Soap Box: Here's my 10-point plan for reproducibility in science.
1. Enhance orthogonal reproducibility - triangulation of @MarcusMunafo and colleagues - as much as direct reproducibility of the same experiment. Don't fetishise narrow reproducible and ignore scientific soundness.
2. Insist on open access publications (It's the 21st century people!) and encourage pre-prints. Support subject based literature access (eg, @EuropePMC_news).
3. Insist on open data on publication, deposited ideally in subject-specific global databases with community standards. The deposition and standards raises the game for everyone and makes everyone have that sneaky thought, "well, what if someone try to reproduce my analysis?"
(as well as enabling the sheer and near constant serendipity of people reusing data for analyses you would never have even dreamt about. Brilliant!).
(The @ELIXIREurope list of Deposition Databases elixir-europe.org/platforms/data… is the right thinking in biology, nearly all of them provided by or in partnership with @emblebi ).
4. Move towards all analysis must be available at source code and/or container level, eventually being made mandatory on publication. There needs to be a lead in time so people can adjust (otherwise the howl of where is my postdoc's scripts will just be everywhere).
5. Have a short section of power analysis in grant proposals (or why the scientist didn't think they needed the power analysis). No need to police the working heavily - you want to see the scientists being honest *with themselves* about what their experiments could discover
6. Encourage and ultimately insist on good Team Science recognition inside research organisations - developing respect and career paths for all aspects of science that is needed for solid results
7. Destigmatize withdrawing papers, analysis and datasets for honest mistakes. Withdrawing / correcting papers should be broadly celebrated, with the minority of outright fraud punished.
8. Have diverse funding schemes (types, duration, criteria) and diverse institutions (missions, sizes, location) populated with diverse people (gender, ethnicity, background) with diverse success criteria. Celebrate all this diversity; never think there is one way to do science.
9. Celebrate science and the scientific method. Each bit of hard won knowledge is a positive for all science. Listen, inquire, fool around, challenge and participate in order to understand the amazing world we inhabit.
10. Proactively and responsibly transfer the hard won knowledge and skills in science for society's benefit - for future generations, for health of citizens, for the health of the environment; use government, private commerce and charities as vehicles for this.
