Skip to content

[R-meta] converting p-values and t-statistics into correlation coefficients

2 messages · Alan Wilson, Wolfgang Viechtbauer

#
Hey everyone - I have a student in my meta-analysis class who wanted convert p-values and t-statistics into correlation coefficients as part of their class project, but we are struggling to get correlation effect size calculations using metafor.  We have found an online resource through the Campbell Collaboration and could create an excel spreadsheet for the calculations, but we would prefer to use metafor for these conversions.

Using this example, escalc(measure="COR", pi=-.07, ni=30), from https://wviechtb.github.io/metafor/reference/escalc.html, we get the following error.  It is not clear that pi is available in escalc.  Any advice how to overcome this hurdle is appreciated and will be shared with my class.  Thanks for any help.  alan

Error in escalc(measure = "COR", pi = -0.07, ni = 30) :
  Cannot compute outcomes. Check that all of the required
  information is specified via the appropriate arguments.
In addition: Warning message:
Extra argument ('pi') disregarded.



---
Alan Wilson - Auburn University
Professor & Graduate Program Officer
Assistant Director for Instruction
School of Fisheries, Aquaculture, and Aquatic Sciences
www.wilsonlab.com<https://www.wilsonlab.com/> - 334.246.1120
#
Hi Alan,

This is stuff I added recently, so it is in the 'devel' version, which you can install as described here:

https://wviechtb.github.io/metafor/#installation

I am quite close to releasing a new version of metafor via CRAN, but in the meantime, you can work with that.

Best,
Wolfgang