Re: running mixed models
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: running mixed models

Dear ROdney,
I don't understand all your question.
> When running a mixed model analysis for the combined analysis of MET with 
> the variables year, district, and year ( district as fixed with constraints 
> to zero
> and varieties, varieties ( district, varieties ( year, and varieties ( 
> district ( year  as random is this the proper model to use?
> X !WT=T~mu  c(Year) c(District) District.Year !r Variety District.Variety 
> District.Year Variety.District.Year !f mv
This is a reasonable model although I would normally include 'Trial'
as a fixed effect (which may equate in your case to Year.District
if there is only one trial in each district in a given year.)

This model is based on equal variances across trials and years
which for some crops is not common.  Wheat can easily generate
a 10 fold difference in means and 100 fold difference in variances.
Therefore, I would be looking at fitting more sophisticated variance

> I did run a similar analysis using c(Year)*c(District), but for some reason 
> dumped the error in the Blup-se for the last year.  Is there an explanation 
> for this?

I have no idea what this means.  You should bet the same variance
components and residual degrees of freedom.  The only difference should be
the fixed effects because they are parameterised differently.  If there
are holes in the Year.District table, especially in the 'last' Year or District,
singularities will make the solutons for c(Year), c(District) and
c(Year).c(District) almost uninterpretable.  The unconstrained 
terms are preferred in this situation.

> Rodney
> --
> Asreml mailinglist archive:

Arthur Gilmour PhD        
Principal Research Scientist (Biometrics)            fax: <61> 2 6391 3899
NSW Agriculture                                           <61> 2 6391 3922
Orange Agricultural Institute             telephone work: <61> 2 6391 3815
Forest Rd, ORANGE, 2800, AUSTRALIA                  home: <61> 2 6362 0046

ASREML is still free by anonymous ftp from pub/aar on
    or point your web browser at 

To join the asreml discussion list, send the message  

To send messages to the list,

Asreml list archive:

                        <> <> <> <> <> <> <>
"Christ Jesus came into the world to save sinners"   I Timothy 1:15.


Asreml mailinglist archive: