[R-sig-ME] Equivalent of covariance pattern model with a heterogeneous compound symmetric

[Pardis Miri]

Hi Ben,

The example I am trying to construct is a 3 x 2 x 3 within-subject design.
I have the following column in my data set
 trial 1:18, Bodysite with 3 levels, Strength with 2 levels,  Pattern with 3 levels, and subject id.
I am interested in SC_slope as a DV.

I just contracted the following but I am not sure if I followed your example correctly. Is this correct?


comsym <- gls(SC_slope~Bodysite+Pattern+Strength,data = df.na.removed.,
                   correlation=corCompSymm(form = ~ trial |id), method="REML", na.action = na.omit,
              weights = varIdent (form = ~1 | trial))

summary(comsym)

Thank you again!
Paris


On Thu, May 30, 2019 at 12:59 AM Ben Pelzer <b.pelzer using maw.ru.nl<mailto:b.pelzer using maw.ru.nl>> wrote:
Hi Pardis,

Four say 4 occasions you could use:

HCS.gls <- gls (read ~ 1+occ2+occ3+occ4, data=DA68, method="REML",

correlation = corCompSymm(form = ~ occasion|id ),

weights = varIdent (form = ~1 | occasion) )



Best regards, Ben.


On 30-5-2019 4:57, Pardis Miri wrote:
> Hi everyone,
>
> I am trying to figure out how to replicate the results I've gotten from SPSS when using the Repeated Covariance Type : Compound Symmetry: Heterogeneous.  Can anyone help me figure this out?
> So far I know the following does't produce the same results.
>
> comsym <- gls(SC_slope~Bodysite+Pattern+Strength,data = df.na.removed.,
>                     correlation=corCompSymm(form = ~ 1 |id), method="REML", na.action = na.omit)
>
> summary(comsym)
>
>
>       [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models using r-project.org<mailto:R-sig-mixed-models using r-project.org> mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models


        [[alternative HTML version deleted]]

_______________________________________________
R-sig-mixed-models using r-project.org<mailto:R-sig-mixed-models using r-project.org> mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

	[[alternative HTML version deleted]]