Education Lead of our department. I am teaching introductory and advanced statistics courses for MSc, Dclinc Psych and PhD students and researchers of the Institute of Psychology, Psychiatry and Neuroscience including Introduction to statistics, Mediation and moderation, Model selection, Multiple testing, Structural equation modelling, Scale development and Statistical learning methods for prognostic models and stratified medicine. I am the statistical supervisor for 60 Dclinc Psych students and involved in the development of an online E-learning module "Research methods and Statistics" and organizer of the summer school "Prediction Modelling".


Statistical Learning and Prediction modelling group

In recent years there has been a shift towards stratified (personalized) medicine in which individ­ual characteristics or biomarkers are used to identify patients who are more likely to respond to treatment. In the era of “Big Data”, prediction modelling cannot rely on classical statistical methods and computer intensive machine learning methods are increasingly needed. However, apply­ing such methods in mental health research involve many methodological challenges such as missing data, unbalanced groups, popula­tion substructure, multi-centre trials, multicollinearity, measurement error, different measures for the same underlying construct and validation of predic­tive models. Many such methods are “black boxes” and have limited potential to explain the underly­ing process. The aim of the Statistical Learning and Prediction modelling group is to combining machine learning methods from computer science and statistical modelling to overcome the prob­lems described above and to apply them to clinical data.



Researcher ID

Google Scholar ID


© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454