Error in glm.nb : if (any(y < 0)) stop("negative values not allowed for the 'Poisson' family") : missing value where TRUE/FALSE needed

[bethevex]

Hello, I am running a negative binomial model using the glm.nb function from the MASS package. When I run the model the following error appears:

Error in if (any(y < 0)) stop("negative values not allowed for the 'Poisson' family") :
missing value where TRUE/FALSE needed
In addition: Warning messages:
1: In model.response(mf, "numeric") :
using type = "numeric" with a factor response will be ignored
2: In Ops.factor(y, 0) : ‘<’ not meaningful for factors

My covariant of interest is numeric count data from 0-1900. I have converted this into numeric using the as.numeric function. My model looks like this:

model <- glm.nb(non_house_contact ~ wave, data = df)

I am not sure why this error is occurring. Any help would be appreciated. Thanks.

[nirgrahamuk]

it seems your conversion to numeric didn't 'take'
Perhaps share the code that you had used for that purpose ?

[bethevex]

Hi, I used the code below to convert to numeric:

as.numeric (non_house_contact)

Here is the str()

$ Wave             : chr [1:7] "wave1" "wave3" "wave5" "wave3" ...
 $ Non_house_contact: num [1:7] 2 1 7 1 1 4 100
 $ Sex              : chr [1:7] "F" "F" "F" "M" ...
 $ Caring           : chr [1:7] "Yes" "Yes" "Yes" "No" ...
 $ Well being       : chr [1:7] "A little" "A lot" "A little" "No change" ...

Thanks!

[nirgrahamuk]

The mysteries are multiplying....

as.numeric (non_house_contact)

this code, has no apparent relation to 'df' which is the object passed to glb.nm in your example.
furthermore the str() if it is of df appears to show a different capitalisation Wave rather than wave etc, and R is case sensitive and will see these as different.

Hopefully these clues will be enough for you to progress.
If not, then you could get additional help by creating a more complete reprex, which you can read about how to do here

[bethevex]

Maybe that is where I am going wrong then! Thanks