[R-sig-ME] Looking for help to test the effect of a grouping factor in repeated measures
th|erry@onke||nx @end|ng |rom |nbo@be
Thu Sep 9 16:19:38 CEST 2021
Please note that you want the binomial distribution. Not the binomial
A model like y ~ treatment * species + (1|date) could make sense if you can
assume that the date could have a common effect on the results. E.g. the
observer being more awake on some days ;-) Use the actual date as a factor
instead of the number of days since the start of the experiment. That is
safer than using the number of days since the start.
If you assume no date effect, then there could be only temporal
autocorrelation within the species. As the strength of the autocorrelation
could be different among species, you could consider fitting a separate
model for every species.
ir. Thierry Onkelinx
Statisticus / Statistician
Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx using inbo.be
Havenlaan 88 bus 73, 1000 Brussel
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
Op do 9 sep. 2021 om 15:43 schreef Léa Fieschi-Méric <
leafieschimeric using gmail.com>:
> I have a dataset to analyse and I am struggling to build my model set and
> was wondering if you could advise me please.
> The experimental plan is as follows: several species of animals were
> observed for a year, but not regularly (some months have many observations,
> others have less). *All individuals belonging to the same species were
> housed together in the same terrarium* (unequal number of individuals per
> species). The response variable is the number of awake individuals per
> terrarium (i.e. per species). The year was divided in 3, unequal, periods
> as follows: treatment / control / treatment.
> I need to determine the effect of the *Period *(treatment / control), of
> the *Species* (12 in total), and of their *Interaction *on the *proportion
> of awake animals*. I wanted to use a GLMM with a binomial error
> distribution. The problem is that I have repeated measures per species, so
> I would like to account for that by using Species as a random effect to
> avoid pseudoreplication, but I think I need Species to be a fixed effect
> because I want to directly test its effect on my response. Therefore, I
> wonder if adding the *Date* (as an integer: number of days since the
> beginning of the experiment) to the model does control for these repeated
> measures (but I can't use it as a random effect because it is continuous,
> and it makes my model too complex to converge when integrated as a fixed
> effect), or if I should sum the response per month to use *Month* as a
> random effect ordered factor. Or if I could just use a linear model without
> bothering, like this: y ~ Treatment * Species !
> I am stuck here and would like to avoid taking a mathematically
> erroneous approach. Could you advise me please?
> I was also wondering what I need to do in order to, in a second phase, be
> able to conclude exactly which species were sensitive to the treatment and
> which were not?
> I am looking forward to hearing back from you, and thank you in advance for
> your help!
> *Léa FIESCHI-MERIC*, PhD student
> Laboratoire d'Ecologie et de Conservation des Amphibiens (LECA)
> Freshwater and OCeanic science Unit of reSearch (FOCUS)
> Université de Liège (Belgique)
> Genetics and Ecology of Amphibians Research Group
> Center for Evolutionary Ecology and Ethical Conservation
> Laurentian University (Canada)
> Tel: (+33) (0)220.127.116.11.15 <+32%204%20366%2050%2077>
> E-mail: leafieschimeric using gmail.com
> [[alternative HTML version deleted]]
> R-sig-mixed-models using r-project.org mailing list
[[alternative HTML version deleted]]
More information about the R-sig-mixed-models