polling data for states & parishes
October 25, 2008 8:16 AM Subscribe
Good sites for political polling data by state county or parish?
Google-fu issues here. I would like to find a good site(s) for state data by county or parish, especially for Louisiana and Texas. Sites with easy to read maps a plus. Previous election years would be nice too. All I seem to find are the national ones and I'd like a little more detail within the individual states. Thanks in advance for any assistance.
Google-fu issues here. I would like to find a good site(s) for state data by county or parish, especially for Louisiana and Texas. Sites with easy to read maps a plus. Previous election years would be nice too. All I seem to find are the national ones and I'd like a little more detail within the individual states. Thanks in advance for any assistance.
Oh wait, do you want opinion polls, like those run by various independent organizations, and not polling as in actual elections? I'm not as sure about those . . .
posted by that girl at 10:56 AM on October 25, 2008
posted by that girl at 10:56 AM on October 25, 2008
Response by poster: that girl those look great for the elections past. Any opinion polls predicting the current election are good too.
posted by dog food sugar at 12:55 PM on October 25, 2008
posted by dog food sugar at 12:55 PM on October 25, 2008
What you're looking for doesn't exist for opinion polls.
There's a good reason national and state polls are not broken down by county. If you broke a state poll down by county, you'd have a few people in every county. That is, you'd be trying to gauge the opinions of County X with a sample of only 5--20 people, and the margin of error would be so large that it would only tell you that between 20 and 90 percent of the people in County X plan to vote for Obama. In a national poll, most counties would have no respondents at all.
About all you'd be able to find is polls of metro areas that happen to ask about presidential preference as well. But you'll only find those for a small percent of counties.
posted by ROU_Xenophobe at 1:53 PM on October 25, 2008
There's a good reason national and state polls are not broken down by county. If you broke a state poll down by county, you'd have a few people in every county. That is, you'd be trying to gauge the opinions of County X with a sample of only 5--20 people, and the margin of error would be so large that it would only tell you that between 20 and 90 percent of the people in County X plan to vote for Obama. In a national poll, most counties would have no respondents at all.
About all you'd be able to find is polls of metro areas that happen to ask about presidential preference as well. But you'll only find those for a small percent of counties.
posted by ROU_Xenophobe at 1:53 PM on October 25, 2008
Response by poster: Ohhh. Well that makes sense. Thanks ROU_Xenophobe. What sparked this was this article in my local paper about a poll in my county, the most populated county in Texas, leaning towards Obama. I was wondering about the rest of the state.
The husband is from Louisiana and we talk about moving there every once in a while, but I don't want to move to a more conservative area.
Thanks both of you for your time and your answers.
posted by dog food sugar at 2:46 PM on October 25, 2008
The husband is from Louisiana and we talk about moving there every once in a while, but I don't want to move to a more conservative area.
Thanks both of you for your time and your answers.
posted by dog food sugar at 2:46 PM on October 25, 2008
I was wondering about the rest of the state.
One of the weird things about polling is that the number of people you need to sample to get a reasonable margin of error doesn't change very much at all with the population of the thing you're sampling.
So if you want a 3% margin of error for Harris County, you need about 1000 people. Which isn't that big a deal -- 1000 adults out of a few million.
But if you want a 3% margin of error from a county with 100,000 adults, you need... about 1000 people. And if you want a 3% margin of error from a county with 10,000 adults, you still need 964 people, or 10% of the population. If you wanted to get a 3% margin of error out of one of those horrifying west Texas counties with 1000 adults and a billion tumbleweed, you'd need to sample just over 500 people, or about half of the population.
It's weird, but it's how the math works itself out.
Anyhow, there are real reasons why you don't see county breakdowns of major polls.
posted by ROU_Xenophobe at 6:13 PM on October 25, 2008
One of the weird things about polling is that the number of people you need to sample to get a reasonable margin of error doesn't change very much at all with the population of the thing you're sampling.
So if you want a 3% margin of error for Harris County, you need about 1000 people. Which isn't that big a deal -- 1000 adults out of a few million.
But if you want a 3% margin of error from a county with 100,000 adults, you need... about 1000 people. And if you want a 3% margin of error from a county with 10,000 adults, you still need 964 people, or 10% of the population. If you wanted to get a 3% margin of error out of one of those horrifying west Texas counties with 1000 adults and a billion tumbleweed, you'd need to sample just over 500 people, or about half of the population.
It's weird, but it's how the math works itself out.
Anyhow, there are real reasons why you don't see county breakdowns of major polls.
posted by ROU_Xenophobe at 6:13 PM on October 25, 2008
« Older How to create a database for personal statistics? | Tennis training aids? Do they work? Newer »
This thread is closed to new comments.
My method is to search for "$STATE elections" or "$STATE board of elections" and go from there. Most have results, likely by county or other political subdivision. If you want even more specific, you can look for a specific county's elections board. I think my city gets as specific as city wards.
Most of these places probably won't have done the work to get pretty maps, but I can't guarantee it one way or another.
posted by that girl at 10:54 AM on October 25, 2008