Log in or create a free Rosenverse account to watch this video.
Log in Create free account100s of community videos are available to free members. Conference talks are generally available to Gold members.
Summary
You've probably heard the old adage "correlation does not imply causation" but at some point we've got to say that drinking boiling hot tea and burning our tongues are more than just "strongly correlated". Enter Causal Inference, a collection of techniques for reasoning about the relationship between effects that can be measured and causes that can be identified. It helps us bridge between what we hear when we talk directly to users and what we observe about their behavior at scale and over time. This talk will introduce some of the key techniques in causal inference and how they can be used by UX Researchers and Designers to understand potential users, current users, and the products we might build.
Key Insights
-
•
Causal inference bridges qualitative reasoning about context with quantitative analysis of data.
-
•
Directed acyclic graphs help visualize causal pathways and common pitfalls like confounders and colliders.
-
•
Randomized controlled experiments require random assignment, blinding, and pre-specified hypotheses to measure causal effects reliably.
-
•
Average treatment effect summarizes impact on an entire population, whereas intent-to-treat reveals experiment adherence issues.
-
•
Quasi-experimental methods like regression discontinuity and difference-in-differences allow causal inference when randomization is not possible.
-
•
Matching is an art, relying on context-aware ranges rather than exact numeric equivalence to equate comparison groups.
-
•
Synthetic control methods build counterfactual scenarios when no good control group exists.
-
•
Understanding and designing for interpretability strengthens collaboration between designers and analytics or marketing teams.
-
•
Qualitative research poses valuable questions; quantitative causal inference sharpens and tests these questions.
-
•
Behavioral causes revealed by causal inference may differ from user self-reported reasons, offering deeper insights.
Notable Quotes
"Correlation doesn’t imply causation, but sometimes two things happening together do imply a causal mechanism."
"Qualitative reasoning about causal mechanisms usually comes from lived experience, either the researcher or participants."
"Designers have a hard time interpreting econometrics-driven quantitative research without bridging approaches."
"Random assignment to treatment or control is essential—if it’s not random, it’s not a true experiment."
"If you cannot remove a treatment, it may not be a true cause."
"Matching is very much an art, not a science; there are no hard and fast rules for what counts as a good match."
"Difference-in-differences lets you study treatment effects in groups that start in very different places by looking at trends."
"Synthetic controls create counterfactuals using historical data when no suitable control group exists."
"Qualitative research really poses questions; quantitative methods let you investigate those questions rigorously."
"You don’t want to overdesign just to make something more data interpretable because that can lead you down a dark path."
Or choose a question:
More Videos
"Throw away the rule book in early ideation phases to get as many ideas out as fast as possible."
Billy CarlsonIdeation tips for Product Managers
December 6, 2022
"If you haven’t done accessibility work yet, it makes sense to get started on mobile because you already have the tools in your pocket."
Sam ProulxMobile Accessibility: Why Moving Accessibility Beyond the Desktop is Critical in a Mobile-first World
November 17, 2022
"Developers are stinking gods among men, generous in their selfishness who keep fixing what’s not broken until it is."
Dan WillisEnterprise Storytelling Sessions
May 13, 2015
"Not a single person on this relationship map reported to me. That’s how I affect design at scale at Autodesk."
Michael PolivkaScaling Design through Relationship Maps
November 7, 2017
"The biggest challenge with LLMs was keeping them on task instead of writing their own Reddit posts."
Karen McGrane Jeff EatonAI for Information Architects: Are the robots coming for our jobs?
November 21, 2024
"One researcher per team for at least three days a week was considered an outrageous luxury at GDS in 2013."
Leisa ReicheltOpening Keynote: Operating in Context
November 7, 2018
"We do this with intention of fostering a more inclusive experience for audience members who might be sight impaired."
Kim Holt Emma Wylds Pearl Koppenhaver Maisee XiongA Salesforce Panel Discussion on Values-Driven DesignOps
September 8, 2022
"Don’t be surprised by change. Expect change."
Jacqui Frey Alison RandSetting the Table for Dynamic Change
October 24, 2019
"Users prefer the simpler and more delightful design even without usability testing."
How to Identify and Increase your "Experience Quotient"
June 15, 2018
