We didn't do much to publicize this blog in the first few weeks, so I almost feel like it's been been the equivalent of a restaurant having a "soft opening"--we're open for business, but it's been mostly our friends who are eating the pizza. Nonetheless, a few other folks are starting to take notice and I'm pleased that we're beginning to get a little bit of coverage from this blog.
Our University PR Office put out this press release on the blog (thanks, Bill). The Smoky Mountain News had a piece on early voting featuring some of our analysis. The Tuck reader did a quick update on our election roundup here and also had a nice post on whether we might be saying "Speaker Shuler" anytime soon with a few quotes from me. Gibbs has been quoted in the Citizen-Times more times than I can possibly link to lately and he'll be doing a radio appearance on WRGC on election night (I might be there, too). I know there are a bunch of links I'm missing, but I need to find a Halloween costume so I'd better sign off for now.
Oh, and don't worry--I won't make a habit of doing posts like this. I'm sure very few of you outside of my mother want to read a post promoting the blog you're already reading.
Saturday, October 30, 2010
The Republicans Will Take Over the Majority of Governorships (probably)
How's that title for a hedge? Just as I did with the House and Senate Forecasts, I collected some of the major forecasts about Gubernatorial elections. Once again, these forecasts are one based on what Political Scientists call the "fundamentals"--factors like the state of the economy, the approval of the President, etc. There's not much candidates can do to influence these forecasts, which is why so many people are frustrated by them. The truth, however, is that despite taking very little about the individual candidates into account, they're pretty accurate.
Each line above is a forecast and the one marked average is..well...the average of the forecasts presented here. The big red line represents the number the Republicans need to reach to take the majority of the Governor's seats.
These forecasts clearly suggest that the Republicans are likely to gain the majority of Governorships, although not by a lot. So, given the nature of these forecasts, why did I hedge so much in my title? The answer: independents. The guy who I consider to be the Michael Jordan of Gubernatorial Scholars, Thad Beyle* recently shared some data with me showing that there are far more independents running for Governor in this election than anytime in recent memory. These independents may very well wreak havoc on election outcomes. If a few get elected, it also moves that big red line a bit.
So--I expect the Republicans will win some governorships, but I would take these forecasts with a little more salt than I would the House or Senate forecasts.
*By that I mean that he's really, really good. Not that he has a gambling problem or that he invests in questionable basketball teams.
These forecasts clearly suggest that the Republicans are likely to gain the majority of Governorships, although not by a lot. So, given the nature of these forecasts, why did I hedge so much in my title? The answer: independents. The guy who I consider to be the Michael Jordan of Gubernatorial Scholars, Thad Beyle* recently shared some data with me showing that there are far more independents running for Governor in this election than anytime in recent memory. These independents may very well wreak havoc on election outcomes. If a few get elected, it also moves that big red line a bit.
So--I expect the Republicans will win some governorships, but I would take these forecasts with a little more salt than I would the House or Senate forecasts.
*By that I mean that he's really, really good. Not that he has a gambling problem or that he invests in questionable basketball teams.
Wednesday, October 27, 2010
Youth Voter Turnout--Looking Back, Looking Forward
A friend of mine recently asked me the following: “Historically, younger voters (18-25?) have not bothered to vote, but they did participate in the Obama election. Is that demographic planning on voting in this election? Are they a block to reckon with, i.e. are politicians paying attention to them now?”
I figured some other people might be interested in the answer so here goes:*
According to the CIRCLE Foundation (the key resource on youth political engagement, turnout among young people was up 2% over 2004 and 9% over 4 years before (although they varied a ton by state). North Carolina is a particularly interesting case. In 2006, we had the 41st highest youth voter turnout in the country, but we had a huge surge in this last election and had the 11th highest turnout (55% youth turnout) following the 2008 Presidential election. I can’t think of any other examples where a state has jumped that much in such a short period of time.
On the one hand, the biggest predictor of whether you'll vote this time is whether you voted last time so simply getting young people to the polls should be a lot easier this time. On the other hand, a lot of people think that this surge was an Obama surge and we'll go back to normal patterns in the midterm.
So what signs do we have about this election? Once again, we have two pieces of evidence to rely on to answer this question: polling and early voting. The best polling data on this issue suggest that there’s a huge enthusiasm gap between young Democrats and Young Republicans. Young Democrats are much less engaged in this election than they were in 2006, while Young Republicans are (not surprisingly) more engaged. Given that more young people are Democrats than Republicans, we might expect to see slightly lower youth turnout this time. Here’s the money graph from the Pew report:
As for early voting, there is no evidence that the early voting population is any younger than it has been in previous years. In the 11th Congressional district, the average age of an early voter is 62. Don’t tell my dad but this doesn’t sound too young to me.
This last part of the question—whether young people are a block to reckon with—I’ll answer with a bit more informed opinion and fewer facts. The bottom line is that I haven’t seen any evidence suggesting that politicians are more likely to target or listen to young people than they were prior to 2008. A 2% increase in one election is great, but it's not exactly huge. And we’re talking about 2% higher among a fairly small proportion of the electorate so the overall influence is not all that big. You add to that that young voters are an inefficient group to reach (how many young people are actually registered in this district? How many have out of area cell phones? How many answer the phone? How many have a good address on file with the board of elections?) and I don't think politicians will spend much more time and effort than they already do trying to reach young voters.
This last part of the question—whether young people are a block to reckon with—I’ll answer with a bit more informed opinion and fewer facts. The bottom line is that I haven’t seen any evidence suggesting that politicians are more likely to target or listen to young people than they were prior to 2008. A 2% increase in one election is great, but it's not exactly huge. And we’re talking about 2% higher among a fairly small proportion of the electorate so the overall influence is not all that big. You add to that that young voters are an inefficient group to reach (how many young people are actually registered in this district? How many have out of area cell phones? How many answer the phone? How many have a good address on file with the board of elections?) and I don't think politicians will spend much more time and effort than they already do trying to reach young voters.
With that long of an answer, I doubt anyone will ask me any more questions.
*All of these data actually refer to people 18-29, which is the way most political scientists define “younger voters.”
The Democrats Will Hold Onto the Senate
Or at least that's what most Political Scientists think. Just as I did with the House Races here , below I plot the predictions of a few different political science forecasts, with a couple of media forecasts thrown in for good measure. The big red line is the line the Republicans are shooting for--the number of seats that marks when the Republicans take over the Senate.
As you can see, no current Political Science forecasts (at least that I can find) predict a Republican takeover. The average of all of these forecasts suggests that the Republicans will gain a little over 6 seats--enough to make the Republicans happy, but not enough to take over the chamber. The Abramowitz and Holbrook forecasts rely on fundamental conditions that are outside of the candidates' control--things like the state of the economy, the President's approval rating, etc. The 538 forecast uses a combination of individual race factors, and bigger, "macro-level" factors. The Sabato forecast relies on his estimation of how individual races will go.
As you can see, no current Political Science forecasts (at least that I can find) predict a Republican takeover. The average of all of these forecasts suggests that the Republicans will gain a little over 6 seats--enough to make the Republicans happy, but not enough to take over the chamber. The Abramowitz and Holbrook forecasts rely on fundamental conditions that are outside of the candidates' control--things like the state of the economy, the President's approval rating, etc. The 538 forecast uses a combination of individual race factors, and bigger, "macro-level" factors. The Sabato forecast relies on his estimation of how individual races will go.
Monday, October 25, 2010
More on Early Voting in WNC
There is a bit of a debate between Michael McDonald, George Mason University Political Scientist (not the Doobie Brother) and Nate Silver, founder of 538.com about how to read early voting data. Essentially, McDonald argues that you've got to compare this year's early voting data to early voting data in previous years to understand what it means, while Silver finds some usefulness in comparing to registration data (as I've been doing). Some are also debating whether early voting in 2008 is a better comparison or whether it's better to compare to 2006 (the last midterm year).
My goal is not to weigh into this debate, but to understand what's going on in WNC so I figured I would give fair time to the other side of this data-debate and compare the 2010 early voting data to 2006 early voting data on the same day of that election cycle to see if we can learn anything new. Each bar below represents the proportion that the group in question makes up of the total early voting electorate. So--to interpret one of the bars--the Democrats represent about .4 of the total early voting turnout in 2010.
The figure above reinforces some of the lessons we've learned by comparing turnout to voter registration, but it does tell us something slightly different: namely, independents seem to be turning out in greater numbers time time than they did in 2006. We could not see this difference using the way I've been looking at it before. Democratic turnout in the 11th district does seem to be down a bit and the Republican turnout is holding pretty steady compared to 2006. Maybe in our district it's not an "enthusiasm gap" with Republicans being more excited and mobilized than Democrats, but rather an "independent enthusiasm gap"
My goal is not to weigh into this debate, but to understand what's going on in WNC so I figured I would give fair time to the other side of this data-debate and compare the 2010 early voting data to 2006 early voting data on the same day of that election cycle to see if we can learn anything new. Each bar below represents the proportion that the group in question makes up of the total early voting electorate. So--to interpret one of the bars--the Democrats represent about .4 of the total early voting turnout in 2010.
The figure above reinforces some of the lessons we've learned by comparing turnout to voter registration, but it does tell us something slightly different: namely, independents seem to be turning out in greater numbers time time than they did in 2006. We could not see this difference using the way I've been looking at it before. Democratic turnout in the 11th district does seem to be down a bit and the Republican turnout is holding pretty steady compared to 2006. Maybe in our district it's not an "enthusiasm gap" with Republicans being more excited and mobilized than Democrats, but rather an "independent enthusiasm gap"
Sunday, October 24, 2010
Early Voting in WNC Update III: Through Oct. 23rd
Before moving into the update of the partisanship information, a reporter this week asked me what the demographic profile of early voters looks like in the 11th Congressional District thus far. It appears that the early voters in the 11th Congressional District are overwhelmingly white (96%), older (average age=62) and fairly split in terms of gender (51% female). The age and race trends are similar to what we have seen in the state as a whole, but interestingly, women seem to be turning out in much larger numbers in WNC than they are in the rest of the state (or at least than they were in the first few days of early voting.
As for partisanship, it looks like Republicans are still turnout out a larger proportion of their voters than are Democrats. Interestingly, however, Democrats still have the raw numbers advantage as they have so many more registered voters in the district to begin with.
This picture is interesting, but the real fun stuff happens when comparing what's going on in different counties. The graph below shows the proportion of registered voters in a county who are Republicans on the X (horizontal) axis, and the proportion of early voters in the county who are Republicans on the Y (vertical) axis. The line is what statisticians call a "line of best fit"--essentially when a county falls above the line, the Republicans are doing better than you'd expect from the proportion of Republicans in the county and the counties that fall below the line are experiencing less Republican turnout than you'd expect. For example, Republicans are doing a good bit better in Polk, Henderson, and Cherokee Counties than you would expect from the percent of voters in the County who are registered Republicans. Similarly, they are underrepresented in Graham County and slightly underrepresented in Madison, Jackson and Haywood. Looks like Democrats are gaining in Clay, Transylvania and Macon as the difference this week is different than it was just a few days before. Those who are more statistically inclined may be interested to know that the R-squared in the graph below is .74— lower than the .82 reported just a few days before and even lower than the .88 from a few days before that. Essentially this means the % of early voters in a county who are Republicans is increasingly deviating from the % of registered voters in a county who are Republican.
So--have the turnout numbers risen dramatically in the last few days? Did Bill Clinton's appearance in Asheville mobilize the Democratic base? The answers: sort of and no. The graph below shows the cumulative turnout by day and I don't see any huge spike (Democratic or Republican) after the former President's visit. Hopefully he at least got to eat at 12 bones while he was here to make his trip worth it.
Here's the full report if you're interested in the specific numbers by county:
http://tinyurl.com/wncearlyvotingoct24
As for partisanship, it looks like Republicans are still turnout out a larger proportion of their voters than are Democrats. Interestingly, however, Democrats still have the raw numbers advantage as they have so many more registered voters in the district to begin with.
This picture is interesting, but the real fun stuff happens when comparing what's going on in different counties. The graph below shows the proportion of registered voters in a county who are Republicans on the X (horizontal) axis, and the proportion of early voters in the county who are Republicans on the Y (vertical) axis. The line is what statisticians call a "line of best fit"--essentially when a county falls above the line, the Republicans are doing better than you'd expect from the proportion of Republicans in the county and the counties that fall below the line are experiencing less Republican turnout than you'd expect. For example, Republicans are doing a good bit better in Polk, Henderson, and Cherokee Counties than you would expect from the percent of voters in the County who are registered Republicans. Similarly, they are underrepresented in Graham County and slightly underrepresented in Madison, Jackson and Haywood. Looks like Democrats are gaining in Clay, Transylvania and Macon as the difference this week is different than it was just a few days before. Those who are more statistically inclined may be interested to know that the R-squared in the graph below is .74— lower than the .82 reported just a few days before and even lower than the .88 from a few days before that. Essentially this means the % of early voters in a county who are Republicans is increasingly deviating from the % of registered voters in a county who are Republican.
So--have the turnout numbers risen dramatically in the last few days? Did Bill Clinton's appearance in Asheville mobilize the Democratic base? The answers: sort of and no. The graph below shows the cumulative turnout by day and I don't see any huge spike (Democratic or Republican) after the former President's visit. Hopefully he at least got to eat at 12 bones while he was here to make his trip worth it.
Here's the full report if you're interested in the specific numbers by county:
http://tinyurl.com/wncearlyvotingoct24
Saturday, October 23, 2010
Every Little Bit Counts
We've been talking about this blog for a while, but it took me being laid up with back surgery to get it online. A couple of folks in the media including WRGC radio and the Tuck Reader have already taken notice so thanks to them. I've also gotten some friendly calls from other media outlets in the region this week asking about the content--and mostly about the early voting data so thanks to them as well. Please keep reading and we'll try to keep the content going.
In the meantime, here's a musical interlude befitting the title of this post.
In the meantime, here's a musical interlude befitting the title of this post.
The Republicans Will Take Over the House
Or at least that's what most political scientists believe. So how do they know so early? Whereas journalists and pollsters rely on last-minute polls and campaign events to read tea-leaves to predict the outcome of elections, Political Scientists believe they can predict the outcome of elections well in advance of the election and often without even knowing who the candidates are. The bad news is that makes following the horse-race a lot less interesting. The good news is that the track record of these forecasts is pretty good and less time following the horse race means more time that can be spent following the College football season.
There are lots of these forecasts out there, but one popular one, by Emory University Political Scientist Alan Abramowitz says you can predict the overall seat share in Midterm elections by knowing two things: (1) the result of the "generic ballot"* in early September and (2) the President's approval in early September. Other forecasts focus more heavily on the state of the economy, but most of the final predictions end up looking fairly similar.
Below I've plotted the predictions of various political science models, along with the forecasts of 538.com, a popular web site by uber-nerd Nate Silver that uses political science methods, but is not written by a political scientist. I also put in a line that is the average of all of these forecasts. As most of you know, the Democrats control the House and the Republicans need 39 seats to take it over. All of the predictions but two show the Republicans moving past the margin number of 39 (yes, that's what the big, red line represents).
It's important to remember that all of these forecasts include lots of caveats and many of them even quantify the probability of error (which I completely ignore in this post). In other words, don't come after these folks if their predictions don't come true. And if you do, please don't tell them that you read it here.
If you want to read more about these predictions, I encourage you to consult the most recent issue of PS: Political Science and Politics, a journal that dedicated a special issue to forecasting the 2010 midterms They've made the whole thing available to the public, so browse away.
On a final note: I intend this blog to mostly address local and state issues since lots of impressive folks already post on national politics, and with with more authority and status than I do. With that said, I can't resist weighing in a little on the national scene from time-to-time. And with that cautionary note: I'll post tomorrow with similar predictions on the Gubernatorial and Senatorial elections.
* Here's the text of a traditional "generic ballot question" "We'd like to get your opinion about how you might vote in the House of Representatives election on November 2nd, 2010. If the House of Representatives election were held today, for whom would you most likely vote? [ASKED OF THOSE WHO HAVE ALREADY VOTED: For which candidate for the House of Representatives did you vote?]?" The Democratic Candidate? The Republican Candidate? Other
There are lots of these forecasts out there, but one popular one, by Emory University Political Scientist Alan Abramowitz says you can predict the overall seat share in Midterm elections by knowing two things: (1) the result of the "generic ballot"* in early September and (2) the President's approval in early September. Other forecasts focus more heavily on the state of the economy, but most of the final predictions end up looking fairly similar.
Below I've plotted the predictions of various political science models, along with the forecasts of 538.com, a popular web site by uber-nerd Nate Silver that uses political science methods, but is not written by a political scientist. I also put in a line that is the average of all of these forecasts. As most of you know, the Democrats control the House and the Republicans need 39 seats to take it over. All of the predictions but two show the Republicans moving past the margin number of 39 (yes, that's what the big, red line represents).
It's important to remember that all of these forecasts include lots of caveats and many of them even quantify the probability of error (which I completely ignore in this post). In other words, don't come after these folks if their predictions don't come true. And if you do, please don't tell them that you read it here.
If you want to read more about these predictions, I encourage you to consult the most recent issue of PS: Political Science and Politics, a journal that dedicated a special issue to forecasting the 2010 midterms They've made the whole thing available to the public, so browse away.
On a final note: I intend this blog to mostly address local and state issues since lots of impressive folks already post on national politics, and with with more authority and status than I do. With that said, I can't resist weighing in a little on the national scene from time-to-time. And with that cautionary note: I'll post tomorrow with similar predictions on the Gubernatorial and Senatorial elections.
* Here's the text of a traditional "generic ballot question" "We'd like to get your opinion about how you might vote in the House of Representatives election on November 2nd, 2010. If the House of Representatives election were held today, for whom would you most likely vote? [ASKED OF THOSE WHO HAVE ALREADY VOTED: For which candidate for the House of Representatives did you vote?]?" The Democratic Candidate? The Republican Candidate? Other
Thursday, October 21, 2010
Predicting NC Election Results
Sometimes predicting the future is pretty easy (Didn't everyone predict that Joan would keep the baby on the season final of Mad Men?); other times it's more difficult (did anyone think Alvin Green would be the Democratic Party's candidate for the South Carolina Senate seat?). So how hard is it to predict the results in the major NC political battles? Looks like the races in North Carolina contain some Joans and some Greenes. Here's the linkup starting from the most certain to the least certain (excluding the House races that are virtual locks):
Burr-Marshall: Richard Burr, our senior Senator was first elected in 2004 by beating my boss', boss' boss' boss' boss, 52-47. After some tight early polling numbers, Burr looks to be almost a lock in November. My favorite political site, 538.com gives Burr a 97% chance of being re-elected. The Cook political report and Real Clear Politics list the race as "likely Republican" and the most recent average polling data predict a 51-35 result. Finally, Political Scientist Tom Holbrook predicts that Marshall will garner a scant 42% of the vote. Barring major disaster, it looks like Burr will coast to victory in November.
Shuler-Miller (NC-11): Heath Shuler, first elected in 2006 can't rest as easily as Richard Burr, but most indications suggest that Shuler is the favorite in this seat. 538.com gives Shuler a 75% chance of holding onto his seat. The Cook Political Report rates the race as one that "Leans Democratic," although Real Clear Politics marks it a "Toss-up." As we draw closer to election day, undecided voters tend to fall towards the incumbent, and the most recent polling data look good for the former quarterback. The race is certainly tighter than the one at the top of the ticket, but Shuler remains the favorite.
McIntyre-Pantano (NC-7): Mike McIntyre has represented the 7th District (Wilmington) since 1996, but he's facing a very tough fight at the hands of Ilarino Pantano. Most forecasters are finding this one tough to predict. 538 gives the incumbent McIntyre a 55% chance of victory while the Cook Political Report says it "Leans Democrat" and Real Clear Politics calls it a toss-up.
Kissell-Johnson (NC-8): Perhaps even tighter is the race for the 8th District that stretches from parts of Charlotte to Fayetteville. Here, the incumbent Larry Kissell is facing tough fight to keep his seat. Cook calls this one a "toss-up" Real Clear Politics says it "Leans Republican" and 538 gives the Republican challenger a 54% chance of victory.
So what about the state legislature? Because there is very little reliable polling data at the state legislative level, there are almost no credible seat-by-seat forecasts. One person, however, has begun to forecast overall partisan control at the state legislative level, using a complicated formula that includes 30 variables ranging from campaign spending to previous voting patterns in the district to the popularity of the Governor. Unfortunately, his forecast in North Carolina is the statistical equivalent of throwing up his hands--he gives the Democrats a 47% of maintaining control of the State House and a 50% chance of maintaining control of the state Senate.
I'll follow-up soon by collecting forecasts at the national level.
Burr-Marshall: Richard Burr, our senior Senator was first elected in 2004 by beating my boss', boss' boss' boss' boss, 52-47. After some tight early polling numbers, Burr looks to be almost a lock in November. My favorite political site, 538.com gives Burr a 97% chance of being re-elected. The Cook political report and Real Clear Politics list the race as "likely Republican" and the most recent average polling data predict a 51-35 result. Finally, Political Scientist Tom Holbrook predicts that Marshall will garner a scant 42% of the vote. Barring major disaster, it looks like Burr will coast to victory in November.
Shuler-Miller (NC-11): Heath Shuler, first elected in 2006 can't rest as easily as Richard Burr, but most indications suggest that Shuler is the favorite in this seat. 538.com gives Shuler a 75% chance of holding onto his seat. The Cook Political Report rates the race as one that "Leans Democratic," although Real Clear Politics marks it a "Toss-up." As we draw closer to election day, undecided voters tend to fall towards the incumbent, and the most recent polling data look good for the former quarterback. The race is certainly tighter than the one at the top of the ticket, but Shuler remains the favorite.
McIntyre-Pantano (NC-7): Mike McIntyre has represented the 7th District (Wilmington) since 1996, but he's facing a very tough fight at the hands of Ilarino Pantano. Most forecasters are finding this one tough to predict. 538 gives the incumbent McIntyre a 55% chance of victory while the Cook Political Report says it "Leans Democrat" and Real Clear Politics calls it a toss-up.
Kissell-Johnson (NC-8): Perhaps even tighter is the race for the 8th District that stretches from parts of Charlotte to Fayetteville. Here, the incumbent Larry Kissell is facing tough fight to keep his seat. Cook calls this one a "toss-up" Real Clear Politics says it "Leans Republican" and 538 gives the Republican challenger a 54% chance of victory.
So what about the state legislature? Because there is very little reliable polling data at the state legislative level, there are almost no credible seat-by-seat forecasts. One person, however, has begun to forecast overall partisan control at the state legislative level, using a complicated formula that includes 30 variables ranging from campaign spending to previous voting patterns in the district to the popularity of the Governor. Unfortunately, his forecast in North Carolina is the statistical equivalent of throwing up his hands--he gives the Democrats a 47% of maintaining control of the State House and a 50% chance of maintaining control of the state Senate.
I'll follow-up soon by collecting forecasts at the national level.
Wednesday, October 20, 2010
Cell Phones and Polling
is a lot of misinformation about cell phones and polling out there and the election season is only going to bring on more polls, so I thought it would be a good time to review some recent work on cell phones and polling.
Many people believe that pollsters cannot include cell phones in their polls. That’s simply not true. Many pollsters (and almost all of the good ones) include cell phones in their sample. Pollsters usually draw samples from what is called “random digit dialing.” Let’s take Sylva for example—many Sylva phone numbers begin with 586. A pollster then can call numbers at random beginning with 586. Cell phone numbers around here often begin with 399, 506, 507, or 508. There’s no reason a good poll can’t do the same thing with cell numbers that they do with land-line numbers—begin with a known prefix and then generate random numbers from there.
The problem with cell phones comes from people like my friend Eric. Eric lives in Sylva, but moved here from Tennessee. When he moved, he kept his Tennessee area code. A pollster who is trying to do a poll of Jackson County residents would not be able to include Eric. Similarly a person who leaves Jackson County may keep her phone number with a 399 prefix, and she might get a call for a survey meant for Jackson County residents.
So—how often does this happen? The Pew Center for the People and the Press recently released a report on this and found that 94% of cell phone users have a number that falls within the same Census region in which they live. 90% have cell phone numbers that match their current state of residence and only 60% have cell numbers that fall within their current county of residence. In other words, the smaller the geographic area you’re polling in, the bigger the problem you’ve got.
So what does this mean for people who want to make sense of polling data? First, make sure that the poll you’re reading about includes cell phones and not just land-lines. Polls that include only land-lines probably under-represent young people and Democrats. Second, the smaller the geographic unit, the more suspect you should be of the poll. As someone who’s conducted a few polls of local areas, I don’t mean to suggest that polls about small geographic areas are bad, but that a little more caution is probably warranted.
I’ll probably follow up with some more on polling soon.
Many people believe that pollsters cannot include cell phones in their polls. That’s simply not true. Many pollsters (and almost all of the good ones) include cell phones in their sample. Pollsters usually draw samples from what is called “random digit dialing.” Let’s take Sylva for example—many Sylva phone numbers begin with 586. A pollster then can call numbers at random beginning with 586. Cell phone numbers around here often begin with 399, 506, 507, or 508. There’s no reason a good poll can’t do the same thing with cell numbers that they do with land-line numbers—begin with a known prefix and then generate random numbers from there.
The problem with cell phones comes from people like my friend Eric. Eric lives in Sylva, but moved here from Tennessee. When he moved, he kept his Tennessee area code. A pollster who is trying to do a poll of Jackson County residents would not be able to include Eric. Similarly a person who leaves Jackson County may keep her phone number with a 399 prefix, and she might get a call for a survey meant for Jackson County residents.
So—how often does this happen? The Pew Center for the People and the Press recently released a report on this and found that 94% of cell phone users have a number that falls within the same Census region in which they live. 90% have cell phone numbers that match their current state of residence and only 60% have cell numbers that fall within their current county of residence. In other words, the smaller the geographic area you’re polling in, the bigger the problem you’ve got.
So what does this mean for people who want to make sense of polling data? First, make sure that the poll you’re reading about includes cell phones and not just land-lines. Polls that include only land-lines probably under-represent young people and Democrats. Second, the smaller the geographic unit, the more suspect you should be of the poll. As someone who’s conducted a few polls of local areas, I don’t mean to suggest that polls about small geographic areas are bad, but that a little more caution is probably warranted.
I’ll probably follow up with some more on polling soon.
Early Voting Update--data through Oct. 19th
I've updated the WNC early voting data through Oct. 19th. As Led Zeppelin once said--"the song remains the same." Republicans are doing slightly better than might be expected based on their share of registered voters, but the Democrats have slightly more votes in all. Check out this link for the full report (including county-level data) or read below for the highlights.
See below for the graph for the 11th Congessional District updated with the new votes
I also created a graph to simulate the one from a few days ago--so you can compare the % of early voters who are Republicans to the % of registered voters who are Republicans by County. Once again, in general, counties with more registered Republicans in the county are seeing higher turnout among Republicans thus far, but there are a few interesting deviations. Republican turnout is Henderson County continues to be better than we might have expected. The story isn't all good for Republicans, however, as they are not doing as well in Madison, Jackson, and Haywood Counties--and particularly in Graham. Those who are more statistically inclined may be interested to know that the R-squared in the graph below is .82—slightly lower than the .88 reported just a few days before.
Next I wanted to begin to graph turnout by day to see which days are the biggest days and if some days look better for one party than others. The graph below tracks the cumulative vote totals for each party by day for the entire 11th Congressional District. Clearly this week has seen a bit pick-up in early voting.
Please continue to check back as the turnout game gets more interesting over the next few weeks. I'm particularly interested to see if Clinton's appearance in Asheville gives the Democrats a temporary boost in Buncombe County on Friday and Saturday.
See below for the graph for the 11th Congessional District updated with the new votes
I also created a graph to simulate the one from a few days ago--so you can compare the % of early voters who are Republicans to the % of registered voters who are Republicans by County. Once again, in general, counties with more registered Republicans in the county are seeing higher turnout among Republicans thus far, but there are a few interesting deviations. Republican turnout is Henderson County continues to be better than we might have expected. The story isn't all good for Republicans, however, as they are not doing as well in Madison, Jackson, and Haywood Counties--and particularly in Graham. Those who are more statistically inclined may be interested to know that the R-squared in the graph below is .82—slightly lower than the .88 reported just a few days before.
Next I wanted to begin to graph turnout by day to see which days are the biggest days and if some days look better for one party than others. The graph below tracks the cumulative vote totals for each party by day for the entire 11th Congressional District. Clearly this week has seen a bit pick-up in early voting.
Please continue to check back as the turnout game gets more interesting over the next few weeks. I'm particularly interested to see if Clinton's appearance in Asheville gives the Democrats a temporary boost in Buncombe County on Friday and Saturday.
Monday, October 18, 2010
Early Voting in WNC as of October 18th
Many of us are curious about what's going to happen in the November elections. Unfortunately the only real data we have are polling data and indications about who's voting early (but hopefully not often). Nationally, there are a lot of good sites that calculate early voting numbers by state. Locally, however, there are no such resources. The NC Board of Elections has the individual level data available for download (down to the individual voter), but it takes some know-how and a considerable amount of time to download, analyze and make sense of these data.
To satisfy my own curiosity and to contribute to our knowledge of politics in WNC, every few days between now and the election, I plan to download and analyze early voting data for each county in the 11th Congressional District on this site. Although I also have data by age and race (and I’ll be happy to make those available), this analysis only shows the percent of early voters by Party Identification. It’s important to remember that this does not tell us who someone is voting for—A Republican may choose to for a Democrat and vice-versa—but it does give us some indication of who’s winning the mobilization game in Western North Carolina. The data below include data from October 17th.
In general, the data don’t show that either party has a large advantage this early, although it does look slightly better for Republicans than Democrats.
The figure above shows that Republicans are enjoying a slight advantage in early voting. Republicans make up a slightly larger percent of the early voters than they do of the registered voters in Western North Carolina. This is not terribly unusual for early voting. It is also not unusual to see that unaffiliated voters are less well-represented in early voting than they are in the electorate.
The following figure shows that in general, counties with more registered Republicans in the county are seeing higher turnout among Republicans thus far. The interesting cases are the ones way off the line. For example, Republicans are doing a good bit better in Henderson County than you would expect from the percent of voters in the County who are registered Republicans. Similarly, they are slightly underrepresented in Graham County. For those who are more statistically inclined, the R-squared for this figure is about .88.
To satisfy my own curiosity and to contribute to our knowledge of politics in WNC, every few days between now and the election, I plan to download and analyze early voting data for each county in the 11th Congressional District on this site. Although I also have data by age and race (and I’ll be happy to make those available), this analysis only shows the percent of early voters by Party Identification. It’s important to remember that this does not tell us who someone is voting for—A Republican may choose to for a Democrat and vice-versa—but it does give us some indication of who’s winning the mobilization game in Western North Carolina. The data below include data from October 17th.
In general, the data don’t show that either party has a large advantage this early, although it does look slightly better for Republicans than Democrats.
The figure above shows that Republicans are enjoying a slight advantage in early voting. Republicans make up a slightly larger percent of the early voters than they do of the registered voters in Western North Carolina. This is not terribly unusual for early voting. It is also not unusual to see that unaffiliated voters are less well-represented in early voting than they are in the electorate.
The following figure shows that in general, counties with more registered Republicans in the county are seeing higher turnout among Republicans thus far. The interesting cases are the ones way off the line. For example, Republicans are doing a good bit better in Henderson County than you would expect from the percent of voters in the County who are registered Republicans. Similarly, they are slightly underrepresented in Graham County. For those who are more statistically inclined, the R-squared for this figure is about .88.
If you'd like to see the exact numbers by county, please check out the brief report here http://tinyurl.com/earlyvotinginwnc
Welcome
Welcome to the Blog for the Public Policy Institute at Western Carolina University. We intend to use this as a platform to add to the policy debate in the Tarheel State. Most of our posts will try to connect findings in Political Science and Public Affairs to the real world of politics and policy in North Carolina. Many of our posts will include original data analysis. Enjoy and please let us know if you have any suggestions.
Subscribe to:
Posts (Atom)