Skip to content

subset factor?

4 messages · Silvano, Sarah Goslee, Peter Ehlers +1 more

Silvano
# 
Hi,

I used subset command, like this:

grupoP = subset(dados, grupos=='P',  select=c(mortos, vivos, 
doses, percevejos, p))

and the variables in select option are numeric.

I tried fit a model with command:

ajuste.logit = glm(cbind(mortos,percevejos)~log10(doses), 
family=binomial(logit), data=grupoP)

and the output is:
Erro em Math.factor(doses) : log10 not meaningful for 
factors

what is wrong with my commands? Why doses aren't numeric?

Thanks,

--------------------------------------
Silvano Cesar da Costa
Departamento de Estat?stica
Universidade Estadual de Londrina
Fone: 3371-4346
Sarah Goslee
# 
Are you certain they are numeric? Have you tried
str() on your data?

Sarah
On Fri, Jan 14, 2011 at 9:08 AM, Silvano <silvano at uel.br> wrote:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Peter Ehlers
    

    

      
      #
    

  


  

      
On 2011-01-14 06:20, Sarah Goslee wrote:

            

      
      
      
    

        
Silvano:

As Sarah says, do str(dados) to see that 'doses' is
NOT numeric. The error message is very clear; it tells you
that 'doses' is a factor. My guess is that you imported
your data from somewhere and somehow managed to import
'doses' as a factor. My advice is to ALWAYS do an str()
after importing/changing data.

Peter Ehlers

            

      
      
      
    

  
  


  


      

    

    

      
      

        


  

        

      Erik Iverson
    

    

      
      #
    

  


  

      
Silvano wrote:

            

      
      
      
    

        
They may *look* numeric, but are they really?  You don't give us
enough information to determine that.  Either give us a sample
of your data with the dput command, or the output of str(grupoP).