I’m afraid I have to take exception to this statement by Goldy:
Oh… and the fact that polls generally show Edwards as being the toughest Democrat to beat… that doesn’t hurt him in my book either.
I suppose Goldy is relying on national head-to-head polls like these. The problem with such national polls is that they don’t reflect the way we elect our Presidents.
Rather than looking at the national head-to-head polls, we should be examining state head-to-head polls and take into consideration the number of votes each state gets in the Electoral College.
In fact, I have been doing just that for a number of months. Essentially, I’ve collected the state head-to-head polls taken in 2007 and have been analyzing the polls as a way of evaluating the relative strength of candidates.
Now I am going to switch into statistical wonk mode and explain my analyses. If you just want to see the results, skip over the Methods section and pick up from the Results.
Methods
To analyze the poll data I take the last month of polls for each state as a way to increase the certainty and (hopefully) minimize biases inherent in individual polls. If there is no polls taken in the last month, I use the most recent poll available in 2007. The analysis could stop at this stage after simply tallying the number of Electoral College votes each candidate would receive for each state based on the poll data.
The one problem with this approach is that it doesn’t account for the uncertainty in the polls. For example, suppose a poll in Pennsylvania of 500 individuals gives Clinton 51% and Giuliani 49% of the vote. Clinton’s lead comes from only five individuals who went for Clinton instead of Giuliani. In fact, statisticians would tell us that there is substantial sampling error because of the small sample size and the very close percentages. The statistician would do some calculations (or simulations) and tell us that the poll indicates that Clinton has only a 69.9% chance of winning, and Giuliani has a 30.1% chance of winning.
In simulating a national election, I do this same evaluation over all states. Here is how it works. I simulate elections using only information from state head-to-head polls (with one exception discussed below). Each single election proceeds state by state, pooling polls from the last month (or the most recent poll if no polls were taken in the last month). For each person polled in the state, I randomly draw votes according to the observed probabilities found by the state’s poll(s).
After conducting such elections in all fifty states (plus Washington D.C.), the electoral vote is totaled and a winner determined from the electoral vote count.
This process is repeated 10,000 times. The result is a distribution of electoral votes for the pair of candidates that fully accounts for the sampling error in the polls used. For example, here is the distribution of electoral votes for a Clinton—McCain match-up from a few days ago:
In this example Clinton won 9,167 simulated elections and McCain won 779 simulated elections. (There were also 54 ties that would go to the House of Representatives and almost certainly result in a Clinton victory.) Thus, the poll data suggests that, if the election were held today, Clinton would have a 92.2% chance of beating McCain.
Oh…about that exception I mentioned above. Some states have had no polls taken at all. In that case, I always assign the electoral votes for the state according to the 2004 presidential election outcome. For the most part, states that have had no polls taken are not likely to hold any surprises. In any case, this procedure slightly favors the Republican candidate (since Bush won in 2004).
Results
Here are the results after simulating a variety of match-ups. (Additionally, I provide a link to my most recent analysis. In most cases the published analysis is slightly older than the analysis from today given in the table below, but the numbers are close.)
| Republican | Democrat | Probability the Democrat wins | Average electoral votes for Democrat | Link |
| Giuliani | Clinton | 100% | 342 | Analysis |
| Huckabee | Clinton | 100% | 335 | Analysis |
| McCain | Clinton | 92.1% | 293 | Analysis |
| Romney | Clinton | 100% | 385 | Analysis |
| Thompson | Clinton | 100% | 354 | Analysis |
| Giuliani | Edwards | 4.90% | 237 | Analysis |
| McCain | Edwards | 99.4 | 303 | — |
| Romney | Edwards | 100% | 388 | — |
| Thompson | Edwards | 100% | 358 | — |
| Giuliani | Obama | 27.7% | 258 | Analysis |
| Huckabee | Obama | 88.7% | 277 | — |
| McCain | Obama | 4.4% | 237 | — |
| Romney | Obama | 100% | 376 | Analysis |
| Thompson | Obama | 100% | 329 | — |
Right now Clinton does better against Republican challengers—she beats every one of them with a high degree of certainty. Edwards does very poorly against Giuliani, although he does a little bit better than Clinton against McCain. Obama doesn’t do well against either Giuliani or McCain right now.
Keep in mind that the analysis only suggests what would happen if the election were held right now. (Interpret this the way you might the speedometer on a long trip—it gives you some idea of your progress even though you know your speed is going to change along the way.)
Things will certainly change in the next ten months, but what we can say now is that Clinton has some advantage over both Obama and Edwards in a general election. Is Clinton’s advantage right now important in the long run? It’s hard to say. It’s not even clear to me that her advantage should be considered over more fundamental characteristics like political philosophy and policy positions. Perhaps some readers will use this information as a tie-breaker.
As for me? I still have no idea who I will support at tonight’s straw caucus. Maybe I’ll pretend to be a Republican….
Slap ’em silly with the bootstrap, Darryl!
P.S. got a comparative analysis of McCain-Clinton vs McCain-Edwards?
I think Ron Sims is the best candidate.
Did you see what Ron is capable of in that story on the Maple Valley Donut Hole property shenanigans??
http://seattletimes.nwsource.c.....le02m.html
Pure BS….but I expect you Lefty’s to come running to Ron’s defense. Sounds like the taxpayers are getting the short-end of the stick…as usual.
Uh huh. Sounds like same kind of “analysis” that lead many to conclude that Dewey was sure to defeat Truman. Way too many variables unaccounted for, not to mention unknowns, for this type of charting to make much sense. Time to sideline such navel-gazing; please
The strongest Dem? HRC can bench press Obama and Edwards plus 75 lbs…with her eyelashes.
Just don’t ask her to do it in Pakistan.
The Piper
RonK @ 2
“got a comparative analysis of McCain-Clinton vs McCain-Edwards?
The table shows the summary of the analysis for both of those match-ups. There is a link to a recent full analysis of McCain v. Clinton. I haven’t yet posted a full version of the McCain v. Edwards analysis yet, so the summary will have to do.
Piper Scott @ 5,
Was that an attempt at humor or wit? If so…very lame.
If not, I hope you get the professional help you need.
You’ve got to be kidding me. A lot of the Edwards/Obama vs. Giuliani stuff just needs to be thrown out, simply because there’s no freakin’ way Giuliani turns NY, CT AND MA red running against them as opposed to Clinton. Not. Going. To. Happen.
@7…Darryl…
Thank you for your advice. Unfortunately, please check your academic credentials since I’m fairly certain you’re not that kind of doc – is it OK to call you doc, doc?
Actually, the kind of doc you are precludes you from comprehending either wit or humor since stuffiness is de rigeuer for academic pomposity to pass itself off as “sophisticated.”
Were you denied Mad Magazine as a child and now express your anxiety and resentment by refusing to laugh at jokes cracked by all save yourself? Alfred E. Neuman has nothing on you!
The Piper
I’d also say that at this point, 10 months out, you are measuring favorability and name recognition more than electability. The candidates Obama and Edwards are getting blown away by are the Rs with the best name recognition (Giuliani and McCain), but once Clinton comes into the picture that goes away.
The way I would read your analysis is, if you figure Clinton as proxy for “Generic Democrat” and McCain/Giuliani as proxy for “Generic Republican”, that right now the D’s stand a good chance of cleaning the R’s clocks in the electoral college.
At this point, I’d also argue Giuliani is very close to being toast. He’s fading rapidly in a lot of polls. The most problematical candidate for the D’s is probably McCain (as he’s the one guy they have other than Giuliani who can pick up independents in any sizeable quantity, and he’ll help a lot in the Mountain West, places that are quite vulnerable to D pickup in the Electoral College). So you are picking up some stuff, I think…
Kiss my Ass
The Tommy
Amazing analysis, Darryl — thanks!
I’m getting to support Edwards too.
Question: I find the fact that Edwards is polling in the 90-something percentiles against other Republicans… but drops to about 5% for Guiliani – quite weird.. What explains that?
Piper Scott @ 9
“Thank you for your advice.”
What advice?
“Unfortunately, please check your academic credentials since I’m fairly certain you’re not that kind of doc”
What kind of doc would that be?
“is it OK to call you doc, doc?”
In fact, the proper title would be Professor.
‘Actually, the kind of doc you are precludes you from comprehending either wit or humor since stuffiness is de rigeuer for academic pomposity to pass itself off as “sophisticated.”’
Oh…that’s fucking rich coming from the biggest pompous phony in the comment threads. Or was that an attempt at satire? (If so, you got the irony part right…you still need work on delivery.)
“Were you denied Mad Magazine as a child and now express your anxiety and resentment by refusing to laugh at jokes cracked by all save yourself? Alfred E. Neuman has nothing on you!”
No plenty of access to Mad, National Lampoon, etc. What was lacking was anything funny in your comment @ 5. Sorry…like so much of your attempts at humor, it fell flat.
10
I think your analysis is on point. McCain is the only one of the Rs that scares me. He actually has a few principles, and I’d like to have a beer with him. With regard to the rest of them, I don’t even believe I’d shake any of their hands. Your point about the Mountain West makes a lot of sense. Ds could take New Mexico, Arizona, Nevada, Colorado, and maybe even Montana. I don’t think that’s true if McCain runs for the Rs.
Interesting.
When drawing per-state votes, what kind of distribution are you using? Also, have you seen name recognition numbers for the candidates? Right now, the results must be heavily skewed by uninformed voters who, by election time, will at least know the R and D candidate names.
Still, you might expect that Hillary’s name recog. advantage would be wiped out by her supposedly negative public image.
eponymous coward @ 10
“I’d also say that at this point, 10 months out, you are measuring favorability and name recognition more than electability.”
Not really…it is measuring who would win a general election today based on people’s stated preferences. Almost all of that is partisan affiliation, as can be clearly seen by the strong correlation in support among different candidates of the same party in a given survey. The few percentage points difference among like-party candidates is based on a complex mix of public perception, sexism, religious and ethnic bigotry, name recognition, prior framing, and a bit of genuine political evaluation. All of these factors will be in play in ten months, as well, but in somewhat different proportions.
“The candidates Obama and Edwards are getting blown away by are the Rs with the best name recognition (Giuliani and McCain), but once Clinton comes into the picture that goes away.”
But Edwards (as a recent VP candidate) should have much better name recognition than Obama. He should probably be doing nearly as well as Clinton if it were only about name recognition. Likewise, Fred Thompson should be doing much better, as he is well known from his acting gigs. And John McCain should be doing much better than he is. In some of my analyses, I show the trends over time, and there is clear evidence of preference switching even for well-known candidates. For example, this trend graph shows that Giuliani peaked relative to Clinton in early summer. I think this was a result of Clinton successfully overcoming years of Wingnut smears and people getting past the “9-11 hero” image for Giuliani—both things that are not about name recognition, per se.
“The way I would read your analysis is, if you figure Clinton as proxy for “Generic Democrat” and McCain/Giuliani as proxy for “Generic Republican”, that right now the D’s stand a good chance of cleaning the R’s clocks in the electoral college.”
Yep…I agree completely that that is a reasonable macro-interpretation.
“At this point, I’d also argue Giuliani is very close to being toast. He’s fading rapidly in a lot of polls.”
I hope you are right, but I doubt it. Giuliani is still ahead in the national GOP primary polls. Furthermore, he is ahead in some important (and some big) states like Nevada, Florida, California, New York, Pennsylvania, and New Jersey.
“The most problematical candidate for the D’s is probably McCain (as he’s the one guy they have other than Giuliani who can pick up independents in any sizeable quantity, and he’ll help a lot in the Mountain West, places that are quite vulnerable to D pickup in the Electoral College). So you are picking up some stuff, I think…”
Yep…McCain is definitely the threat here. Republicans have only recently started to realize it, even though he has always matched-up against Dems more strongly than any other Republican. Part of the problem is that he looked like an old fool supporting “the surge” and Bush this summer, he sucks at fundraising which looked embarrassing earlier this summer, and the fundies really don’t like him at all.
I suspect his recent resurgence is more about an implosion of the rest of the Republican field….
Weird, a liberal posting on a Liberal site doing an analysis of poll data and the Dem wins almost 80% of the time. Didn’t you guys predict the same in 04 and during the Reichert/bruner race? Why not try for some actual objectivity? The thing about statistics, are if you look at them enough, you can see whatever you want.
Darryl, have you run this exercise against poll data from the 2004 election? How did the model do at predicting the actual outcome?
It would be interesting to hear your evaluation of this method’s predictive capabilities.
The Giulani-Edwards and McCain-Obama numbers sort of jump out. Why would Giulani beat Edwards so badly and Obama pretty handily and strike out against Clinton? And why would McCain beat Obama so badly and strike out against both Clinton and Edwards? It is not like any of these Democrats have vastly different positions, it is more a matter of shading and emphasis. Assuming all the calculations are correct and I’m sure they are, I would question some of the underlying poll data.
Scotto @ 14
“When drawing per-state votes, what kind of distribution are you using?”
It is, effectively, a binomial distribution. With large numbers I would draw the appropriate quantile from a beta distribution. But since the poll numbers are typically small within a state (i.e. usually less than 1,000), it is fast enough (and more fool-proof) to simply loop through each of the “voters” as a series of Bernoulli trials. So that’s what I do.
“Also, have you seen name recognition numbers for the candidates? Right now, the results must be heavily skewed by uninformed voters who, by election time, will at least know the R and D candidate names.”
See my comment at 15. My philosophy is that these results are what we would get if the general were held today. There will always be a name recognition effect and a negative framing effect. What we hope is that at least some of the voters will get past these things, learn about the candidates’ positions and philosophy as the season goes on. (Clearly this is the part that failed in a big way for 2000 and 2004!)
“Still, you might expect that Hillary’s name recog. advantage would be wiped out by her supposedly negative public image.”
Yep. She has clearly shed some of that image since June. (See the trend graph linked to in comment 15). I doubt her name recognition has changed much over the same period.
IAFF “Fireman” at 16,
“Weird, a liberal posting on a Liberal site doing an analysis of poll data and the Dem wins almost 80% of the time. “
Nothing weird about it. The polls are all publically available. (I.e. You are free to replicate my analyses and you would get the same results.)
“Didn’t you guys predict the same in 04 and during the Reichert/bruner race?”
No. There was no Reichert—Burner race in 04. Here was my last analysis of the polls in the 2006 Reichert—Burner race:
“Why not try for some actual objectivity?”
I know it is hard for you Wingdings to understand this, but the actual numerical analyses (and the direct interpretation of the analyses) I do are done completely objectively. That said, commentary around the analysis is sometimes partisan.
“The thing about statistics, are if you look at them enough, you can see whatever you want.”
I think this speaks more about your ignorance of statistics than anything else….
Ah. The pipe has blesed us all over again.
Go Willard!
Go Hickabee!
@20…Darryl…
The problem with head-to-head matchups and stats today are that they are…today, not relevent to November 2008.
Heard this not too long ago about HRC…to the effect that her positives drop and negatives rise the more people get to know her or are reminded what they already know about her.
Keep in mind that 90% plus of the American people aren’t paying attention to this stuff right now, and won’t until maybe late summer ’08. How many people will participate in the Iowa caucuses tomorrow? Less than 300,000? In a state with a population of 3 million?
Getting too bogged down in predicting this stuff right now is like predicting which cards will fall face up in a game of 52-card pick-up. Wonks and hard core political junkies live and die on this stuff – it’s mother’s milk to me – but most people think those of us who fixate on this stuff are completely off our rockers!
So, don’t get too far out on a limb with predictions that, Friday, will sound foolish and meaningless. This from an award winning political prognosticator!
The Piper
Oneman @ 17,
“Darryl, have you run this exercise against poll data from the 2004 election? How did the model do at predicting the actual outcome?”
I have, and one of these days I’ll do a detailed post on it. The Reader’s Digest Condensed version of it is that there was a shift from a D win to an R win a few days before the election. My read of the numbers is that the Republican’s voter contact system was highly effective and made a real difference in 2004.
“It would be interesting to hear your evaluation of this method’s predictive capabilities.”
If multiple polls are available for swing states (as actually happens just before the general election), the method should be pretty good at (1) minimizing some biases inherent in polls, and (2) characterizing the uncertainty in the poll-driven outcome.
There are some biases that will get worse with time…like land line telephone-based polling becomes less representative. Also, last minute events (like a tape from OBL) can shift the election results in a way that is too fast for polls to measure.
Piper Scott @ 22
“The problem with head-to-head matchups and stats today are that they are…today, not relevent to November 2008.”
Did you think that up all by yourself? Or did you read it in my post above? :-)
“Heard this not too long ago about HRC…to the effect that her positives drop and negatives rise the more people get to know her or are reminded what they already know about her.”
Empirically, the exact opposite seems to have happened. For example, this graph of her performance versus Rudy, or this graph of her (relatively constant) performance against McCain, or this graph of her performance versus Fred. (Unfortunately, I haven’t generated a similar graph for Clinton v. Romney, and there is not enough older polling data to generate a graph of Clinton v. Huckabee).
“Keep in mind that 90% plus of the American people aren’t paying attention to this stuff right now, and won’t until maybe late summer ‘08.”
Yep…but the majority can still rank candidates in the state match-ups. Recent polls typically have only 10% of people who cannot or won’t offer an opinion when asked.
“How many people will participate in the Iowa caucuses tomorrow? Less than 300,000? In a state with a population of 3 million?”
This comment is completely irrelevant to your other point.
“Getting too bogged down in predicting this stuff right now is like predicting which cards will fall face up in a game of 52-card pick-up.”
But, I am not predicting the outcome of the 2008 election, am I? I mean if you read the post, that ought to be clear.
“Wonks and hard core political junkies live and die on this stuff – it’s mother’s milk to me – but most people think those of us who fixate on this stuff are completely off our rockers!”
Agreed!
“So, don’t get too far out on a limb with predictions that, Friday, will sound foolish and meaningless. This from an award winning political prognosticator!”
Since I only predicted “what would happen if the general election were held today” I feel we are on the same page here.
@24…Darryl…
“Since I only predicted “what would happen if the general election were held today” I feel we are on the same page here.”
But that’s the problem with all these polls…the election won’t be held today or tomorrow or anytime soon; in politics, a week is a lifetime, so we’ve got more lifetimes than a pack of wild dogs before anyone casts a ballot.
So…how soon before there’s a gambling scandel involving UCLA football?
The Piper
Piper Scott @ 25,
“But that’s the problem with all these polls…the election won’t be held today or tomorrow or anytime soon; in politics, a week is a lifetime, so we’ve got more lifetimes than a pack of wild dogs before anyone casts a ballot.”
Indeed…in the same way your car speedometer does not typically tell you how many miles you will go in the next 60 minutes. Many of us find value in speedometers nonetheless.
I presume nobody is holding a gun to your head forcing you read my analyses, so if the analyses don’t offer you any insight, feel free to skip over them.
@26…Darryl…
“I presume nobody is holding a gun to your head forcing you read my analyses, so if the analyses don’t offer you any insight, feel free to skip over them.”
Well, the Gilligan’s Island film festival isn’t running this evening, so I wanted something on that level…ergo I’m playing with you, Lil’ Buddy!
The Piper
“Well, the Gilligan’s Island film festival isn’t running this evening, so I wanted something on that level…ergo I’m playing with you, Lil’ Buddy!”
Ah, a mouse who believes it is chasing the cat! Or, are you really Bob Denver’s bastard child?
ooooo Piper said ” ergo ” – is this your deluded interpretation of of a Big Voccabulary? Pooper, you think you talk a Big Game ( kinda like Dorki Monson ) but you’re easily self-impressed and truly are a waste of bandwidth.
“I think this speaks more about your ignorance of statistics than anything else….”
Not really. For example, Push polls. Or if your statistics were to examine polls that are looking for a democratic winner.
You specifically site Survey USA polls, but during their California Prop 76 Poll, the outcomes were vastly different when they used a three sentence description versus a 1 sentence description. So while I may be ignorant of statistics, it seems you might be a little ignorant on about your own methods. The simple fact is that statistics can be manipulated.
“I know it is hard for you Wingdings to understand this, but the actual numerical analyses (and the direct interpretation of the analyses) I do are done completely objectively. That said, commentary around the analysis is sometimes partisan.”
Sure, perusing your blog shows the objectivity. Look, I admit that my political beliefs actually influence how I interpret certain information. Why is that so hard for you to admit?
@28…PTBAA…
Long time no annoyed by you (said marginally in jest, you know!)!
Only if Bob’s estate was flush with assets…Besides, who wouldn’t want to be the illegitimate son of Maynard G. Krebs?
The Piper
@30…IAFF-F…
“Sure, perusing your blog shows the objectivity. Look, I admit that my political beliefs actually influence how I interpret certain information. Why is that so hard for you to admit?”
Because conceit is a fault, and PROFESSOR Darryl obviously has no faults.
The Piper
Piper,
Because conceit is a fault, and PROFESSOR Darryl obviously has no faults.
Not that you should stoop to allowing a few facts to interrupt a good diatribe (or should that be “wit and charm”?), but you were the one who made an issue out of the official letters after his name. At that point, correcting your misstatement isn’t exactly conceit.
But, I will grant that you are an expert in conceit.
re 31: Bob Denver was gay.
IAFF Fireman
“Not really. For example, Push polls.”
Umm…push polls don’t publish results. They are designed to change the mind of the “pollee,” not to generate polling results.
“Or if your statistics were to examine polls that are looking for a democratic winner.”
The poll data I’ve collected are well known pollsters who use very standard methods. The polling community is pretty quick to identify fraudulent pollsters. There are even well known partisan pollsters (i.e. they typically work for one side) whose results are highly respected.
“You specifically site Survey USA polls, but during their California Prop 76 Poll, the outcomes were vastly different when they used a three sentence description versus a 1 sentence description.”
I recall this happened with Survey USA in a recent Washington state initiative. Yet, they were quick to point out their error and re-poll. The fact is, Survey USA is a highly respected pollster in the polling community. One other comment on that: the language for polling on an initiative is a much thornier problem (and less standard) than doing candidate preference polls.
“So while I may be ignorant of statistics, it seems you might be a little ignorant on about your own methods.”
Any examples of that, or are you just pulling that out of your ass?
“The simple fact is that statistics can be manipulated.”
Sure…bias in a poll or a pollster is alway an issue and there is a good community of polling professional and enthusiasts that makes it pretty easy to learn the reputation of different pollsters. BTW: One argument for pooling multiple polls for a state is to minimize any such biases.
“Sure, perusing your blog shows the objectivity. Look, I admit that my political beliefs actually influence how I interpret certain information. Why is that so hard for you to admit?”
I believe I did admit that my commentary is frequently non-objective. The poll analyses, however, are done by a completely objective numerical simulation done on the computer. The computer doesn’t care if it is “rolling the dice” for a D or an R. The program simulates elections based directly on data from the polls and writes most of the web page, colors the map, and lists the state results. For my part, I post links to each new poll I include in the database, so the methods and data are completely transparent for anyone to scrutinize. So, I think it is fair to claim that the analyses themselves are completely objective. And I have incentive to be squeaky clean, since it is trivial for anybody with a little statistical knowledge and programming skills to verify my results.
Piper Scott,
“Because conceit is a fault, and PROFESSOR Darryl obviously has no faults.”
Crap…do I have to explain everything to you? IAFF Fireman wasn’t talking about admitting faults. He was discussing admitting bias. As I point out in the comment above, I freely admitted to bias in my commentary (Yet, I am scrupulous in keeping the poll analyses objective.)
Try to keep up.
@3 Anything Ron Sims does is insignificant compared to the havoc you Repukes have wreaked on the United States.
@9 This is your answer to sober analysis? No wonder you’re self-employed — and not as a lawyer.
@16 See #38.
I see my rib has been on this site while I was at DL … 37 through 39 were posted by Roger Rabbit.
rib = rabbit in burrow
If Obama wins this evening, the state tracking polls will change dramatically. Same for NH. All that came before was interesting but will no longer mean anything.
Every Democratic contender promises to change America..If change is what the Americans want..We challenge all the activists who talk about change to put their money where their mouth is and support Vladimir Putin for American President
Please, I’m tired of hearing we need a change! I don’t see anyone here talking about the past wasted 8 years and fixing what was neglected. To me this needs fixing and then it’s time to look at change. No one is addressing this past mess the little shrub has us in. No one standing up there has a clue of the mess they are truly walking into, believe me!! Let’s get serious about the needs of the American People, jobs that were shipped over sea for profit, bring our guys home and take a serious look at the enviroment, curb some of the control that was let slide to the wrong folks and the list goes on. Gov for and by the people, not some power hungry folks that hide behind closed doors and take our rights from us.