Curriculum Vitae
Open Methodology
Open Methodology refers to a set of research practices that allow full transparency of methods and analysis. Like the movements for Open Access publication and Open Source software, Open Methodology has the potential to eliminate barriers to the efficient exchange of ideas. In research those barriers slow research progress and invite ethical transgression. Full openness requires three elements not usually required by journals in the social sciences:Freely available raw data
The low marginal cost of publishing materials online means there is no reason to be limited to what can reasonably fit into a journal. Data is the most valuable piece of any empirical work and sharing it publicly allows more minds to interpret and learn from the results.Freely available statistical code
Statistical methods are routinely described in published product of research, but with analysis the devil is always in the details. Even with the raw data and the published description of analyses it can be difficult to replicate the exact results published. By publishing the full code from data manipulation to significance tests, there need not be any ambiguity about the reported results. This has benefits of transparency but also of establishing stronger norms and standardizing procedures within a field.Open Source statistical computing
Having access to freely available data and statistical code is important, but there are limits to its usefulness if the code is written for use with proprietary software I don't own. Most proprietary statistical packages come with significant costs that introduce real barriers to scholars fully understanding one another's work. This problem is particularly large in interdisciplinary research where different fields have different software norms that decrease the likelihood of smooth statistical communication. With open source statistical computing software, anyone in the world has the tools to join a scholarly discussion.R is the most widely used open source statistical package/language and is freely available and runs on all operating systems. I recommend these resources for those looking to try it out:
- R Project homepage - The web presence is not as impressive as the software. Download R here and then for the most part, you can move on down the list of resources.
- R Studio - A very attractive front-end through that makes it much easier to get comfortable in R. Also fully cross-platform and free.
- UCLA Statistical Computing - UCLA hosts a really helpful collection of examples from all the popular analysis approaches.
- ggplot2 - The fanciest library of graphical tools to make your results looks extra impressive.
- Crantastic - A searchable, taggable, rateable directory of R packages (plugins). There are 3,000+ pacakges to do whatever statistics you need to do.