I will be offering a workshop at the Global Amphibian and Reptile Disease conference in Knoxville this August. You can find more information about it here: https://utconferences.eventsair.com/gard-conference/workshops–field-trips
Our statistics, from regressions to SEMs, are association machines. They cannot, by themselves, tell us anything about causation, but only the degree and direction of statistical associations. We can, however, use Directed Acyclic Graphs to help us understand and test the consequence of assumed causal relationships, which can help us infer causation from statistical associations. That is, just by drawing a box-and-arrow sort of graph, we can make better sense of our data, related to amphibian and reptile health and beyond!
In this workshop we’ll cover the basics of DAGs, identify their testable implications, and learn why throwing all of our variables into a regression often makes our inference worse, not better. Our primary goal will be to place you, the researcher, back in the driver seat of your statistical inference. We will work with a handful of examples relevant to infectious disease research and practice drawing and thinking about causal diagrams based on our own research questions.
This workshop is ideal for students beginning to design their own studies, but is also intended for those wishing to develop a more formal version of their hard-won intuition for their own research or teaching.