[R-meta] P-values one or two-sided

Tina Dudenhöffer t|n@@dudenhoe||er @end|ng |rom gm@||@com
Mon Jun 28 15:43:28 CEST 2021


Thank you, all clear.

And yes, I was referring to the QM-test.

Thanks for the quick response,
Tina

> On 28 Jun 2021, at 15:40, Viechtbauer, Wolfgang (SP) <wolfgang.viechtbauer using maastrichtuniversity.nl> wrote:
> 
> Dear Tina,
> 
> Please see my responses below.
> 
> Best,
> Wolfgang
> 
>> -----Original Message-----
>> From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces using r-project.org] On
>> Behalf Of Tina Dudenhöffer
>> Sent: Monday, 28 June, 2021 15:32
>> To: r-sig-meta-analysis using r-project.org
>> Subject: [R-meta] P-values one or two-sided
>> 
>> Dear all,
>> 
>> I am working with the Metafor package using the rma.mv function.
>> 
>> I read that, by default, the p-values are reported for two-sided tests, so I am
>> wondering whether I should divide my p-values by two.
>> 
>> I am asking about the test of moderators specifically (when performing the Q-test
>> of Moderation).
> 
> The p-value given there is a one-sided p-value (since this is for a chi-square / F-test). There is no need to adjust this p-value.
> 
>> Similarly, when removing the intercept (i.e putting a -1 behind the moderator) I
>> assume that p-value is not two-sided?
> 
> You mean for the QM-test? That's always reported as a one-sided p-value.
> 
>> I am just not sure how much of the output is depending on R default and how much
>> is part of the Metafor package and how exactly it is set up.
>> 
>> Thanks for clarifying.
>> 
>> Kind regards,
>> Tina



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