[R-meta] Goodness of fit test for dose-response meta-analysis

[Michael Dewey]

Dear Marimuthu

According to the paper cited in the help file for gof() the degrees of 
freedom only take into account n and p. No mention is made of the 
covariance matrix of the random effects. The paper is well worth a look 
as it is much more extensive than what you see after ?gof

Michael

On 04/12/2024 20:19, Marimuthu S via R-sig-meta-analysis wrote:
> Dear all,
> 
> I am currently working on random effect one-stage dose-response 
> meta-analysis (DRMA), and I am trying to assess goodness of fit using 
> the R package "dosresmeta".
> 
> Here are the R code I used and the results:
> 
> *# Goodness fits statistics for Random effect quadratic polynomial model*
> library(dosresmeta)
> data(alcohol_cvd)
> 
> lin.R <- dosresmeta(formula = logrr ~ dose+I(dose^2), type = type, id = 
> id,  se = se, cases = cases, n = n,proc = "1stage", data = alcohol_cvd, 
> method="reml", control = list(maxiter =1000))
> 
>  >*gof(lin.R, fixed =FALSE)* Goodness-of-fit statistics: Deviance test: 
> D = 19.920 (df = 17), p-value = 0.278            
> *# Goodness fits statistics for fixed effect quadratic polynomial model*
> lin.F <- dosresmeta(formula = logrr ~ dose+I(dose^2), type = type, id = 
> id,  se = se, cases = cases, n = n,proc = "1stage", data = alcohol_cvd, 
> method="fixed")
>  >*gof(lin.F)* Goodness-of-fit statistics: Deviance test: D = 40.992 (df 
> = 17), p-value = 0.001
> The deviances are different for fixed vs. random but the degrees of 
> freedom are identical. Since the random effect model includes the 
> between study variance-covariance components (which should be 
> estimated),, I expected the degrees of freedom for random effect to be 
> smaller.
> 
> I would appreciate if anyone could share their thoughts.
> 
> Warm Regards,
> 
> *Marimuthu S,*
> Ph.D. Student (Biostatistics)
> 
> 
> 
> <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient>	Virus-free.www.avg.com <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient>
> 
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> 
> _______________________________________________
> 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

-- 
Michael