[R-sig-ME] Time as both fixed and random term

[Pantelis Hadjipantelis]

Hello Joe,

I think that discretising a continuous variable can be unnatural and 
make a model hard to interpreter. AIC while extremely helpful is not a 
panacea.
Unless I had distinctly clustered measurement times and I could 
additionally assume that the time-dependence of the data is very 
weak/non-existent I would actively avoid treating time as factor.

Best regards,
Pantelis

On 11/24/2015 07:08 PM, shi_peijian wrote:
> Dear all,
>
>
> Could I additonally ask a question?
>      fit1 <- Biomass ~ Treatment + Time + (1|Plot), where 'Time' is a continuous covariate
>      fit2 <- Biomass ~ Treatment  + (1|Time/Plot), where 'Time' is a factor variable
>
>
>      fit3 <- Biomass ~ Treatment + (1|Time), where 'Time' is also a factor variable
>
>
> If AIC(fit3) is smaller than AIC(fit1) and AIC(fit2), can we choose fit3 rather than fit1?
>
>
> Thanks a lot!
>
>
> Best regards,
>
>
> Joe
>
>
>
> --
>
>
> Peijian (Joe)  Shi, Ph.D.
>
> Research interests: forest ecology; theoretical ecology; thermal biology
>
> Member of China Ornithological Society from 2005 up to the present
>
> Bamboo Research Institute, Nanjing Forestry University, P.R. China
>
> 159 Longpan Road, Nanjing City, Jiangsu Province 210037
>
> Office:  60817  Biotechnology Building
>
> Tel:  +86 25 85427231
>
> E-mail addresses:  peijianshi at gmail.com
>
>                                 shi_peijian at 163.com
>
>
>
>
> At 2015-11-25 06:06:51, "Lionel" <hughes.dupond at gmx.de> wrote:
>> Dear List,
>>
>> In my work we usually deals with measures sampled repeatedly on
>> experimental units over several time points and with specific
>> treatments. For example we sampled plant biomass on 100 experimental
>> plots at 5 different time point (say season or year). Some people argue
>> that in this context we should model time as both a fixed effect term
>> (as continuous variable) and random effect term in order to compute the
>> correct numbers of degrees of freedom to test our treatment effects
>> (usually considered as a continuous variables).
>>
>> This is how such a model would look like:
>>
>> Biomass ~ Treatment + Time + (1|Plot) + (1|Time)
>>
>> In my experience having the same term has both fixed and random results
>> in very low estimated standard deviation for the random term, which
>> makes me skeptical about this approach. But having very little knowledge
>> about how to correctly estimate the numbers of degrees of freedom I
>> would like to ask you:
>>
>> (i) if such a model makes sense,
>> (ii) if the argument "we need to have time as both fixed and random term
>> to get the correct number of degrees of freedom" is valid
>> (iii) if such an alternative model: "Biomass ~ Treatment + Time +
>> (1|Plot)" would be more appropriate.
>>
>> Thanks for your input,
>> Lionel
>>
>> _______________________________________________
>> R-sig-mixed-models at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models