I think poll watchers would do well to accept that the field is in an era of transition, and one consequence of that is a frustrating amount of noise in subgroup estimates.

Your name:*Required

Your email address:*Required

Pick a user name:*Required