What can we learn from the polls, if anything?

The polling for the 2008 Presidential Election is all over the map, though generally agreeing on Obama having a lead from anywhere between 4 to 7 points, depending on the poll and how much you trust the average.

I’ve read numerous articles trying to “explain” the polls, though none from someone objective outside of a single polling organization. Last night I found a great article which went much more in depth and was able to look at all the polls collectively without the bias of a pollster defending his or her individual poll. The article comes from RealClearPolitics.com, a site I frequent many times a day, written by Jay Cost on why the polls differ.

Here are the important parts:

I’ve received several emails from people asking about the polls. The national polls do seem pretty variable, so I thought I would toss in my two cents on them.

First, we need a short primer on basic statistics. Real Clear Politics offers an unweighted average, or mean, of the polls. As long as there is more than one poll in the average, we can also calculate the standard deviation, which is one of the most important concepts in inferential statistics. The standard deviation simply tells us how much the polls are disagreeing with one another.

With this stuff in mind, let’s focus on some hard numbers. As of this writing, Barack Obama’s share of the vote in the RCP average is 50.3%. His standard deviation is 2.7. For McCain, whose average is 42.5%, the standard deviation is 2.3. For comparative purposes, I looked at the polls RCP was using from its 2004 averages. For roughly the same time in that cycle (10/17/04 to 10/24/04) Bush’s standard deviation was 1.8; Kerry’s was 1.7. This means that there is more disagreement among pollsters now than there was in 2004.

Well obviously we know there are large discrepancies between polls. As Cost discusses, the pollsters all disagree much more than they did during the 2004 election.

Cost continues:

So, let’s push the analysis a little bit further by looking at specific polls. We can test to see if the polls are separated from the average by a statistically significant amount. Again, since we’re dealing with each candidate’s individual poll positions – we’ll test each candidate’s number in an individual poll against the RCP average. To make sure we dot all our “i’s” and cross all our “t’s,” we’ll supplement the RCP average with a weighted average of the polls, which takes into account the number of observations when averaging the polls together.

Of the fifteen polls in the RCP average, four fall significantly outside the average for Obama and five do so for McCain. Meanwhile, three polls are right at the boundary of significance (one for Obama, two for McCain). The rules of statistics being what they are, we should expect a few polls here or there to fall outside the average by a statistically significant amount. But this is a lot. 40% of all our tests produced results around or outside the acceptable range.

So, we have made three observations: (a) relative to 2004, the standard deviation for Obama and McCain’s polls are high, indicating more disagreement among pollsters at a similar point in this cycle; (b) the shape of the distribution of each candidate’s poll position is not what we might expect; (c) multiple polls are separated from the RCP average by statistically significant differences.

Combined, these considerations suggest that this variation cannot be chalked up to typical statistical “noise.” Instead, it is more likely that pollsters are disagreeing with each other in their sampling methodologies. In other words, different pollsters have different “visions” of what the electorate will look like on November 4th, and these visions are affecting their results.

The bottom line here is that polls showing Obama at +15 could be correct, or they could be very wrong. The same goes for the polls which have Obama at +1, they could be correct or very wrong, nobody really knows.

It all depends on the sampling the pollster decides to use, how many Democrats, Republicans, and Independents they will include based on who they think will turn out to vote.

Something else which is interesting from the Gallup organization:

PRINCETON, NJ — Gallup finds 13% of registered voters saying they will vote for president for the first time in 2008. That matches the figure Gallup found in its final 2004 pre-election poll.

The current data are based on interviews with more than 2,700 registered voters as part of Oct. 17-19 Gallup Poll Daily tracking. Gallup asked these voters a question it had asked in its 2004 election polling: whether this would be the first time they had voted in a presidential election, or whether they had voted for president before. Despite much discussion of the possibility of large numbers of new voters in 2008, the percentage of “first time” voters in Gallup polling this election cycle is no higher than it was at approximately the same time in 2004.

The vast majority of these “new voters” are of the 18-30 age range, they are the “youth” voters who the Obama campaign is counting in, in some areas of the country. This information is interesting since we watched as the “youth” vote didn’t materialize for John Kerry in 2004. It remains to be seen if a similar trend will happen again this year.

Another interesting tidbit from Gallup about the breakdown of early voters:

Gallup reports this morning that about 11 percent of registered voters who plan to vote already have done so, and that they’re split almost evenly between supporters of John McCain and Barack Obama.

The 11 percent early voting rate is just a little higher than the 9 percent who’d voted at this stage in 2004, according to Gallup.

But another 19 percent tell the polling organization they plan to vote before this Election Day, meaning three out of 10 voters would have voted before then.

According to Gallup, the early voting has been very split between Obama and McCain. This is significant since it does not show a landslide of support in either direction, it merely confirms that the race may be tighter than indicated by some polls.

However, it’s difficult to draw concrete conclusion from early voting and we really won’t know until the rest of us vote on November 4th.

  • JD

    Well, I don’t know if any one wants to say if they have or not, but I have voted as of thurday of last week.

    +1 for Obama.

    Maybe we can start a poll here on Youdecide2008.com and compare it against the national polls.

  • rzzzzz

    this is the only site that i’ve seen that has the early voting breaking evenly. i have seen some FLA reports that the large advantage that Obama has in early voting is nearly offset by the number of absentee ballots that have traditionally gone republican, but this still means, overall, an encouraging sign to the democrat.

    i just hope that a lot of folks do vote early to avoid any election day shenanigans that would suppress the vote in any way. hopefully every citizen who wants to will get to cast a vote.

  • rzzzzz,

    It is Gallup which put out the number on early voting breaking somewhat even, though they are speaking of early voting on a national level, which makes sense.

    Some states may be split more one side or another.

    Here’s a direct link to Gallup’s report:

    http://www.gallup.com/poll/111430/Early-Voting-Now-11-Could-Reach-30.aspx

    This is what they say:

    Roughly equal percentages of Barack Obama supporters and John McCain supporters have taken advantage of the early voting opportunity — so far.

    So who knows, they’re one polling organization and the make it a point to say “so far” since we have many days of early voting to go.

  • EricF

    Obama supporters are blowing their load early. come election day the dem vote will be spent. McCain wins in landslide.

  • Bill Hedges

    EricF- Hope you are right. Ex-Clinto Morris said full effect of Joe the Plumber should be seen in polls Monday and Tuesday, if they are good, then we have chance. Of course never know what may come out about Obama. With his unknown past something may become public.

  • b

    EricF that is a pretty funny comment…
    I think it is pretty clear that Obama will win the national popular vote by 4 – 5%, that seems to me to be a conservative mean of nearly all the national polls over the last month really, realclear politics and pollsters collections. The same is true for many state polls, if you look at the trends and the numbers over the last month in VA for example, Obama is ahead in every poll since 10/3 by anywhere from 3-11 points, you gotta imagine that a conservative mean out of all these polls will hold true on election day… say 4-6 points.

    My official prediction (which means nothing… kind of the politcal version of the march madness map) for this election is a landslide at the state level
    Obama: 381 electoral votes. The entire northeast from ME to PA, inclding NH,… MN, MI, IL, IN, OH, IA, MO, ND, MT… VA, NC, the Left Coast including NV, CO and NM.

    McCain: 157

    I would also like to add that if the McCain camp continues to implode and wastes more of its time and money in states like Iowa and New Mexico (which are going to go Dem), then I think that Georgia, Arkansas and Louisiana are gonna come more into play in the last few days for Obama.
    PS can someone tell me why in the hell McCain was wasting his money on adds in Maine up until this week?

    I am 1 more vote for Obama in MO on Nov 4.

  • EricF

    beware the PUMA

  • b

    If I were one of McCain strategists I would scrap the strategy of competing for NM or CO or IA. I would put him on one Straight Talk Express bus in Florida and Palin in another one in OH and go 2000 New Hampshire Primary style, crossing up and down the East Coast from OH, PA area down to Florida until election day. Stop in every town possible, shake as many hands as possible, and instead of trying to match Obama’s mega rallies hold as many small intimate rallies as possible, and finally on each bus have a travelling press pool, and when both of them are on the road just allow open free flwoing q & a sessions. Winning VA, NC, FL, OH, PA and WV will win him the election given that he can also hang onto ND and MT, or win MO.

  • Bill Hedges

    b- I appreciate your concern for McCain’s money. No need to worry, Cindy is rich. Polls don’t always judge what they measasure. Knowing staistcs as you do, you know exactly what I mean !

  • b

    Bill- That is true but the same could also be said for Obama. From what I understand not many of these polls include in thier models first-time voters or young voters, two groups who I think are gonna show up big for Obama and leave a lot of pundits wowed on election night and the days of analysis that will follow.

  • EricF

    just me or did Obama seem a little nervous today?

  • Bill Hedges

    b- I don’y keep up with polls. Don’t know how good a track record they have. You entions some varibles. Add cell phones, caller id that causes some not to answer calls. A good across the board study group may be deficult. News medial channels muct affect too.

  • Pats

    The polls are as unreliable as the people they sample. Don’t trust them because people change with situation and time. The concentration has been on battle states forgetting that the most reliable states can change with the slightest twist of events. Lets leave Nov 4 to decide.
    Right now each candidates are piling a lot of lies against each other. Time will tell.

  • dale

    The explanation by Jay Cost from RCP is very interesting. What I find more interesting than that is that although they know about the extenuating factors as any good statistician should, their ratings system doesn’t reflect them.

    I’m not sure how common knowledge this is but this site, Five Thirty Eight uses a far more sophisticated method of analysing the polling data on a national and state level. It’s run by a guy (or group of guys, I’m not actually sure) who specialise in baseball statistics and use a similar set of statistical models to give a really impressive set of statistics that take in a LOT of variables.

  • Thanks for that link Dale.

    It doesn’t really matter which polls we look at, Obama is well ahead and McCain needs nothing short of Palins witch doctor to perform a miracle for Obama to lose.