We tested habitat preference and competition of two species by placing them in 'arenas' and monitoring their position every 20 mins- recording the the habitat in which the individual was found. This was repeated for 6 arenas for 24 hours each.
6 arenas:
2 arenas with 16 individuals of sppA
2 arenas with 16 individuals of sppB
2 arenas with 8 + 8 of both species A&B
Released the same time and given time to aclimatise before start the clock.
Every 20 mins record the location of each individual.
The research questions:
1) does sppA show habitat preferences?
2) is the presence of sppB affecting sppA habitat preference?
(repeat questions for spp B or is it easier to test them together?).
See attached example data.
How do I answer my two research questions with this data?
Second step - dealing with issues:
If I want to use a chi square I need to add up the total counts in each habitat for all individuals pooled together? But I have several individuals - so the data is nested or repeated.. How do I deal with this?
Missing Data: In some arenas, individuals were not seen (hidden) in some observations.. so the total was not always exactly 16 individuals in all arenas and the no. of observations for each individual was not always every 20mins for 24hrs.
Thanks
Olá, Tânia. Depende do que você quer analisar. Existem vários tipos de teste Qui-quadrado, o mais tradicional te dá a resposta se existe uma relação de associação entre suas variáveis. Esse teste te informa se seu X e Y apresentam ou não relação de associação, e no SPSS (statistical package for social science) tem-se como H0: não apresenta relação e Ha: Apresenta relação de associação.
Com isso você irá verificar se especie e habito alimentar possuem relação de associação ou não.
Ou também se arena e habito alimentar possuem associação ou não
Ou ainda se arena e especie possuem relação de associação ou não.
Pelo o que entendi você possui 3 tipos de arenas, duas espécies animais e observa se eles se alimentam sozinho ou não, me parece que você vai ter 2 variáveis dicotômicas: especie 0 e especie 1 e também se alimenta sozinho (1) ou não (0). É isso?
-----------------------------------------------------------
Hi, Tanya. It depends on what you want to analyze. There are several types of Chi-square test, the more traditional test gives you the answer if there is an association relationship between your variables. This test tells you if your X and Y have or do not have an association relation, and in SPSS (statistical package for social science) you have as H0: no relation and Ha: It presents association relation.
With this you will check if species and food habits have association relationship or not.
Or if the arena and food habit have an association or not
Or even if arena and species have an association relation or not.
From what I understand you have 3 types of arenas, two animal species and see if they feed alone or not, I think you will have 2 dichotomous variables: species 0 and species 1 and also feeds alone (1) or not ( 0). It is right?
I'd probably think about this in terms of a multilevel generalised linear model (also called linear mixed models) of the binomial observations or possibly of counts. Simple chi-square tests don't tend to extend well to repeated measures (though one can think of these multilevel generalised linear models as generalisations of simpler chi-square models it is probably more useful to start with an understanding of logistic or count (e.g., Poisson) style models).
If I understand it correctly, for each species type (sspA, sspB or sspA&B), you had 16 individuals. You recorded/counted how many of these were found in which of the 3 habitats in the arenas. This was repeated 3 time per hour, during 24 hours. This would imply that the maximum is (ignoring the 'missings') 16 x 3 x 24 = 1152 counts.
I think you can run a chi-square test on the totals of the observed counts per habitat. See the example in the R script in attachment.
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training