No. Definitely don't.
If you do as.factor() on a covariate it creates dummy variables for you.
Note the output has 6 coefficients for education. Thus when you try to set
"educ=7" Zelig doesn't know quite you are trying to do since there is no
education variable any more- there are 6 of them.
We are assuming linearity on the effects of education and income. This
might be a terrible assumption for many reasons but its what we intended you
to do.
Brandon
On Wed, Mar 30, 2011 at 3:29 PM, Slawa Rokicki <slawa.rokicki(a)gmail.com>wrote;wrote:
Hi Brandon,
Do we have to make education and income as.factor levels for the first
question? It works to run zelig like this, but i am getting this strange
error when i try to do the first differences:
x.high <- setx(z.out, data=final.data, educ=7)
Error in `contrasts<-`(`*tmp*`, value = "contr.treatment") :
contrasts can be applied only to factors with 2 or more levels
Thanks!
Slawa
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l
_______________________________________________
gov2001-l mailing list
gov2001-l(a)lists.fas.harvard.edu
http://lists.fas.harvard.edu/mailman/listinfo/gov2001-l