Hello Daniel,

I have a data frame of 127 studies with the mean and SD that I want to transform into Hedges' g. But now I am updating my R to version 4.3.1, and the function is not working. The code worked previously in version 4.2.0. This is the code I used:

transform = esc_mean_sd(grp1m = df$mean1, grp1sd = df$sd1, grp1n = df$n1, grp2m = df$m2, grp2sd = df$sd2, grp2n = df$n2, es.type = "g", study = df$author)

This is the error message:
Error in missing(grp1sd) || is.null(grp1sd) || is.na(grp1sd) : 'length = 127' in coercion to 'logical(1)'

Do you have any suggestion on how to fix issue? Thank you.

It says on the documentation that for esc_rpb I can ask it to give me the effect size r in the output. When I try it, it says that is has to be any of the other types listed (d, g, logodds, etc). the default for it seems to be to spit out a cohen's d, which isn't what I thought the default would be. either way, it'd be nice if it would allow me to get output with the effect size as an r.

When I want convert (un)standardized beta to r, I can't use vectors with esc_beta or esc_B , but it is ok if I convert beta in d or g. It may be a limit, when I want convert numerous beta in r.

Does'nt work
esc_B(b = df_beta$beta, sdy = df_beta$SD_DV, grp1n = df_beta$Obsevations, grp2n = df_beta$Obsevations, es.type = "r", study = df_beta$unique_id)

works
esc_B(b = df_beta$beta, sdy = df_beta$SD_DV, grp1n = df_beta$Obsevations, grp2n = df_beta$Obsevations, es.type = "d", study = df_beta$unique_id)

Thanks for this handy package! I was trying to use it programmatically, and constructed a call to convert multiple effect size types at once - which then results in unexpected / unpredictable output. There is probably no meaningful way to do this in one call, but then an error would be great.

library(esc)
pearsons_r(d = 1)
#> [1] 0.4472136
# Multiple arguments result in least expected behaviour - 
# where later arguments override earlier ones.
pearsons_r(d = 1, or = 2)
#> [1] 0.1876806

^{Created on 2024-02-14 with reprex v2.0.2}

Hi Daniel & all,

Does anyone know whether group1 should always be treatment group and group2 as control? For some functions like convert_d2r and esc_B, it is stipulated to enter experimental group values as group 1 and control group as group 2. However, for functions like esc_mean_sd and esc_mean_gain, this is not stipulated.

I've been using the function esc_mean_gain entering treatment group as group1 and getting opposite results from what is anticipated. Just wondering whether there is a necessary order for these two functions and what computations they are doing to obtain the result?

Thank you for your input

Hello,

Thank you very much for the nice package. I have a question when using the package.

I was using the function esc_mean_sd() to convert means and sds to a point-biserial correlation coefficient. I wonder how the standard deviation for the Fisher's z transformation was calculated.

I learn that the formula is sqrt(1/(n-3)) from here, also from the book "Practical Meta-analysis", but when I double check the output of function esc_mean_sd(), it does not match.

Could you please help me explain why?

Thank you very much!
Best regards,
Trang Bui.

I ran the following code and got a different result than with the web-based 'Practical Meta-Analysis Effect Size
Calculator' from David B. Wilson (http://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-Home.php)

esc_2x2(
grp1yes = 205,
grp1no = 98,
grp2yes = 512,
grp2no = 178,
es.type = c("r")
)

Could you please let me know why the formula is different?
Thank you!
Diana

Hi! I am having trouble installing the package and I was hoping you could help. This is what I get in the console:

Installing package into ‘C:/Users/smith/Documents/R/win-library/3.3’
(as ‘lib’ is unspecified)
also installing the dependencies ‘haven’, ‘sjlabelled’, ‘sjmisc’


  There are binary versions available but the source versions are later:
           binary source needs_compilation
haven       1.1.1  2.0.0              TRUE
sjlabelled  1.0.8 1.0.15             FALSE
sjmisc      2.7.1  2.7.6             FALSE
esc         0.4.0  0.4.1             FALSE

  Binaries will be installed
trying URL 'https://cran.rstudio.com/bin/windows/contrib/3.3/haven_1.1.1.zip'
Content type 'application/zip' length 881609 bytes (860 KB)
downloaded 860 KB

package ‘haven’ successfully unpacked and MD5 sums checked

The downloaded binary packages are in
	C:\Users\smith\AppData\Local\Temp\RtmpQLPCZ4\downloaded_packages
installing the source packages ‘sjlabelled’, ‘sjmisc’, ‘esc’

trying URL 'https://cran.rstudio.com/src/contrib/sjlabelled_1.0.15.tar.gz'
Content type 'application/x-gzip' length 132525 bytes (129 KB)
downloaded 129 KB

trying URL 'https://cran.rstudio.com/src/contrib/sjmisc_2.7.6.tar.gz'
Content type 'application/x-gzip' length 199341 bytes (194 KB)
downloaded 194 KB

trying URL 'https://cran.rstudio.com/src/contrib/esc_0.4.1.tar.gz'
Content type 'application/x-gzip' length 52565 bytes (51 KB)
downloaded 51 KB

* installing *source* package 'sjlabelled' ...
** Paket 'sjlabelled' erfolgreich entpackt und MD5 Summen überprüft
** R
** data
** inst
** preparing package for lazy loading
Error in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = vI[[j]]) : 
  namespace 'haven' 1.1.1 is being loaded, but >= 1.1.2 is required
ERROR: lazy loading failed for package 'sjlabelled'
* removing 'C:/Users/smith/Documents/R/win-library/3.3/sjlabelled'
Warning in install.packages :
  running command '"C:/PROGRA~1/R/R-33~1.2/bin/x64/R" CMD INSTALL -l "C:\Users\smith\Documents\R\win-library\3.3" C:\Users\olderbak\AppData\Local\Temp\RtmpQLPCZ4/downloaded_packages/sjlabelled_1.0.15.tar.gz' had status 1
Warning in install.packages :
  installation of package ‘sjlabelled’ had non-zero exit status
ERROR: dependency 'sjlabelled' is not available for package 'sjmisc'
* removing 'C:/Users/smith/Documents/R/win-library/3.3/sjmisc'
Warning in install.packages :
  running command '"C:/PROGRA~1/R/R-33~1.2/bin/x64/R" CMD INSTALL -l "C:\Users\smith\Documents\R\win-library\3.3" C:\Users\smith\AppData\Local\Temp\RtmpQLPCZ4/downloaded_packages/sjmisc_2.7.6.tar.gz' had status 1
Warning in install.packages :
  installation of package ‘sjmisc’ had non-zero exit status
ERROR: dependency 'sjmisc' is not available for package 'esc'
* removing 'C:/Users/smith/Documents/R/win-library/3.3/esc'
Warning in install.packages :
  running command '"C:/PROGRA~1/R/R-33~1.2/bin/x64/R" CMD INSTALL -l "C:\Users\smith\Documents\R\win-library\3.3" C:\Users\smith\AppData\Local\Temp\RtmpQLPCZ4/downloaded_packages/esc_0.4.1.tar.gz' had status 1
Warning in install.packages :
  installation of package ‘esc’ had non-zero exit status

The downloaded source packages are in
	‘C:\Users\smith\AppData\Local\Temp\RtmpQLPCZ4\downloaded_packages’
> library(esc)
Error in library(esc) : es gibt kein Paket namens ‘esc’

Any idea what is going on?

Nice package. I have studies to meta-analyze that report mean differences between the treatment and control groups rather than the mean and SD for each group.

Is there a way to do this with the esc package?

thanks

library(esc)
esc_B(3.3, 5, 100, 150, es.type = "f")
#> 
#> Effect Size Calculation for Meta Analysis
#> 
#>      Conversion: unstandardized regression coefficient to effect size f
#>     Effect Size:   0.3481
#>  Standard Error:   0.1328
#>        Variance:   0.0176
#>        Lower CI:   1.0918
#>        Upper CI:   1.8375
#>          Weight:  56.7018

Effect size is out of CI bounds.

I recently reviewed an old script but now I got this error when I run it on a new computer with R version 4.3.1:

Error in missing(grp1sd) || is.null(grp1sd) || is.na(grp1sd) :
'length = 59' in coercion to 'logical(1)'

I wonder if the error is related to this what is said here kaz-yos/tableone#93:
"Message from the CRAN maintainer:

NEWS for R-devel has

• Calling && or || with either argument of length greater than one
now gives a warning (which it is intended will become an error)."

Hi Daniel & all,
I tried to load the esc package and get the following error message:
Error in loadNamespace(j <- i[[1L]], c(lib.loc, .libPaths()), versionCheck = vI[[j]]) :
namespace ‘sjmisc’ 2.0.0 is being loaded, but >= 2.5.0 is required
In addition: Warning message:
package ‘esc’ was built under R version 3.3.3
Error: package or namespace load failed for ‘esc’

Any suggestions for what might be going on? I am relatively new to R, so still trying to work my way around issues.
Many thanks in advance
ron

(via mail)

I noticed in the esc package that the formula for converting unstandardized b to d is incorrect as it utilizes the total sample variance/SD
rather than the pooled SD. I see the on-line calculator has the same error, but that the source document from Wilson is correct:
https://mason.gmu.edu/~dwilsonb/downloads/esformulas.pdf