At the moment I am running logistic regressions on individual data with weights. Here is my model:

model <- glm(Dummy_WBT_17~ hinc2 + tmp_b4p_bildung_quote + benefits_7 + benefits_10 + Dummy_PartyID + gender + age_rec + region_rec + urban, weights = weight, data=new3, family=binomial)

In a next step I wanted to calculate standardized regression coefficients with beta glm from 'reghelper' package . But I receive the following error:

Error in eval(extras, data, env) : object 'weight' not found

There is no error if I exclude the weights from my model. So I think the error is due to using weights in my model. Do you have any idea how to calculate standardized regression coefficients inculding the weights?

Thank you!

  • $\begingroup$ do you have an object or variable in your workspace called 'weight'? The error appears to be that r is unable to find an object with that exact name. Have you checked capitalisation and plural/singular is consistent? $\endgroup$ – ReneBt May 29 '18 at 8:45
  • $\begingroup$ I checked everything. Capitalisation and plural/singular is consistent. The variable in my data 'new3' is called 'weight'. So I wonder why R claims that there is no object 'weight'. Maybe it's a more general problem, such that you can't calculate standardized coefficients if you weight your data... $\endgroup$ – Lars Bischoff May 29 '18 at 9:42

Well the problem seems to be a little bit strange. We tried the following:

 [1] "Dummy_WBT_17"         
 [2] "hinc2"                
 [3] "tmp_b4p_bildung_quote"
 [4] "benefits_7"           
 [5] "benefits_10"          
 [6] "Dummy_PartyID"        
 [7] "gender"               
 [8] "age_rec"              
 [9] "region_rec"           
[10] "urban"                
[11] "(weights)"

It seems like the glm function renamed automatically the variable weight into "(weights)". That explains why beta function claims: "object 'weight' not found". Since it wasn't possible to rename "(weight)" into weights, we created a new variable which is simply a copy of "(weight)" and renamed it.

model$model[,12] = model$model[,11]

And now the beta function works.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.