Abstract
We believe at the core scientists want a causal answer to their question. It is indeed
true that for some systems - due to their enormous complexity - it is extremely hard to
get to a causal answer. But our literature survey showed that sometimes scientists end
up settling for correlation-based results. In the workshop, we will go over “Ten Simple
Rules for doing Causal Research”. Here, we will equip you with simple mathematical
tools (shown programmatically with python notebooks distributed to students) that
you can use in your research especially, while designing experiments, to ensure you can
make a causal claim. For example, what are the observations you will need, what are
things you must control for, how can your randomize to remove the effects of certain
confounders, and so on. In addition, we also go over some empirical tools developed and
used by economists, who also work with a complex dynamically changing the system,
while making policies and how we can borrow some of them to do causal neuroscience
research.
