User:OpenScientist/Open grant writing/Testing the efficiency of open science - Wikiversity

"This page is part of a workshop at Open Science, Open Issues, aimed at drafting a skeleton for a research proposal to put open science to an efficiency test in comparison to how science is usually performed. See this blogpost for background and this Etherpad for notes from an earlier workshop on the same subject ... Basic idea - I think the aim of this project is to convince researchers and funders that research projects that are usually carried out in a closed fashion can be open. So, projects which are inherently open (e.g. based on thousands of volunteers gathering or analyzing data) are not useful to test the hypothesis. I think the funds should be directed to projects that could indeed be carried out as traditional closed research, so that we can later extrapolate our results.  - use a significant amount of money (think USD 1M or more) exclusively for open science - using crowdsourcing? in principle an open project with volunteers can be very cost effective in strict monetary terms because the volunteers donate their time. Good example is Galaxy Zoo where Arfon Smith would have catalogued 25K galaxies in one PhD, but volunteers catalogued 1 million in same time. Allowing for infrastructure development Can we say this was >10 times more costs effective. (We can do the calculation in hindsight - and they probably have been done). - analyse its output How are projects usually evaluated? by peers. Peers evaluate projects all the time, using a mixture of quantitative metrics (publications, phds, etc) and their opinion on the results. I think one could design a double-blind test with a lot of evaluators that have to grade projects, without knowing if those projects are open or closed research (maybe that´s difficult) ..."