Pollster.com

Articles and Analysis

 

More Cell Phone Data from Gallup

Topics: Cell Phones

In my column this week, I summarized some recent data provided by both the Pew Research Center and Gallup on the interviews both have been conducting this year via cell phone. Both pollsters have seen a similar pattern. With the interviews among "cell phone only" respondents (those who live in households without landline telephones) included, Obama does a point or two better, McCain does a point or two worse.

One of the more significant findings in this data was from Pew. They found big differences between 18-to-29-year-old voters with landline phone service an those reachable only by cell phone. The cell phone only younger voters favored Obama by a much bigger margin (62% to 27%, n=250) than the landline younger voters (52% to 39%, n=146). While those differences were just large enough to be statistically significant, the sample sizes involved are relatively small, so it is hard to be certain about the magnitude of the difference.

081017 pew cell phone.png

Curious, I asked Gallup's Jeff Jones if they could replicate the Pew table, and he kindly sent along the following data. You do not see differences by party identification. The ballot numbers show the same general pattern, but the differences are not as large as they appear on the Pew survey. Obama, runs better among the Gallup cell-phone only 18-to-29-year-olds (+28) than among those reachable by landline (+20). Those differences still appear to be statistically significant given the larger sample sizes.

081017 gallup cell phone.png

The bottom line? Gallup and Pew have produced data supporting the theory of a likely "cell phone effect" in Obama's favor that weighting by age may not eliminate. However, the Gallup data implies a smaller effect than Pew. Of course, without seeing a weighted comparison for all voters by Gallup (something they may not have time to produce until after the election), we won't know for sure.

 

Comments
Darwin Please:

Interesting, but who knows if they will even vote.

____________________

s.b.:

I find it amazing that gallup is so concerned about this when it is putting out a poll that ssumes an 82% turnout of Rv's that you are chosing to use. The 2004 election at 60% RV turnout had the third highest turnout in modern US history behind 1960 at 63% and 1968 at 62%. Does Gallup really believe that 20% more Registered voters will turnout this year they did during Vietnam? 2004 turnout was historically high, but Gallup and pollster think that voting pattern isn't good enough to help determine likely voters???? I would be surprised if this election reaches 60% Rv turnout expecially since a substantial amount of voter registrations may have been made without the voters knowing about it, if they even exist. Gallup is overestimating voter turnout by at least 20%. When all tighter screen show Obama losing support to McCain, I think the lack of oppropriate screenign is a heck of a lot more of a concern than cell phone only users not being reached in a few polls. I think it accounts for a whole lot more than 2 points. In fact we know that tightening the screen by 7%, adds 4% to the spread in favour of Obama. So you do the math. How much advantage does Mccain have when you take away the 20% who won't go to the polls even in the traditional Gallup screen? Nov 4th could be a very interesting day indeed.

____________________

s.b.:

Sorry about the typos and losenign the screen by 7% adds a 4% spread for Obama in gallup, not the other way around.

My point is if voter turnout is 63% of RV it will be the highest in modern US history, and I dont think it will be.

Different screens show very different results. Pollster is choosing the Gallup screen that assumes an 82% likely voter turnout. Total garbage.

____________________

mattn:

Regarding turnout, it's a little more complicated than you (s.b.) make out, because what matters isn't overall turnout, but rather turnout in the battleground states. In that case, the picture is a bit more mixed, even in 2004 (http://elections.gmu.edu/Turnout_2004G.html).

But of course, what this really shows is that it's the individual state polls that matters here. Why anyone looks at the national polls (which, frankly, if they get it right it's just the sheerest amount of luck) is beyond me.

____________________

s.b.:

Yes Matt turnout in battleground states is higher. Minnesota I believe was the highest at 78% registered voter turnout in 2004. I am however talking about a national poll and a national likely voter screen.

Gallup's "New" model assumes an 82% likely voter turnout. It's garbage. As I have said with the third highest voter turnout in US history, 2004 is a fine election to use as a likely voter parameter.

Do the math, a 7% looser screen gives Obama 4%, so take away the 20% of RV's that won't vote even if this election has the highest voter turnout in US history, and what happens to the numbers?

I think pollsters are being completely irresponsible right now and drinking Obama koolaid instead of sticking by models that have worked for 50 years.

____________________

mattn:

I think you're mistaking percentage of eligible voters and percentage of registered voters. Typically, turnout of RV is in the 80's (2004 was 89%, according to the census bureau: http://www.census.gov/Press-Release/www/releases/archives/voting/004986.html). It is turnout of *all* adults that was around 60%.

____________________

Stonecreek:

s.b.,

I think you're probably smarter than your postings let on. Why obfuscate simply to make things look better for McCain?

The number of voters included in the LV screen, whether 82% of the RV's or 60% of the RV's, is irrlevant. What is relavent is the percentage of each candidate's RV's that turn out. Voter fall off from any screen will vary with any number of factors (weather, competing sports events, work or family obligations, illness, etc.), but the most important factor is voter enthusiasm.

When you compare the "strong favorables" for Obama with the same figure for McCain, you find about a 5/3 ratio; similarly, you find a higher ratio of "committed" voters who say they will definately vote for Obama than will definately vote for McCain. Some 70% of Obama voters say they are enthusiastic about their choice compared to something like 40% of McCain voters. All of this points to a larger drop off of McCain's RV's than of Obama's RV's.

Simply because the actual voter turnout may only be 60% of the registrations rather than 80% of the registrations doesn't make a voter screen which uses 80% of the RVs as LVs invalid. That's like saying that a poll which only samples 1% of the electorate is invalid. The best LV screen is NOT the one which most accurately predicts the actual number of voters, but rather the one which most accurately predicts the voting patterns of the actual voters. But, of course, you already knew this, didn't you?

____________________

As a 41-year old CPO I like to think I'm in this special "18-to-29-year-old" demographic. And yes, I'm for Obama.

See who's been inducted into the Hypocrites Hall of Infamy.

____________________

wagthedog1001:

Stonecreek:

Thanks for explaining percentages, ratios... et al to our friend.

You are right and the most important factor is voter enthusiasm. I've said all along, if the non-white and young voters come out like I think they will, these polls will mean nothing. If they don't come out, they may get what they deserve. Until we wake up and vote for strengthening the middle class, we will continue to have the Richie Ritches run our country.

____________________

BridgeportJoe:

I think you're mistaking percentage of eligible voters and percentage of registered voters. Typically, turnout of RV is in the 80's (2004 was 89%, according to the census bureau: http://www.census.gov/Press-Release/www/releases/archives/voting/004986.html). It is turnout of *all* adults that was around 60%.

Ouch. The following is a more direct route to that data. (And yes, the percent of voting registered (as opposed to eligible) voters is usually in the mid-80% range.)

http://www.census.gov/population/www/socdemo/voting/cps2004.html

____________________

RS:

Assuming the differences in cell-phone-user-voter-preference is real, I wonder if the difference is because of where cell-phones are used.
I am guessing that cell-phones work best in cities, and coverage is spotty in rural areas. So cell-phone-only folks are likely to be more urban... and thus probably more Democratic/liberal. So they prefer Obama by a bigger margin.

As for s.b. - I don't see where Gallup says their LV-II model uses 82% turnout. Here's the relevant piece of information when they first started using the LV models:

Among typical "likely voters" -- the subset of registered voters who appear most likely to vote on Election Day according to their current voting intentions and past voting behavior -- Obama's lead is a slightly narrower seven points, 51% to 44%. This assumes that about 60% of the voting age population will vote, slightly higher than the 55% who turned out in 2004.

Among a more broadly defined likely voter group that only takes into account current voting intentions -- not past voting behavior -- Obama's lead is the same 10 points as among all registered voters, 53% to 43%. This group represents approximately two-thirds of the general public, a significantly higher proportion than has turned out in any recent election.


[emphasis added]
Link:
http://www.gallup.com/poll/111112/Gallup-Daily-Obama-Ahead-51-41.aspx
____________________

s.b.:

No I'm not confusing Rv's with all adults. The numbers for voter percentages of all adults is in the low 50%'s. Voter turnout in 2004 was 60% of registered voters, the third highest in mdern US history.

____________________

BridgeportJoe:

No I'm not confusing Rv's with all adults. The numbers for voter percentages of all adults is in the low 50%'s. Voter turnout in 2004 was 60% of registered voters, the third highest in mdern US history.

You're just flatly wrong. 2+2=5 wrong.

In the 2004 election, there were roughly 215,694,000 adults 18+ (the Census numbers are in 1000s, so that's as exact as I can get) in the US. Of those, 142,070,000 (65.2%) were registered to vote. 125,736,000 people actually voted -- that's 58.3% of all adults, and 88.5% of registered voters.

Again, these are Census figures. I'll repost the same link.

http://www.census.gov/population/www/socdemo/voting/cps2004.html

You really just need to admit you are dead wrong and let it drop.

____________________

s.b.:

Stonecreek, I've seen enough elections to know that unfortunately voter enthusiasm for turnout pales in comparison to negative voting in turning out voters.

Nothing motivates people like negative voting. ie voting against something. As much as their is enthusiasm for Obama, I do not think it will generate more voter turnout than the negative voting turnout against George Bush at the height of the war in 2004. In 1968 there was 62% voter turnout, the second highest in modern history and that was against the war in Vietnam.

Do Pollster and Gallup and all the other polsters who have changed their likely voter models think that there is more enthusiasm among young people to vote than there was to vote against the Vietnam war or against Bush for that matter?

Again I will say, it's totally wreckless and irresponsible and completely unfounded for polling companies to be doing this. It compromises their professionalism and calls into question the legitimacy of all their polls and models.

Honestly I think this election may have a lower voter turnout than 2004, at the height of the war and the protest vote against Bush.

Like i said, i've yet to see an enthusiasm vote trump a protest vote for voter turnout, ever, anywhere in the world.

And Gallup's traditional model still assumes a 20% higher RV turnout than at the height of the Vietnam war, but pollster thinks voter turnout might be higher than that and has to use a model with a looser screen?

And the assumption that the turnout will favour Obama is also unfounded. Again Obama has higher stronger negatives than McCain. Negative enthusiasm motivates people to come out to vote against someone more than positives. These negative voters will also have much higher likelihoods of voting and a more consistent voting patern. Just a thought.

____________________

As others have already pointed out, s.b. is confused between registered and eligible voters. I should know, so I am the one that computes the eligible voter turnout numbers.

http://elections.gmu.edu/voter_turnout.htm

But if one wanted to compute turnout rates among registered voters, the Census' Current Population Survey is one way to go, but note how in 2004, 89% of North Dakotans reported being registered (the state is the only one without voter registration). So, one must take these self-reported numbers with a grain of salt. And if one wanted to use aggregate registration statistics as the denominator, the problem becomes that states vary on how well they purge their voter rolls, both across states and time.

It is for these reasons that I view registration turnout rates with much skepticism and prefer to calculate turnout rates for eligible voters. However, many pollsters use registration as a screen. Judging the value of that screen is actually much more difficult than one might imagine.

____________________

s.b.:

btime now. I am NOT talking about over 18 year old adults in the poulation. i am talking about the percentage of Registered voters who actually vote. I don't give a rats ass about people who aren't registered or can't vote and they aren't in the Gallup sample i am discussing.

The % of Registered voters who voted aka voter turnout for 2004 was 60%. if someone is not eliglble to vote they don't countin voter turnout, nor do they count in the math used, very simple math i might add that shows the % of Gallup Registered voters who didn't make the likely voter screens.

You take the LV number and you divide it by the RV number Gallup sampled. If their screen isn't getting them to 60%, it's too loose. Let's assume the answering of the poll itself is a screen and that you can't be perfect, and that their might be slightly higher turnout this year (which i doubt, especially given the voter registration fraud going on) Anyways, a screen that gets you to 70% might be reasonable, loose but reasonable. A screen that leaves 82% of Rv's in the poll is absurd and goes against 50 years of Gallup's own polling practice.

Again with the third highest voter turnout in modern history, 2004 is a fine parameter for likely voters who are over 22.

____________________

s.b.:

Bridgeport US voter turnout is not 88.5%.

____________________

BridgeportJoe:

And Gallup's traditional model still assumes a 20% higher RV turnout than at the height of the Vietnam war, but pollster thinks voter turnout might be higher than that and has to use a model with a looser screen?

In 1968 there were 116,535,000 Americans of voting age. 74.3% of them (86,585,000) were registered to vote. 72,957,000 actually voted -- so 62% of all adults, and 84.3% of all registered voters. Gallop's projected turnout for 2008 is a few percentage points lower than that.

Wrong, wrong, wrong.

____________________

s.b.:

If you have to compute something for a university then you are just playing with numbers. I am taking the numbers of people who voted vs the number of people who were registered to vote in the US according to the secretaries of state. It's on lots of web sites. It is not 88.5%. One assumes since there is only one state that does not require registration to vote that these numbers are not something that needs to be computed unless you count addition as computing. Goggle it.

I assume by eligible you mean over 18 a citizen and not a felon in certain jurisdictions. That is not what I am talking about. I am talking about turnout of registered voters.

I assume Gallup is using a list of registered voters. If they aren't then yes it would be self reported and possibly inaccurate, but that would also mean their screen is too loose if unregistered voters are counted in their numbers of LV's.

____________________

BridgeportJoe:

Bridgeport US voter turnout is not 88.5%.

Seriously, are you illiterate? Sorry to be uncivil, but I've given you the exact numbers from Census records. You are just plain mistaken. You thought that the turnout figure was voters/registered voters. It's not. It's voters/eligible adults. Bleating the same disproven point over and over and over again is not very convincing.

Please supply sourced numbers indicating that contrary to the Census data, there weren't approximately 142m registered voters in 2004, and that approximately 126m of them (88-89%) didn't actually vote. Otherwise admit you are dead wrong and let it drop.

____________________

BridgeportJoe:

It's on lots of web sites.

Please supply one with data inconsistent to what has been posted above -- i.e., there were approximately 142m registered voters in 2004, and that approximately 126m of them (88-89%) actually voted. Just one.

____________________

s.b.:

Sorry Michel, I am talking about what you call eligible voters. Why they would be registered i have no idea, well actually I do after this week.

Is what you are saying is that these people who aren't elegible to vote actually vote and then have their ballots discounted?

So, the voters I am referring to are citzens registered to vote who are not felons in certain jurisdictions, who actually vote.

http://elections.gmu.edu/voter_turnout.htm

There's a link.

____________________

s.b.:

Bridgeport Joe, national voter turnout in the US is not 88.5%

____________________

s.b.:

There was just an election in Canada Joe. the voter turnout was 60%. Discussion of this turnout cited that it is exactly the same turnout as in the US. Voter turnout in the US is not 88.5%, never has been.

____________________

sulthernao:

@s.b.

Turnout in 2004 was 60.32% of VEP, not registered voters. Not all VEP are registered voters. Turnout of registered voters was much higher than VEP. Do you dispute that? Because then you are just flat out wrong.

____________________

s.b.:

I assume that elegible voters are registered voters. If you aren't registered then you aren't elegible.

This 60% figure I assumed was the number of registered voters who voted. But I see what you are saying, many of them may not registered.

____________________

sulthernao:

Yep. The way that election trackers look at it is any citizen over the age of 18 is considered part of the "voter eligible population" (for the most part).

BTW, I'd estimate that the turnout this year may reach 65% of VEP.

____________________

wjbill49:

wow, I feel for you Bridgeport

____________________

Stonecreek:

Well, s.b., you've been pretty well flogged here and I don't want to seem to be piling on, but your attempting to differentiate between voters who are enthusiastic about voting due to their positive feelings about a candidate versus their negative feelings about the opponent is a fool's errand. Either set is motivated and either motivation drives turnout.

Among the 70% or so of commited Obama voters exists some significant number of each category -- some are inspired by Obama, while others are primarily registering a negative reaction to the party in power. John McCain's pointing out that he is not George Bush is of little influence on the anti-Republican vote at work in the current election.

I may not disagree with you on the relative motivational power of positive versus negative motivation, however, the "enthusiasm gap" between Obama and McCain reflects both the positive and the negative.

The Republican strategists obviously concentrate much more on negative motivation than on positive motiviation, i.e., the Robocalls and 100% negative TV buys. But despite this, the enthusiasm gap remains, which is yet another steep hill for McCain's flailing campaign to overcome. Making the hill even steeper is the factor that perhaps close to a third of all of the votes will have been cast before election day, meaning his timeline has already begun to slip away.

One more word about the 1968 election: Despite voter emotions having been heightened by the Vietnam war, the choice voters had was between two candidates who, until the very last days of the election, both FAVORED the war. Virtually everyone agrees that had HHH given his Salt Lake City speech a week earlier, he would have won. Added to the vote-depressing campaign, the hero of the anti-war activists, Robert Kennedy, had just been assasinated, and many otherwise motivated would-be voters simply "dropped out" of participating in the system at all. Believe me, I know because I lived that election. And believe me, voters have never been both as motivated, and voting (despite Republican efforts to the contrary) has never been easier, than in this very election of 2008.

You can wish turnout down, and Repubican goon squads can attempt to intimidate voters, but it won't change the fact that voter turnout as a percentage of EVP will hit a historic high this year. JUST LOOK AT THE PRIMARIES, DUMMY! What was that, chopped liver?

____________________

mattn:

I think it's probably a waste of time with s.b., who has a clear reading comprehension problem, but what the hell:

eligible voters = all adult citizens over 18 who are legally allowed to register to vote (e.g. not felons, not dead)

registered voters = all eligible voters who have signed the paperwork making them officially registered (names on the voting rolls), which allows them to be able to vote in the upcoming election if they so choose

likely voters = some fraction of registered voters

So that:

# eligible voters >= registered voters >= people who actually vote

In 2004, about 60% of *eligible* voters voted, which is the same number (different rate) as the 80+% of *registered* voters who voted.

Thus, the current Gallup screen defining likely voters as 82% of *registered* voters seems reasonable.

____________________

s.b., you're not the only one confused as I have had this discussion with many reporters. So, let's all cut s.b. a little break. Maybe some reflection will help clear his/her head.

In 2004, I wrote a memo that lays out some of the issues of self-reported registration vs. actual registration statistics. In particular, look at Table 1:

http://www.brookings.edu/views/papers/mcdonald/20041005.pdf

Regrettably, I did not calculate in this memo the percent of the citizen voting-age population registered to vote according to the CPS. Once we start getting final registration numbers from the early registration deadline states I'll try to update these statistics in a new memo.

[For the interested hardcore here, this article discusses the difference between VEP and the CPS sample universe:

Michael P. McDonald. 2003. "On the Over-Report Bias of the National Election Study." Political Analysis 11(2): 180-186.]

____________________

Aaron_in_TX:

I've been thoroughly confused by this entire thread. This is the first I've heard of 80% turnout of registered voters. However, I have heard of the argument s.b. makes about all pollsters using faulty LV screens vs. what they've used in years prior. I find it hard to believe that all pollsters are "in the tank" for Obama and are not judging likely voters as accurately as they can.

Mr. McDonald seems to half-agree with s.b. that the RV turnout statistics are not terribly reliable.

So my question is, does the percentage of RV turnout matter? McDonald's paper indicates that RV statistics are misleading. In that case, s.b would be correct that pollsters are overestimating turnout for some reason.

____________________

RS:

@Aaron_in_TX:
Gallup says 60% (or 2/3 in LV-II) of the voting-eligible-population (VEP) will vote. See Michael McDonald's brookings link, Table 1 - that says only 80.9% of the VEP is registered.
So an 80% turnout of RVs means only 64% of the VEP votes - which is close to Gallup's "expanded LV" model.

The key to understanding the situation is that not all eligible, 18+, US citizens are registered. Hence the Obama campaign's voter registration drive...

____________________

exrocketsci:

A data point for all of you to consider - the Minnesota Sec of State announced a few days ago that voters in Minnesota reached a historical high of 90% of all eligible voters (your VEP), AND a few days were left to register. AND Minnesota has election day registration. AND I will personally march 5 people I found not registered to the polls. AND this will be done all over the state.

Turn out will be VERY high here. I am not so sure how high it will be in Crawford, Texas

____________________

DCWylly:

@ s.b.

I haven't looked up Gallup's figures, because I don't want to make a career out of this. But here are some ideas about voter turnout.

I believe that this election is going to break records in voter turnout...people are much more energized to vote. The examples you give (2004 and 1968) are simply incomparable to what we are seeing now. 1968 in particular is an awful example...people were alienated...the left was protesting its own convention. In 2004, the level of dissatisfaction with Bush was rising, but not to the levels it is now. Furthermore, there was still a politics of fear which may have kept people from going out to vote against the status quo. Oh...and don't forget that the dems nominated a man with close to zero personality.

My opinion is that, as with the primaries, we are going to shatter the turnout record. I am an Obama supporter but a hopeless cynic (as many dems are). I still think that Obama is going to dominate the ground game because he has armies of volunteers. If anything is going to keep us below 80% in battleground states it is going to be low turnout by republicans.

____________________

RS:

@DCWylly:

NO! Obama is a useless candidate, an empty suit! The youth and African-Americans will NOT turn out to vote because they are lazy! Only Republicans and PUMAs will come out to vote! Obama will lose by a margin wider than Mondale '84!!

/snark... which could just as easily pass for a serious reply from s.b.

____________________

UpstateProgressive:

I think some US Census data can help us here. It seems from this info that both the percentage of eligible voters who voted in 2004 and the percentage of registered voters who voted have been understated by sb:

Sixty-four percent of U.S. citizens age 18 and over voted in the 2004 presidential election, up from 60 percent in 2000, the U.S. Census Bureau reported today. Tables from a November survey also show that of 197 million citizens, 72 percent (142 million) reported they were registered to vote. Among those registered, 89 percent (126 million) said they voted. In the 2000 election, 70 percent of citizens were registered; and among them, 86 percent voted.

sb has suggested that we use 2004 as a model for turnout. If we do, we should expect 89 percent of RV's to vote and 64% of all voters (including those not registered) to vote.

Now one might argue that, given the high intensity associated with this election, the percentage of RVs voting might be up from 2004 (as 2004 was up from 2000). We currently have two wars and an economic meltdown going on, along with severe ideological polarization between the parties. That seems likely to drive turnout higher, especially given, as well, the enthusiasm for the Obama campaign and its unprescedented GOTV effort.

But, even if we take the 2004 numbers, we can expect all but 11% of all RV's to vote. So the percent of them on cellphones that some polls are currently not catching would appear to be relevant. Only time will tell how the 2008 election will turn out, but my guess is that Obama will at least equal, and perhaps surpass by a a percentage point or two, the pollster.com consensus the the national poll on Nov 4.

____________________

UpstateProgressive:

BTW, here's the link to the census data I was citing. Sorry for omitting it.


____________________

UpstateProgressive:
____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR