Dynamics and causality in the social and behavioural sciences

Course goals

After taking this course, students will be able to:

a. recognize the difference between causal, predictive, and descriptive research questions, and use directed acyclical graphs (DAGs) and the potential outcomes framework to make decisions about design and analysis
b. perform the identification and estimation techniques taught in this course
c. interpret the results in light of the underlying research question
d. critically evaluate cross-sectional and longitudinal research in the literature with respect to the research question that is posed.

Assessment
Two group assignments and a final exam.

The  assignments both count for 20% of the final mark, the exam for 60%. 

Each group assignment should be graded with at least a 5.5 to pass.

To qualify for a repair of the final result the mark needs to be at least a 4, or “AANV”.

Content

In this course students will learn the important difference between causal questions versus predictive or descriptive questions, and they will get acquainted with diverse techniques to study causality in the context of both cross-sectional and longitudinal data.

A particularly challenging task is how to properly use non-experimental (also referred to as observational or correlational) data for causal inference.
The reason for this is that there may be initial differences between groups to begin with.
For instance, suppose we are interested in studying whether children who grow up in poverty have lower high school grades on average than children who do not grow up in poverty.
Even if we find such a relation, we cannot conclude that poverty caused these children to do less well; we need to consider possible alternative explanations such as parental characteristics, genetic factors, and social norms, which may be the reason of the relation between poverty and school grades.
Similarly, if we are interested in the effect of recreational drug use on happiness among young adults, we need to consider other factors that may play a role, such as personality, level of happiness to begin with, employment status, et cetera.
In the past decades, researchers from various disciplines—including statistics, econometrics, epidemiology, and computer sciences—have developed formal frameworks and techniques for causal inference, which we will consider in this course. 

We begin with an introduction to the core areas within the  interventionist framework for causal inference, and discuss:
a) potential outcomes and how to define a causal effect;
b) how to account for confounding in our analyses;
c) graphical approaches to causal modelling based on directed acyclical graphs (DAGs) and how these can be used to guide design and analysis choices; and
d) how causal models themselves can be learned from data.

Students obtain hands-on experience with different techniques, including regression, stratification, matching, inverse probability weighting, d-separation, and conditional independence testing.
Subsequently, we see how these techniques can be used if we have repeated measures of the outcome, and when we have repeated measures of the cause (i.e., treatment) and the outcome.
Students will obtain hands-on experience with change score analysis (difference-in-differences), joint treatment effects, g-estimation, and multilevel modeling. 

Course form
Weekly lectures and lab sessions, both with mandatory attendance.
The lectures will introduce students to the various frameworks and techniques of causal inference with cross-sectional and longitudinal data.
Lab exercises and group assignments will allow students to gain hands-on experience with the techniques, and to deepen insight in the materials covered in the lectures (aims a-c).
In addition, the exercises and group assignments stimulate students  to critically evaluating the diverse approaches and their appropriateness in different contexts (aim d).

Additional information

Enrolment in this course is only available by using the pre-enrolment form that will be forwarded to all ADSM students in the second week of September.