Our panel starts with the following related questions:
- When we talk about the “science of data analysis,” what different things might we mean?
- In statistics and computer science there is a often a gap between theoretical models and applied methodology. In the context of the “science of data analysis” how might we bridge this gap?
We have two concrete examples in which to think about these questions.
- How should the current debate over the value of reporting statistical significance influence the development of methods for analyzing data privately? Reference: American Statistician, Vol. 73, Issue 1. https://www.tandfonline.
- How do we reconcile different views/approaches to the problem of adaptive data analysis suggested in different literatures with statistical practice? References: http://www.stat.
columbia.edu/~gelman/research/ unpublished/p_hacking.pdf and http://science.sciencemag.org/ content/349/6248/636.full
If you have other ideas/examples/questions that you would like to see addressed by the panel, please email Aleksandra Slavkovic (sesa at psu.edu).