Democrats lead Republicans 45-37 percent on the generic congressional ballot ahead of the midterm elections in November, according to the latest Harvard CAPS/Harris poll released exclusively to The Hill.
The numbers were in line with July's survey, when Democrats were ahead of Republicans by 43-36 percent in terms of which party they would prefer in November.
The findings also showed Democrats had a 47-31 percent advantage among female voters in the generic ballot question.
"Democrats continue to have an edge in the midterms driven by a widening gender gap as the Democrats have gained considerably with women voters," said Harvard CAPS/Harris Poll co-director Mark Penn.
http://thehill.com/homenews/campaign-polls/403491-democrats-continue-to-hold-lead-on-generic-ballot-ahead-of-midterms-----------
Let me outline why I'm skeptical of these polls by comparing their sample assumptions against actual voter turnout here in Iowa in '14
Here are Harvard-Harris':
Here is Iowa:
Notice the discrepancies:
1. Reps outvoted Dems by 12 points, outvoted Indies by 31 points, or 39% R, 33% D, 26% I, compared to 32% R, 37% D, 29% I in the poll.
2. Age demographics are totally skewed. Actual voting was- 18-34:15%, 35-49: 20%, 50-64: 33%, 65+: 32% in Iowa v. 26/26/28/20% for the poll.
3. Suburban voters swamped the sample. I can tell you by many years experience that unless you have a district dominated by a large metro area, rural voters drastically outvote urban/ suburban voters.
4. Ethnic demographics won't hold up district to district, state to state. Maybe in the wash, but each competitive district would have to be broken down to determine it's makeup.
Most polls are sampling this way right now. Compared to the actual vote, they represent a very Democrat friendly scenario. If it doesn't hold true, they will get stomped in Nov.
