If asked, we all say we aim to to good research and make sensible decisions. In mathematics, the choice of criteria to optimise is often explicit, and we know there is no complete ordering in more than one dimension.
Statisticians involved in multi-disciplinary research need to reflect on how their understanding of uncertainty and statistical methods can contribute to reliable and reproducible research. The ISI Declaration of Professional Ethics provides a framework for statisticians. Judging what is "normal" and what is "best" requires an appreciation of the assumptions and guidelines of other disciplines.
I will briefly discuss the requirements for design and analysis in medical research, and relate this to debates on reproducible research and p-values in social science research. Issues arising from informed and uninformed consent will be outlined.
Examples might include medical research in developing countries, toxic tort or wrongful birth claims, big data and use of routine administrative or commercial data.
- Maths in Society Seminar