We seek to appoint a PhD candidate who will investigate the multilevel explicit-duration hidden Markov model for real time behavioural data. Due to technological advances, it becomes increasingly easy within the Social Sciences to collect data on behavior as it unfolds in real time. These new data enable a novel perspective on investigating behavior: studying the dynamics of behavior over time. This in contrast to the static summaries of behavior that are currently typically obtained. To extract the dynamics of behavior over time, the statistical model of choice is the hidden Markov model. HMMs are a machine learning method that have been used for several decades in many different scientific fields, such as speech recognition and DNA segmentation. To make the HMM the perfect match for real time behavioral data, the conventional HMM must be extended in two ways: 1) the HMM is extended to the multilevel framework such that we can model the observed sequences of multiple subjects simultaneously, 2) the durations of the latent behavioral states need to be explicitly modeled and allowed to deviate from a geometric distribution by using an explicit duration hidden Markov model (ED-HMM). The multilevel ED-HMM provides the perfect match to summarize real time behavioral data and extract novel information: it allows one to model the dynamics of behavior over time, and quantify and predict heterogeneity in behavioral dynamics between subjects. The ED-HMM within the multilevel framework is a novel method and not yet described in literature, and is a viable method as shown by extensive preliminary results.

Research goals are:
1. Improve the algorithm for estimating the multilevel ED-HMM to reduce computational intensity while maintaining robust and unbiased estimation performance.
2. Develop a user friendly and open source software package such that applied researchers can use the developed statistical method.
3. Investigate on how many subjects observational sequences should be collected and how long these observational sequences should be when applying the multilevel ED-HMM.
4. Investigate how the required sample size of the ED-HMM depends on the complexity (e.g., number of hidden states, number of dependent variables) of the data.

The multilevel ED-HMM is implemented within the statistical package R (and partly in C++). Regarding the specific research questions:

• a literature study will be conducted to narrow down optimal possibilities to improve the algorithm for estimating the multilevel ED-HMM. A small number of possibilities will be implemented and tested using simulation studies;
• an official R package will be developed, including an extensive tutorial and workshop.
• simulation studies will be conducted.

This project will be embedded in the Department of Methodology and Statistics, and the candidate will be supervised by Prof. Dr. Irene Klugkist and Dr. Emmeke Aarts. Read more about the project here (.pdf).

Job tasks/responsibilities

• Conducting the research (literature research, develop and implement a novel analysis method, setup and conduct simulation studies, reporting the results), resulting in international scientific publications and a dissertation.
• Develop an user friendly and open source software package that implements the developed statistical method.
• To present results at national and international scientific conferences.
• Teaching (max. 10% per year).
• Active participation in the research group of the project and of the Department.
• Ambition to collaborate with (inter)national partners and to pay visits to these partners.

We are looking for someone who:

• holds (or nearly holds) a Master’s degree in Methodology and Statistics or a related field;
• is excellent in programming with R;
• preferably has affinity with data science and intense longitudinal data;
• has good social skills;
• is effective and efficient, and able to think conceptually;
• is able to meet deadlines, and conduct research independently and as part of a team;
• has good communication skills (written and oral) and statistical skills in English.

We offer a temporary position (1.0 FTE) for one year, starting preferably September 1, 2019. Upon a positive performance, the appointment will be extended for three further years. The gross salary - depending on previous qualifications and experience - ranges between €2,325 and €2,972 (scale P according to the Collective Labour Agreement Dutch Universities) gross per month for a full-time employment. Salaries are supplemented with a holiday bonus of 8 % and a year-end bonus of 8.3 % per year. We offer a pension scheme, (partly paid) parental leave, collective insurance schemes and flexible employment conditions (multiple choice model). More information is available at our website: working at Utrecht University.

A better future for everyone. This ambition motivates our scientists in executing their leading research and inspiring teaching. At Utrecht University, the various disciplines collaborate intensively towards major societal themes. Our focus is on Dynamics of Youth, Institutions for Open Societies, Life Sciences and Sustainability.

Utrecht is a young and vibrant city with a large academic population, around 30 minutes south of Amsterdam. It combines a beautiful old city centre with a modern university. Utrecht has an excellent quality of life, with plenty of green space and a strong bicycle culture.

The Faculty of Social and Behavioural Sciences is one of the leading Faculties in Europe providing research and academic teaching in Interdisciplinary Social Science, Cultural Anthropology, Educational Sciences, Pedagogical Sciences, Psychology, and Sociology. More than 5,600 students are enrolled in a broad range of undergraduate and graduate programmes. The Faculty of Social and Behavioural Sciences has some 850 faculty and staff members, all providing their individual contribution to the training and education of young talent and to the research into and finding solutions for scientific and societal issues.