[R-sig-ME] mixed model for nested factor?

[Thierry Onkelinx]

In this case, use the coding as I suggested. The interaction with season
will take care of the difference in warming.

Is a linear trend along year a reasonable assumption? If yes, you can use
year as continuous. If no, use it as a factor.

ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey

2017-05-09 4:25 GMT+02:00 Jinsong Zhao <jszhao op yeah.net>:

> Thank you very much for the reply. The 2, 1 and 0 are not the amount of
> warming. They are just codes for the warming strength. Since the warming
> pattern is different in Spring and Fall, I wonder that warming pattern
> (Temp) is nested in Season.
>
> Another concern is how to deal with Year. I hope to test the time effect
> on SOC. Should Year be treated as continuous variable or factor? I recorded
> the exact time of soil sampling.
>
> Thanks again.
>
> Best regards,
> Jinsong
>
>
> On 2017/5/9 2:35, Thierry Onkelinx wrote:
>
>> Dear Jinsong,
>>
>> IMHO you want to detect a potential effect of warming. The amount of
>> warming seems to be less relevant. Therefore I'd recode temperature into a
>> factor with two levels: warming = yes and warming = no.
>> Then add the interaction between warming and season to the model. That
>> will
>> take care of the difference in warming between spring and fall.
>>
>> Best regards,
>>
>> ir. Thierry Onkelinx
>> Instituut voor natuur- en bosonderzoek / Research Institute for Nature and
>> Forest
>> team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
>> Kliniekstraat 25
>> 1070 Anderlecht
>> Belgium
>>
>> To call in the statistician after the experiment is done may be no more
>> than asking him to perform a post-mortem examination: he may be able to
>> say
>> what the experiment died of. ~ Sir Ronald Aylmer Fisher
>> The plural of anecdote is not data. ~ Roger Brinner
>> The combination of some data and an aching desire for an answer does not
>> ensure that a reasonable answer can be extracted from a given body of
>> data.
>> ~ John Tukey
>>
>> 2017-05-08 16:13 GMT+02:00 Jinsong Zhao <jszhao op yeah.net>:
>>
>> Hi there,
>>>
>>> We have conducted an experiments for five years. The experiment is a
>>> completed block design with repeated measures. In detail, there is only
>>> one
>>> factor, i.e., temperature, and the experiments is performed in four
>>> blocks.
>>> In each block, the plot was split into two subplots. One subplot received
>>> two degree warming in Spring, and one degree warming in Fall, the other
>>> subplot didn't received any changes in Spring and Fall. We measured SOC
>>> every Spring and Fall. So we have got data something like:
>>>
>>> SOC, Temp, Season, Year, block
>>> 2.032, 2, Spring, 2007, 1
>>> 1.988, 0, Spring, 2007, 1
>>> 1.977, 1, Fall, 2007, 1
>>> 1.881, 0, Fall, 2007, 1
>>> ....
>>> ....
>>> It seems that mixed model might be the right way to analyze the data.
>>> However, we have encountered the difficulties that Temp seems to be
>>> nested
>>> with Season, and don't know how to code a formula for lme() in nlme
>>> package.
>>>
>>> Any suggestions or comments will be really appreciated. Thanks in
>>> advance.
>>>
>>> Best regards,
>>> Jinsong
>>>
>>
> _______________________________________________
> R-sig-mixed-models op r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>

	[[alternative HTML version deleted]]