[R-meta] lmer, SAS and rma.mv
Viechtbauer, Wolfgang (SP)
wolfgang.viechtbauer at maastrichtuniversity.nl
Sun Apr 15 18:52:13 CEST 2018
Thanks. I thought SAS would be clever enough to use 10000 for all fixed effects, but I guess one has to explicitly repeat this for each one.
As for the fit statistics:
rma.mv() computes the 'full' residual log likelihood, while SAS omits one term (that does not involve the parameters). If you want to do the same with rma.mv(), then use:
rma.mv(..., control=list(REMLf=FALSE))
Then the ll and the deviance should be the same. However, the AIC, AICc, and BIC will still differ. SAS does not count the fixed effects as parameters (when using REML estimation), so for the AIC, it computes 2177.0 + 2*2 = 2181.0. On the other hand, rma.mv() does count the fixed effects, so you would get 2177.0 + 2*4 = 2185.0.
Best,
Wolfgang
-----Original Message-----
From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces at r-project.org] On Behalf Of Martineau, Roger
Sent: Friday, 13 April, 2018 14:58
To: r-sig-meta-analysis at r-project.org
Subject: Re: [R-meta] lmer, SAS and rma.mv
Thanks Dr Viechtbauer,
I asked our statistician (Steve Méthot) to run the suggested model in SAS for me and he had to make 1 modification to get P-value and CIs for x1 (not reported in the output otherwise).
model y = intcpt x1 / solution cl noint ddf=10000;
was changed for :
model y = intcpt x1 / solution cl noint ddf=10000,10000;
We got the exact same results except for fit statistics which differ a little bit but I guess this depends on the way each program calculates them.
Fit Statistics
-2 Res Log Likelihood 2177.0
AIC (Smaller is Better) 2181.0
AICc (Smaller is Better) 2181.0
BIC (Smaller is Better) 2185.7
Multivariate Meta-Analysis Model (k = 192; method: REML)
logLik Deviance AIC BIC AICc
-1083.9130 2167.8260 2175.8260 2188.8141 2176.0422
Thanks again,
roger ☺
Roger Martineau, mv Ph.D.
Centre de recherche et de développement
sur le bovin laitier et le porc
Agriculture et agroalimentaire Canada/Agriculture and Agri-Food Canada
Téléphone/Telephone: 819-780-7319
Télécopieur/Facsimile: 819-564-5507
2000, Rue Collège / 2000, College Street
Sherbrooke (Québec) J1M 0C8
Canada
roger.martineau at agr.gc.ca
More information about the R-sig-meta-analysis
mailing list