Session title: Philosophy of Science and the New Paradigm of Data-Driven Science
Organizer: Todd Kuffner (Washington U)
Chair: Todd Kuffner (Washington U)
Time: June 4th, 1:45pm – 3:15pm
Location: VEC 902/903
Speech 1: Your Data-Driven Claims Must Still be Probed Severely
Speaker: Deborah Mayo (Virginia Tech)
Abstract: I analyse some of the philosophical claims from leaders of the “new paradigm” of data-driven science regarding the end of theory and the obsoleteness of scientific method. Science may be data intensive, but data are theory laden, so conceptual assumptions and biases must still be examined. Science may be question driven, but good science can’t opt out of probing flaws in any potential answers to those questions.
Speech 2: On the replicability of scientific studies
Speaker: Ian McKeague (Columbia)
Abstract:
I will discuss a number of issues, both statistical and philosophical, related to the replicability and verification of scientific results. In particular, I will discuss a recent proposal of Munafo and Davey Smith (Nature, 2018) that such verification requires disparate lines of evidence, a technique that they call triangulation.
Speech 3: Conducting Highly Principled Data Science: A Statistician’s Job and Joy
Speaker: Xiao-Li Meng (Harvard)
Abstract:
This talk is based on a contribution to the special issue on “The Role of Statistics in the Era of Big Data” organized by Statistics and Probability Letters,
with the title above and the following abstract: “Highly Principled Data Science insists on methodologies that are: (1) scientifically justified; (2) statistically principled; and (3) computationally efficient. An astrostatistics collaboration, together with some reminiscences, illustrates the increased roles statisticians can and should play to ensure this trio, and to advance the science of data along the way. “