SER is pleased to offer a Mid-Year Meeting held via web conference. This event is free of charge for current SER members.
2020 SER Mid Year Meeting!
Live web-conference from Salt Lake City held on March 6, 2020
“Theory + methods + training: reflections on the future of epidemiology”
SYMPOSIUM #1 - 10:15am - 11:30am MDT
Where do DAGs come from?Read more
It is well established that selection of covariates into an adjustment set for valid causal effect estimation is premised on correct specification of the DAG, but where does the correctly specified DAG come from? In Cox’s “Principles of Statistical Inference” (2006) he wrote that “how the translation from subject-matter problem to statistical model is done is often the most critical part of an analysis”, and yet it is the most neglected aspect of epidemiologic analyses in traditional curricula. This symposium explores the process by which previous observation coalesces into theory, which guides the formulation and arrangement of the DAG to allow for specification of statistical models that yield valid causal estimates. Given that epidemiologic theory is often underdeveloped relative to other disciplines like physics or chemistry, we often resign ourselves to Box’s pessimism that “all models are wrong”. How does this weakness of theory translate into uncertainty in the model outputs and their public health implications? Epidemiologists seem to be desperately in need of some overarching theory for how to develop and apply theories in routine analysis and inference.
Model selection as a philosophical and practical problem, and critiques of common covariate discovery and screening algorithms.
Model fit is the criterion used to guide model specification, but it is subject to “over-fitting” and other errors, such a Freedman’s Paradox, when investigators conduct brute-force specification searches on the dataset. Read more
Using Social Theory to form Epidemiologic Theory:
Incidence of violence and intentional injury are driven more by pathological psychological and social processes than by innate physiological determinants, Read more
Using Biological Theory to form Epidemiologic Theory:
The epidemiology of pregnancy is a morass of adjustment conundrums, with complicated judgments about censoring, competing events, and time windows. Read more
How “in silico” experiments can be used to develop the theories that populate DAGs with dependencies and independencies. Read more
SYMPOSIUM #2 - 11:45am - 1:00pm MDT
How should future epidemiologists learn epidemiology?Read more
“Authentic instruction: Making epidemiology relevant to high school students through student-centered, community engaged learning”
“Innovative, inquiry-based learning techniques for undergraduate epidemiology and heterogeneous classrooms”
“Social media and podcasts: New media as mechanisms for engaging with peers and seasoned scholars”
“Epidemiology by design: integrating causal ideas into fundamental methods”