There must be something else happening so you might have to send me the example.
From your description, there are only 9 patterns of Year/Treat in the data.
i.e. Year Treat
I.e. There are NO Treat 1 and Treat 2 associated ith years 1-6,8
and NO missing trat for year 7
It should not matter whether the * is 0, the result should be
And you would get the same answer if the last line was just 'treat'
or Year.treat or c(treat) or Year.c(treat)
If however, there are additional combinations, you will get
additional degrees of freedom.
> X-Authentication-Warning: lamb.chiswick.anprod.csiro.au: petidomo set sender
to email@example.com using -f
> Date: Thu, 27 Jan 2000 14:19:19 +1100
> To: firstname.lastname@example.org
> From: Kim Bunter <email@example.com>
> Subject: DF's with nested effects
> Mime-Version: 1.0
> Hi all,
> I am having trouble coming up with the right DFs (and apparently labels in
> the solution file) for nested effects. For example, I have 8 years of data
> and a treatment (2 levels) has been applied in year 7 only. I have tried
> fitting this as:
> trait ~ mu year at(year,7).(treat)
> which gives me 7DF for year and still 2DF for treatment. The problem seems
> to arise due to missing values for treatment - even though these only occur
> in the years where this effect is not fit, and there are no missing values
> in the year where this factor is fit. I can (apparently?) fix it by making
> sure the factor levels are zero rather than missing. Well, at least the DF
> for the treatment become 1 which is what I expected. Previously defining
> the treatment seems to have no impact. It is not an option to delete the
> records with missing treatments unless I want to throw out an entire years
> data (which I don't).
> However, I am not sure that making factor levels not missing (but all
> zeros) is actually the correct solution. For both the above cases in the
> solution file the label seems to show that treatment is fit within year 1
> (which can't be right as all factors in year 1=0).
> at(myear,7).csupp 1.0001 0.000 0.000
> at(myear,7).csupp 1.0002 7.248 4.146
The fact that the first solutuion is 0.000 +- 0.000 is intruiging.
It is apparently saying there is no effect [which is what you expect]
but has not been detected as missing.
It may just be a failure to detect the singularity [ although I do not
why coding csupp 0 instead of * should make any difference.
You may be able to confirm the madel by hand calculating some
predicted values [not hard if you do not have a more complicated model].
Is the example huge?
> As the data is ordered from 1 to 8, I don't think the confusion is arising
> for year. Why does the label not state that the appropriate year is 7?
The 1 in 1.0002 refers to the first level of the factor at(myear,7)
[there is only one level]. I hadn't thought of trying to call this year 7
> What am I missing here?
I can't see what is happening yet. Please send me the example.
> Kim Bunter (M.Rur.Sc)
> PhD Student
> Animal Genetics and Breeding Unit
> University of New England
> Armidale, NSW, 2351
> Ph: (02) 6773 3788
> Fax: (02) 6773 3266
> email: firstname.lastname@example.org
> Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml
Arthur Gilmour PhD mailto:Arthur.Gilmour@agric.nsw.gov.au
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 ftp.res.bbsrc.ac.uk
or point your web browser at ftp://ftp.res.bbsrc.ac.uk/pub/aar/
To join the asreml discussion list, send the message
To send messages to the list, mailto:email@example.com.CSIRO.au
Asreml list archive: http://www.chiswick.anprod.csiro.au/lists/asreml
<> <> <> <> <> <> <>
"Christ Jesus came into the world to save sinners" I Timothy 1:15.
Asreml mailinglist archive: http://www.chiswick.anprod.csiro.au/lists/asreml