PhD position in machine learning scientific inference behavioural science

PhD position in machine learning for scientific inference for behavioural science

Deadline 13 Mar ’26 Published 12 Feb ’26 Vacancy

We are looking for a PhD candidate (1.0 fte) who will make machine learning more scientifically useful. At the intersection of statistics, machine learning, and behavioural science, you will develop meta analysis for results from machine learning.

ML methods provide unprecedented flexibility and powerful predictions, which are critical for modeling the complex and often high-dimensional associations underlying human behavior. However, results of ML models are more difficult to interpret than those of traditional statistical methods and uncertainty quantification is rarely provided, making it difficult to obtain generalizable scientific conclusions. We will address this challenge by developing a) methods that produce valid, generalizable and interpretable effect sizes with accurate uncertainty estimates;

b) ML-based meta‑analysis, in which results can be compared and combined across studies. The project is led by Dr. Marjolein Fokkema and funded by the Dutch Research Council (NWO).

Your tasks will involve:

• Developing new statistical methods and implementing them in open‑source software.
• Testing the new methods through simulation studies and applying them to real‑world behavioural science data.
• Publishing results in scientific journals and presenting them at national and international conferences.
• Collaborating with researchers from behavioural science and related fields to better understand studies and datasets the methods could be applied to, and facilitating such applications, for example through methodological or software improvements, or the writing of software documentation and tutorials.
• Taking courses and workshops tailored to your development, for example (but not limited to) those offered within the Graduate School of Social and Behavioural Sciences.

The Faculty of Social and Behavioural Sciences consists of five institutes: the Centre for Science and Technology Studies, Cultural Anthropology and Development Sociology, Education and Child Studies, Political Science, and Psychology. The faculty is home to approximately 7,000 students and 1,000 staff members. Our institutes are dedicated not only to education but also to groundbreaking research that expands our understanding of human behavior and societal structures.

What makes our faculty unique is the diversity of research topics, the variety of teaching approaches, and the structure of professional support. This provides you with the opportunity to explore and develop your interests and expertise. For an introduction to our faculty, visit our website:
Welcome to the Leiden Faculty of Social and Behavioural Sciences – Leiden University.

You will be working within the Methodology and Statistics Unit, a dynamic unit focusing on Neuroimaging Statistics, Statistical Learning and Artificial Intelligence, Applied Psychometric and Sociometric Modelling, and Responsible Research Methods. The successful candidate will have the opportunity to work with a dynamic, international team of researchers and contribute to cutting‑edge research in their field. The team values scientific integrity, open science, and inclusiveness.

If you are a highly motivated individual with a passion for data analysis and research, please apply with your resume and motivation letter.

• A completed (research) master's degree in statistics, data science, psychology, or a related quantitative field.
• Strong programming skills in R, experience with data analysis and Monte Carlo simulation studies.
• Strong written and spoken English, clear communication skills (both written and oral).
• A genuine interest in behavioural science, for example illustrated by relevant coursework, projects, or other academic or extracurricular activities.

In addition, the following skills are desirable (but not required):

• A background in methods for Bayesian (high-dimensional) regression, interpretable machine learning, and meta‑analytic techniques, as evidenced, for example, by relevant coursework and research or thesis projects.
• Experie…