0

I would like to obtain a general p-value for each of my model variables and not an specific p-value for each level of every variable of my model. I think it is possible to obtain a p-value by performing a chi-square test but I do not know how to do it with my model variables. I would need this p-value to show the variable significance when I publish the results.

The two independent variables of the model are factors. This is my model:

 model <- zeroinfl(data=abundanceAfter ,
   formula = total ~ sav_open+period|1, dist = 'poisson')
Ben Bolker
  • 211,554
  • 25
  • 370
  • 453
Daniel Sánchez García
  • 1
  • 1
  • 1
    It's easier to help you if you include a simple [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) with sample input and desired output that can be used to test and verify possible solutions. What type of test do you want to perform to get a p-value? A likelihood ratio test? – MrFlick Jan 23 '19 at 20:00
  • I'm guessing that the p values you are talking about refer to the output for categorical predictors from `summary(model)`. Try running `drop1(model)` or `anova(model)`. – ekstroem Jan 23 '19 at 20:44
  • @ekstroem you are right. That you say about using `drop1(model)` or `anova(model)` is that i want, but this functions do not work with `zeroinfl` function. – Daniel Sánchez García Jan 24 '19 at 21:13
  • @MrFlick Yes, it is possible that performing a likelihood ratio test is that i am looking for, but i just know how to do it to compare models, not to obtain the p-values of my variables. – Daniel Sánchez García Jan 24 '19 at 21:16
  • 2
    fit a reduced model with one variable missing. Then the likelihood ratio test p-value for the comparison between the full and reduced models will be the p-value for that variable. – Ben Bolker Jan 24 '19 at 21:48

0 Answers0