Skip to content

[R-meta] MA of regression coefficients (as accuracy indicators) in pre-post test

2 messages · Emerson M. Del Ponte, James Pustejovsky

#
James,
1) Yes, and I have the Lin?s CCC for about 15% of the articles for which I have the raw data, or all estimates for the sample of leaves for each rater.
2) No, what I have is only the regression coefficients and correlation coefficient (Pearson?s r)  (estimates regressed again actual values) for each rater. Wonder if there is a way to calculate the Lin?s bias coefficient (Cb) from the intercept and slope? The Lin?s CCC is r * Cb, and I have only r.
Thanks,
Emerson

Emerson,

Two quick questions to see if there is an easy solution:
1) Do you mean Lin's concordance correlation (DOI: 10.2307/2532051)?
2) Do you have (or can you extract) the means of the pre-test or post-test
values for each rater?
If the answers to both questions are "yes" then it is possible to compute
an estimate of the concordance correlation from the available summary
statistics.

James

On Fri, Dec 1, 2017 at 6:06 AM, Emerson M. Del Ponte <edelponte at gmail.com <https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis>>
wrote:

  
  
#
Unfortunately I don?t think there is a way to calculate Cb (or the ccc) without knowing the mean of one of the variables. Without one or the other, I can?t see any way to determine the degree of shift (above or below the line y = x) in the regression line.