Tom Stafford<p>So far at this conference I have seen reports of true experiments, natural experiments, difference in difference analysis and regression discontinuity design - but no instrumental variable analysis </p><p>I wonder why?</p><p>I was hoping for the full set of causal inference methods</p><p><a href="https://mastodon.online/tags/ICSSI2025" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>ICSSI2025</span></a> <a href="https://mastodon.online/tags/CausalInference" class="mention hashtag" rel="nofollow noopener" target="_blank">#<span>CausalInference</span></a></p>
#CausalInference
0 posts0 participants0 posts today
Angela Carollo, Aapo Hiilamo and Mikko Myrskylä recently hosted a workshop to "demystify machine learning for population researchers" @MPIDR. More than 50 researchers came together to learn about new research techniques.
https://demogr.mpg.de/en/news_events_6123/news_press_releases_4630/news/demystifying_machine_learning_for_population_researchers_13655
#statstab #212 Eight basic rules for causal inference
Thoughts: Good explanation of the basics; confounder, colliders, randomization, and when to adjust.
#DAGs #causalinference #causal #collider #correlation #r
https://pedermisager.org/blog/seven_basic_rules_for_causal_inference/
Peder M. IsagerEight basic rules for causal inference | Peder M. IsagerPersonal website of Dr. Peder M. Isager
To fully realize the potential of our clinical trials, we must go beyond randomization, and use causal inference and pharmacometric modelling and simulation. Advancing both we show that non-linear mixed effects modelling implements the equivalent of standardization in causal inference. Dive into this if you're into #causal #causalinference #DAGs #pharmacometrics, or clinical development #stats.
My old boss is hiring for Causal Inference work.
"Do you have some causal inference experience? Do you want a role that helps you build the science roadmap that impacts millions of customers every single day? I'm hiring in NYC and would love to talk to you."
#CausalInference #whypy #dowhy #fedihire #fedijob
https://www.amazon.jobs/en/jobs/2775999/senior-data-scientist-ring-data-science-and-engineering
amazon.jobsSenior Data Scientist, Ring Data Science and EngineeringCome build the future of smart security with us. Are you interested in helping shape the future of devices and services designed to keep people close to what’s important?ABOUT RINGWe started in a garage in 2012 when our founder asked a simple question: what if you could answer the front door from your phone? What if you could be there without needing to actually, you know, be there? After many late nights and endless tinkering, our first Video Doorbell was born.That invention has grown into over a decade of groundbreaking products and next-level features. And at the core of all that, everything we’ve done and everything we’ve yet to build, is that same inventor's spirit and drive to bridge the distance between people and what they care about. Whatever it is, at Ring we’re committed to helping you be there for it.(https://www.ring.com)ABOUT THE ROLEThe Senior Data Scientist within Ring Data Science and Engineering plays a pivotal role in shaping how we carry the voice of our customers. We strive to understand their behaviors and preferences in order to provide them with the best experience connecting with the places, people and things that matter to them. This role will build scalable solutions and models to support our business functions (Subscriptions, Product, Customer Service). By leveraging a range of methods including statistical analysis and machine learning, you will explain, quantify, predict and prescribe in support of informing critical business decisions. You will translate business goals into agile, insightful analytics. You will seek to create value for both stakeholders and customers and inform findings in a clear, actionable way to managers and senior leaders.Key job responsibilities- Drive shared understanding among business, engineering, and science teams of domain knowledge of processes, system structures, and business requirements.- Apply domain knowledge to identify product roadmap, growth, engagement, and retention opportunities; quantify impact; and inform prioritization.- Advocate technical solutions to business stakeholders, engineering teams, and executive level decision makers.- Lead development and validation of state-of-the-art technical designs (data pipelines, data models, causal inference, predictive models, data insights/visualizations, etc)- Contribute to the hiring and development of others- Communicate strategy, progress, and impact to senior leadershipA day in the lifeTranslate/Interpret • Complex and interrelated datasets describing customer behavior, messaging, content, product design and financial impact.Measure/Quantify/Expand • Retrieve, synthesize, and present critical data in a format that is immediately useful to answering specific questions or improving system performance. • Analyze historical data to identify trends and support decision making. • Improve upon existing methodologies by developing new data sources, testing model enhancements, and fine-tuning model parameters. • Provide requirements to develop analytic capabilities, platforms, and pipelines. • Apply statistical or machine learning knowledge to specific business problems and data.Explore/Enlighten • Formalize assumptions about how users are expected to behave, create statistical definition of the outlier, and develop methods to systematically identify these outliers. Work out why such examples are outliers and define if any actions needed. • Given anecdotes about anomalies or generate automatic scripts to define anomalies, deep dive to explain why they happen, and identify fixes. • Make decisions and recommendations. • Build decision-making models and propose solution for the business problem you defined. • Conduct written and verbal presentation to share insights and recommendations to audiences of varying levels of technical sophistication. • Utilize code (Python/R/SQL) for data analyzing and modeling algorithms.
One of my favorite science blogs, the 100% CI, has published a new blog post: „Sometimes a #causaleffect is just a causal effect (regardless of how it’s mediated or moderated)“ (by Julia Rohrer).
I like the opener: "It’s probably fair to say that many psychological researchers are somewhat confused about causal inference.“
Causal Inference in R
https://www.r-causal.org/
The book is in the making. What is available so far looks useful for teaching and accessibly written #rstats #CausalInference
www.r-causal.orgCausal Inference in R
Confusion of the day:
In electronics, an "open circuit" is one where electrons cannot flow (at least one switch is open) and a "closed circuit" one where electrons can flow (all switches are closed).
In causal inference, an "open path" is one where information can flow (all variables can vary) and a "closed path" one where it cannot flow (at least one variable is fixed / controlled).
Isn't this confusing? Is there maybe another way of thinking about it that makes more sense?
#statstab #116 An Evaluation of Four Solutions to the Forking Paths Problem from @MarkRubin
Thoughts: Forking paths is an interesting inference problem. Solutions: Adjusted Alpha, Preregistration, Sensitivity Analyses, and Fisherian p-value.
A great read from @rlmcelreath detailing some failures of the research industry's current practices and how we can move forward.
#science #ecology #causalinference #modeling #statistics #Evolution
https://elevanth.org/blog/2023/06/14/some-dumpster-fires-for-your-consideration/
Elements of Evolutionary Anthropology · Some Dumpster Fires for Your ConsiderationIn the interest of showing some reasons that the sciences are in bad shape, here are some raging dumpster fires for your consideration.
Replied in thread
@joakinen Also from this linked post, "(...) asking the right questions is one of the most important skills he’s learned", which is precisely the first step in #causalinference: ask a #causal question. The overlap between (computer science) #engineering and #philosophy through #causality may be one of the clearest examples of this needed change of mindset [1]. @Jose_A_Alonso
[1] https://cs.ulb.ac.be/conferences/ebiss2023/slides/EBISS2023_slides_JordiVitria_1.IntroCausality.pdf
Rethinking the pros and cons of randomized controlled trials and observational studies in the era of big data and advanced methods: a panel discussion https://doi.org/10.1186/s12919-023-00285-8
Fernainy et al (2024). BMC Proc.
via the #CasualInference pod https://casualinfer.libsyn.com/