[R-meta] F test vs QM test for test of moderators

[Huang Wu]

Hi Yefeng,

Thank you very much for your reply.

In other examples on the metafor package website, the results of Test of
Moderators are QM values (e.g., *QM*(df = 3) = 15.9842, p-val = 0.0011) but
mine was F values. I wonder if the QM test was based on chi-square
distribution, and the reason that causes this difference. I would highly
appreciate your feedback.

Best regards,
Huang




On Sat, Apr 29, 2023 at 12:49 AM Yefeng Yang <yefeng.yang1 using unsw.edu.au>
wrote:

> Dear Huang
>
> If you understand what QM test is, it is not difficult to find out "Test
> of Moderators (coefficients 2:3)" printed below your output is the QM test
> results. Your test statistic was tested against F distribution. You can
> also use chi-square distribution (but not recommended). In essence, QM
> test is a sort of omnibus test or joint null-hypothesis test.
>
> Best,
> Yefeng
>
>
> ------------------------------
> *From:* R-sig-meta-analysis <r-sig-meta-analysis-bounces using r-project.org>
> on behalf of Huang Wu via R-sig-meta-analysis <
> r-sig-meta-analysis using r-project.org>
> *Sent:* Saturday, 29 April 2023 12:34
> *To:* R meta <r-sig-meta-analysis using r-project.org>
> *Cc:* Huang Wu <wuhuang0421 using gmail.com>
> *Subject:* [R-meta] F test vs QM test for test of moderators
>
> Dear all,
>
> I am writing to ask about the Test of Moderators in meta-analysis.
> Specifically, I am curious about the appropriate test to use between the F
> test and QM test. I ran the following code and obtained results using the F
> test for the Test of Moderators. However, I would like to explore how to
> obtain QM test results.
>
> Could you kindly advise me on the suitable test to use and how to obtain QM
> test results using the metafor package?
>
> Thank you for your assistance.
> Huang
> ----------------------------------------------------
>
> USnew_c.Dnoctl.model <- rma.mv(yi=effect_d, #effect size
>                                V = VUSnew_c.Dnoctl, #variance (tHIS IS WHAt
> CHANGES FROM HEmodel)
>                                mods = ~ grade_level,
>                                random = ~1 | ID/eid, #nesting structure
>                                test= "t", #use t-tests
>                                data=USnew_c.Dnoctl, #define data
>                                method="REML") #estimate variances using
> REML
>
> summary(USnew_c.Dnoctl.model)
>
> ----------------------------------------------------
> Multivariate Meta-Analysis Model (k = 142; method: REML)
>
>   logLik  Deviance       AIC       BIC      AICc
>  16.9680  -33.9361  -23.9361   -9.2637  -23.4849
>
> Variance Components:
>
>             estim    sqrt  nlvls  fixed  factor
> sigma^2.1  0.0300  0.1733     19     no      ID
> sigma^2.2  0.0161  0.1270    142     no  ID/eid
>
> Test for Residual Heterogeneity:
> QE(df = 139) = 344.3018, p-val < .0001
>
> Test of Moderators (coefficients 2:3):
> F(df1 = 2, df2 = 139) = 1.1327, p-val = 0.3251
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-meta-analysis mailing list @ R-sig-meta-analysis using r-project.org
> To manage your subscription to this mailing list, go to:
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis
>

	[[alternative HTML version deleted]]