Courses

1. Learning outcomes
   On a general level the student shall acquire practical knowledge of the important components of formal demography. Specifically, students will be able to:
   • use infinitesimal, differential, integral, and matrix calculus in their future practice;
   • acquire an overview of usual formalization in demography.
2. Course content
   The course is divided into several modules:

• • Causation
    The course aims to introduce students to principles of counterfactual causal inference, providing them with an overview of some methods in this field. Focus is placed on methods for conditioning on observables (i.e. regression adjustment, g-computation, propensity score methods, imputation methods for missing data), with a possibility, if there is interest, to also delve into methods for conditioning unobservables (i.e. individual intercepts, regression discontinuity, difference-in-differences, and instrumental variables).
  • Event History Analysis
    During these courses, students shall acquire practical knowledge of the important components of formal demography. Specifically, students will be able to use infinitesimal, differential, integral, and matrix calculus in their future practice and acquire an overview of usual formalization in demography. The module will deal primarily with so-called event history models, statistical techniques used to analyze the occurrence of events in time, such as death, marriage, childbirth, entry into retirement, etc. The following topics will be covered:
    • Characterizing duration distributions and common parametric families
    • Observation schemes: censoring and truncation
    • Nonparametric approaches
    • Basic hazard regression (proportional hazards)
    • The Cox PH model, model diagnostics
    • Discrete-time hazard regression
    • The piece-wise constant hazard model; aggregate event-data
    • Non-proportional hazards models
    • Unobserved heterogeneity, repeated events, competing risks, multistate models

• • Sequence Analysis
    The increasing availability of (longitudinal) data and the explosion of big data in demography and other social sciences present methodological challenges in relation to the significant reduction of information that allows the isolation and identification of crucial properties and relationships. At the same time, there is growing interest in understanding what are the relevant patterns in time-related processes that have recently gained complexity such as professional careers, family trajectories or migration pathways. The course aims to provide a concise introduction to sequence analysis, its origins and its applications, which will be demonstrated with hands-on practice using the software R (package TraMineR). The course covers basics of data management, major techniques of algorithmic sequence comparison, sequence visualization, grouping of sequences in typologies, as well as recent developments and applications in the social sciences.

• • Qualitative Research in Demography
    The rise in qualitative contributions to research endeavours has been mirrored in Demography, traditionally a quantitative discipline. Since 1980s, populations studies scientists have increasingly used qualitative methods. The methods used have themselves shifted over tome, from ethnographic approaches, towards increasing reliance on focus groups and in-depths interviews. Qualitative approaches are used in demography at multiple points in the production of knowledge, including: design and testing of quantitative questionnaires; to understand unexpected survey results; and, to grasp sensitive issues, perceptions, “cultural contexts”, and other elements of the social world which are difficult to measure quantitatively.
  • Multilevel Data Analysis
    This course will teach you a basic conceptual understanding of the multilevel (also known as mixed or hierarchical) analysis. The seminars will cover theoretical fundamentals of the multilevel approach to the data and show potential applications, paying particular attention to the solid understanding of key concepts (such as levels, fixed and random coefficients, variability or shrinkage), reasonable approach to model building, and interpretation of the results. Issues related to contextual effects, within-between models, and cross-level interactions will also be discussed. Further, the course will devote time to specific applications, such as repeated measures (panel data analysis) or non-linear models, and the Bayesian approach to multilevel analysis.
  • Agent-based Modelling and Simulation
    The aim of the course is to connect microsimulation, agent-based modeling (ABM), and probability theory. Microsimulation models and ABM rely on random variables and their probability distributions. One particular function, the quantile function or inverse distribution function, is the workhorse of both microsimulation and ABM. Different quantile functions are covered and their use in stochastic agent-based modeling discussed. Illustrations are in R.

3. Assessment
The course is designed as a series of lectures and seminars. Grading is based on individual performance, via written assignments, oral presentation or group activities.