Thu, 30 July
|Future of Work Institute
Using DAGs to aid causal inference in field research
Introducing the latest tool for aiding causal inference, Directed Acyclic Graphs (or DAGs), and show how DAGs can be used to help identify variables you must measure, variables you must control for, variables you must not control for, & importantly, what DAGs can teach us about mediation analysis.
Time & Location
30 July 2020, 1:00 pm – 3:00 pm
Future of Work Institute, 78 Murray St, Perth WA 6000, Australia
About the Event
Using DAGs to aid causal inference in field research
Thursday July 30, 2020, 1PM-3PM (Online & Face2Face - see below).
The Future of Work Institute’s mission is to promote thriving at work in the digital age. To succeed, we must show that the interventions we recommend have a causal effect on worker thriving. And yet, in field studies, establishing that any ‘X’ causes ‘Y’ is extremely difficult. In a previous workshop, Pat Dunlop discussed how either failing to condition on confounding variables, or conditioning on collider variables will introduce bias to causal estimates. In this workshop, Pat and Stijn Masschelein, from UWA, will introduce attendees to the latest tool, developed by Prof. Judea Pearl and his team, for aiding causal inference, namely Directed Acyclic Graphs (or DAGs), and show how DAGs can be used to help identify variables you must measure, variables you must control for, variables you must not control for, and importantly, what DAGs can teach us about mediation analysis. This workshop will include a recap, a “DAGs 101” presentation with a tutorial, and some simulations designed to illustrate what can go wrong if our assumptions are wrong, but we don’t know it. We recommend watching the video of Pat’s first causality workshop if you were not there. Delivering this workshop is part of our own learning journeys as we navigate this difficult topic!
Presenters
- Patrick Dunlop is an Associate Professor at Curtin Future of Work Institute. He obtained his PhD at the University of Western Australia in 2012. His research interests include personnel recruitment, assessment, and selection. He also dabbles in research methods.
- Stijn Masschelein is a Senior Lecturer at the University of Western Australia. He obtained his PhD at Leuven University in 2010. Stijn's main research interests are the role of performance measures and informal information on incentives and decision making.
Venue and RSVP
This session will be run both via live stream, over the web, AND face to face at 78 Murray Street, in Level, 2, Room 212. Due to social distancing restrictions, spaces in the face to face session are extremely limited. To reserve your spot, and to obtain the link to the stream, please contact Diane.Garnham@curtin.edu.au as soon as possible with your preference.