[R-meta] converting escalc ROM outputs to the log-2 scale

Viechtbauer, Wolfgang (SP) wo||g@ng@v|echtb@uer @end|ng |rom m@@@tr|chtun|ver@|ty@n|
Thu Dec 17 10:35:51 CET 2020

Dear Divya,

That's just how variances change under multiplicative transformations. The same works with raw data:

x <- rnorm(10)
2^2 * var(x)


>-----Original Message-----
>From: Divya Ravichandar [mailto:divya using secondgenome.com]
>Sent: Saturday, 12 December, 2020 1:10
>To: Viechtbauer, Wolfgang (SP)
>Cc: r-sig-meta-analysis using r-project.org
>Subject: Re: [R-meta] converting escalc ROM outputs to the log-2 scale
>Hi Dr. Wolfgang
>Thank you for getting back to me. Could you help me understand more on why
>we square the log(2) in the vi calculation?
>On Fri, Dec 11, 2020 at 6:37 PM Viechtbauer, Wolfgang (SP)
><wolfgang.viechtbauer using maastrichtuniversity.nl> wrote:
>Dear Divya,
>If y_i ~ N(theta_i, v_i), then for any constant c, c*y_i ~ N(c*theta_i, c^2
>* v_i).
>So, it would be yi_new = yi/log(2) and vi_new = vi/log(2)^2.
>>-----Original Message-----
>>From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces using r-
>>On Behalf Of Divya Ravichandar
>>Sent: Friday, 11 December, 2020 23:36
>>To: r-sig-meta-analysis using r-project.org
>>Subject: [R-meta] converting escalc ROM outputs to the log-2 scale
>>Hi all
>>I have a question around changing the scale of *yi *and *vi *exported
>>from *escalc()
>>*function when using *measure=ROM.* Based on the documentation, this
>>function exports *yi *in natural log (ln) scale. --> yi =
>>ln(grp1-mean/grp2-mean) and then the associated vi.
>>I am interested in obtaining the *log-2 fold change. *Is there a way to
>>convert yi & vi from *escalc(measure = ROM,...) *to this scale. Would it be
>>reasonable to simply transform the values by converting from log-e -> log-2
>>scale by
>>yi_new = yi/log(2) and vi_new = vi/log(2)
>>Thank you
>>*Divya Ravichandar*
>>Second Genome

More information about the R-sig-meta-analysis mailing list