I have behavioral data (feeding latency) which is the dependent variable. There are 4 populations from which the behavioral data is collected. So population becomes a random effect. I have various environmental parameters like dissolved oxygen, water velocity, temperature, fish diversity index, habitat complexity etc. as the independent variables (continuous). I want to see which of these variables or combination of variables will have significant effect on the behavior.
I think Dr.Paul R. Yarnold explaine and referate very good the problem. You can use the models regression linear as( ALM) and no linera as neural network and evaluate the accurate model.
I agree with Paul to run separate analyses first. If the response variable ( or rather, the errors) are non-normal, consider transforming it to a normal quantile first. Consider adding three indicator variables for the four populations, and then observe the role of these indicator variables in the regression. Are they significant?
Separately, consider doing a factor analysis. This will help you in identifying combinations of environmental factors, and you may be able to fi d out even more useful thi gs. I would normalize all variables before running the factor analysis so that no variable dominates the analysis.
good luck.
Thank you all for the suggestions. There are two points to worry about and they are multicollinearity and just 4 data-point for four populations. The independent variables used in my studies should not have been collinearly connected. But that is not happening because these are environmental cum ecological factors which are interdependent and hence there is a correlation between many of the. Also, I just have 4 data points from 4 populations which does not allow me to form any distribution necessary for building a kind o f model. Please suggest.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling