Polls of important counties

[mb123]

    I am trying to put together an electoral college map. I was trying to place who was going to win ohio and I was looking at 538's projection then over at the atlas map for 2012 general election. It seems like ohio was carried for Obama by like 3 counties. Considering how many people live in Cuyahoga county [645k people] and since ohio is such a bell-weather for if a republican will be in the white house why arent there polls of this place? It would be nice to find some details of how different demographics are feeling this year vs 2012 in some of these particular counties that have huge amounts of the votes in important battle ground states right?

[LLR]

All polls of counties, congressional districts, and the like have tiny sample sizes and are meaningless.

[HillOfANight]

There are a few internal poll at the county level if you look around. Here's one of Hillsborough, without cell phones.

http://www.tampabay.com/blogs/the-buzz-florida-politics/poll-donald-trump-edging-hillary-clinton-in-bellwether-florida-county/2293913

http://axiomstrategies.com/abc/
Remington Research also has a lot of polls by county, but they're also landline only polls.

The state polls were right in 2008 and 2012, no reason to believe county polls would be more accurate.

[Alcon]

All polls of counties, congressional districts, and the like have tiny sample sizes and are meaningless.

That's not necessarily true.  We occasionally get polls at the county level (mostly for novelty), and it's not uncommon to get a full-fledged, standard poll of a congressional district.  Although, like HillOfANight says, there's really no reason to take the county polls over state polls when predicting the state's presidential results.

[mb123]

Well as far as sample size all the polls I see on RCP/538 have sample sizes of like 900-1600 likely/registered voters so I dont see how you couldnt pull together the same sample size for a county and have it be more accurate (for that county's results) than for the a proportionately smaller sample of a larger body.

   I figured like a local news organization or school would want to do some for major counties (cuyahoga and like 2 other counties next to it have 10% of the votes in the state and I was trying to see if the partisan strength was the same there or not) in the state. There are a lot of things that could influence the outcomes in a state and rather than just take RCP avg + 538 avg and click a state into a column I was trying to look at last presidential election result [candidate and strength], current senate polls, whether bernie or hillary won the primary etc. In ohio I saw that a handful of counties helped put Obama at +3% but if Hillary doesnt perform as well as Obama with minority voters or just doesnt motivate democrats like Obama did in those counties the state could go differently. I guess I just wanted to wonk too much   Thanks for the link to the tampa bay polls.