All of the analyses, functions, data and text for this paper are available here. The easiest way to access this compendium is as an R package which may be installed using the following command:

FOR NOW:

You'll also need a beta version of our RItools package (soon to be uploaded to CRAN but just for now hosted on my own webserver):

install.packages("RItools",lib=Sys.getenv("R_LIBS_USER"),repos="http://jakebowers.org/R/",type="source",dependencies="Depends")

THEN:

install.packages("ManifestEffects",lib=Sys.getenv("R_LIBS_USER"),repos=c("http://jakebowers.org/R/","http://cran.r-project.org"),type="source",dependencies="Depends")

will install the compendium plus the other packages on which it depends.

The argument lib=Sys.getenv("R_LIBS_USER") indicates that the library should not be installed in the main library tree but in the default user library tree. You could also say lib="." to install into the current working directory, or choose some other place to install the package.

You can download the package directly from within R using the following command:

download.packages("ManifestEffects",destdir="~/TEMP/",repos="http://jakebowers.org/R/",type="source")

In this case, I asked R to download it to the "~/TEMP" directory. You could choose another directory.

Or you can download the Compendium directly using your web-browser at http://jakebowers.org/R/src/contrib/ManifestEffects_1.0.tar.gz.

To see the pdf of the paper as directly generated by the R packaging process do:

library(ManifestEffects)

if you installed the package in Sys.getenv("R_LIBS_USER") or using the lib.loc argument to the library() function to point to the place you installed the package.

Once R has loaded the library, to see the pdf of the paper type:

vignette("manifesteffects",package="ManifestEffects")

*Ways to look at the source of the paper:

If you want to look at just the R code do:

v1<-vignette("manifesteffects",package="ManifestEffects") 
edit(v1)

If you want to see the Sweave file (text and code):

file.edit(file=v1$file)

You can see that v1$file is a pointer to the actual Sweave document and thus you can use v1$file to open the Sweave document in your own R/LaTeX/Sweave editor of choice. In the GUI for Windows and OS X you can change the default editor under options/preferences. If you have configured R to use an external editor then edit(v1) will not cause R to wait for you.

*Re-doing the analyses:

If you want to re-run the all of analysis in the paper type:

Sweave(v1$file)

At the end of such a run, your R environment will be full of the objects (functions, data) created in the paper and you can inspect them at the command line if you wish.

Re-running the analyses shouldn't take too long because I have packaged the results of the most time-consuming simulations inside the vignette. If you want to re-run the simulations yourself, perhaps using a different seed, or perhaps fiddling with other settings, you'll need to move or rename the files referenced in the if(inherits(try(load(...)))) statements and be prepared to wait a while (one of the analyses takes advantage of parallel processing on multicore machines, the others do not).

*Uninstalling/Removing the compendium

You can remove the compendium from your user R package library by doing:

remove.packages("ManifestEffects")