[R] Including entire dataframe in svyglm formula

cerman at u.washington.edu cerman at u.washington.edu
Tue Mar 16 07:01:57 CET 2010

I am attempting to use svyglm {survey package} with a large set of predictors and would like to be able to automatically use all the variables in my data as the predictors without having to type out their names explicitly, similar to the way one can do:

y = x1+x2+rnorm(4,0,1)
xdataframe = data.frame(x1,x2)
lm(y~., data=xdataframe) 

and the period or dataframe name will cause all predictors to be included. I have been trying that with the following code (and all sorts of permutations thereof):

svd = svydesign(id =~PSUSCID, weights=~GSWGT1, data=data.frame(cbind(y1, xfull)))
model = svyglm(y1~., data=svd$variables[,-1], design=svd)

where y1 is a vector of my independent variable and xfull is a matrix of my independent variables. I receive the following error:

Error in svyglm.survey.design(y1 ~ ., data = svd$variables, design = svd) : 
  all variables must be in design= argument

Is there any way I can avoid explicitly typing out the names of all the predictors in my formula in svyglm? (I have 60+.) Thanks so much!

-nate cermak
student, university of washington

More information about the R-help mailing list