I have a one year data of catch per unit effort (i.e. fisheries data). The data contains many true zero values and it is right skewed. Thus, I need to run delta Log-normal GLM in order to compare the catch per unit effort (CPUE) across sampling locations and seasons. The dependent variable is CPUE in Kg , whereas, the independent variables are location (four sampling sites) and season (dry and wet). However, I faced a problem to run this in SPSS. Therefore, is there anyone who can suggest me the procedures please?
You may want to try one of the zero-inflated models. I have a preference for the so-called hurdle model, for which there is a single source of 0's: you jointly model the proba to get a non-null catch (the source of 0's), and the catch (> 0). Each sub-model may have its own covariates. See http://ftp.uni-bayreuth.de/math/statlib/R/CRAN/doc/vignettes/pscl/countreg.pdf
Dear Renaud Lancelot I thank you very much for the suggestion and attached link. It is very useful to me.
Dear Ette Etuk
Thank you very much for your information. Could you tell me something about how can I run this things in r.
Hi Gebremedhin, regards. Sorry, did you manage to get a routine in R to run a model incorporating the zero values?
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology