Update: New graphs here.
Following what we did for 2000 and 2004, we took the Pew pre-election poll data and broke it down by state and income. The numbers are noisy in some of the states (for example, New Hampshire had 15 respondents in the lowest income category and only 11 in the highest category), so we piped the data through a hierarchical model to get a more stable estimate for the voting patterns among rich, middle-income, and poor within each state. Here's what we got (red and blue states are those McCain or Obama would've won):
Update: My original maps were slightly wrong--they were based on unadjusted poll numbers. I re-posted the corrected maps 6.28pm.
The most striking pattern is our estimate that Obama would've won almost all the states, if only low-income voters were counted. This is our best estimate for each individual state; however, given the uncertainty in these estimates in small states such as Idaho and North Dakota, we wouldn't claim with certainty that there was such a clean sweep.
Among rich voters, Obama won in California and some northeastern and midwestern states--"blue America," if you will. (I don't really know what to say about New Mexico, but there is such a thing as sampling variability!)
The five income categories I used in the analysis are: 0-20,000; 20-40,000; 40-75,000; 75-150,000; over 150,000. The graphs above show the estimates for the highest, middle, and lowest of these five categories. I assume the numbers represent family income (as reported by the survey respondent).
P.S. I don't know what sort of technical detail people want on this blog--I got enough complaints on my last couple of posts--so I'm posting the R code for the fitted model here.
60 comments
Interesting! And wow, Kansas is really sticking out there.
P.S.: I hope many people catch this post. Maybe a little more coordination between Sean, Nate and Andrew is needed ;-)
This is really impressive.
Really? Rhode Island was won for Obama by poor voters alone? I am stunned.
I appreciate the graphs. They're fairly informative on their own and I don't know that any analysis is needed.
Not surprising. Rich people are smarter than poor people (typically) and poor people are fearful (and typically gullible)and found Ovomit's socialist propaganda appealing. Next!
I like the level of technical data that you provide. Good article!
Lesson to be learned - If you want a hand out, vote for the man that will take it away from someone else and give it to you.
Thanks for another insightful post. It's especially intriguing that Obama was so competitive in the midwest with middle income voters. The spin on Palin and later Joe the Plumber (after first claiming she would help with women) was to make the GOP ticket more appealing to those "blue collar whites" Obama supposedly couldn't convince in places like Ohio and Pennsylvania. Yet more evidence that people are pretty good at identifying and voting in line with their real interests, political dress up games notwithstanding.
I, too, am surprised to see Kansas going blue on the middle-income map. How would you explain this?
Lesson to be learned -- rich people who live in intellectually and culturally progressive areas like California and the Northeast support Obama.
Also, if you look at Maine, you can get an inkling into why Senators Snowe and Collins get re-elected. They are moderate enough to get cross-over support, DESPITE the hugh Obama effect.
I may not LIKE Senator Collins, butif she can get re-elected in 2008 with President Obama at the top of the OTHER party's ticket, she has staying power.
PK, Obama won the most educated demographic.
An interesting post with nice graphics that make this site unique. :)
JeffF says...
Hmm.. I wonder how much this would change if the categories were based on state income percentiles rather than absolute dollar amounts.
Median incomes in the US vary from around 66k(New Jersey) down to 35k(Mississippi).
So in NJ a median earner is near the top of the "middle income" category used in this analysis while IN Mississippi the "middle income" category is composed entirely of people making 5k or more more than the state median.
Interesting grouping of graphs, I'd have liked a little more analysis- for instance, what proportion of the electorate is represented in each graph and state? If only the rich voted, how many electoral votes does Obama take? Does this breakdown tell us anything about the future, or on whom the GOP or Dems should focus their attention in 2012?
I'm a lazy reader, I won't glean these things myself.
I like your additions to the site, learning curve for the style notwithstanding.
Thanks Andrew... what about the other 2 income buckets that you didn't include ($20k - 40k, $75k-$150k) -- I would love to see what those brackets include.
A question for Nate: Are you normalizing to compare comparable socio-economic strata in the various states? For example, did you divide the population into quintiles based on national income distribution or state income distribution? If you used national distribution, the lowest quintile would correspond to a larger share of the population in Arkansas, for example, than in Connecticut. The opposite would be true for the highest quintile. If you did not normalize, that would make the maps even more impressive.
Interesting... KS middle classers go dem but RI doesn't. Nice analysis, but I share MWG's concern.
btw... Nice posting the R code (it doesn't mean too much for me, but I appreciate the openness).
Nice post, can you do graphs like these looking only at education levels?
Wow, thanks for posting the R code! Open-source blogging - I love it. It would be great to make this a standard practice.
Thanks for an interesting post. I like me some R code, but I would be very surprised if more than 10% of site viewers feel the same way.
Hey, sorry about Kansas! The comments here and at the other blog served as useful beta-testing and led me back to my code to fix that.
Wow PK, way to grossly (and incorrectly) oversimplify. Never mind that Obama won in every single education catagory, including college grads and those with postgrad degrees. Never mind that he also won among those who make 200K+. Look at the post again. "Poor" voters are voters who make 0-$20000 annually. Now think, what group of likely voters falls into this income range?
Young people! Especially those currently in college or just out of college. A lot of these people who have just started out in the world and will likely be moving up in income as they age.
Of course, given what we know about Party ID, most of them will likely be holding on to their Democratic affliation for a long time.
As far as New Mexico: there's a lot of rich liberals around Santa Fe and Taos, and not many rich people at all in much of the rest of the state.
Pete Kent -
At the end of the day, you're just an old-fashioned dickhead.
Your party has become the party of dickheads.
It started as a party that contained some dickheads.
Then it became the party that contained all the dickheads, but some other people.
Now it's just the party of nothing but dickheads.
And that's why it's dead in the water.
Not surprising. Rich people are smarter than poor people (typically) and poor people are fearful (and typically gullible)and found Ovomit's socialist propaganda appealing. Next!
As recently as 2000, before he became drunk on power, even George W. Bush himself was careful to distance himself from this kind of obnoxious, stupid, self-defeating language.
And yes, I said stupid -
1) You may be rich, although I profoundly doubt it, but half of what you do is repeat memorized right wing soundbites.
2) And the other half of what you do is throw out childishly simplistic yet offense and obnoxious name-calling attacks.
You think referring the the president of the United States as "Ovomit" is going to win people to your cause?
I shouldn't even tell you this - but it's okay, because I know it won't make you stop.
It would be useful to include information about the proportion of the population that is in each of those income brackets. Nationally, but even more useful, state-by-state.
In other words, is it possible to include a proportion of the state's population that is in that income bracket as a little number inside each state?
The maps say they're from 2008 -- but they must be from 1958, since Alaska and Hawaii were apparently not states when this data was compiled...
Here is your takeaway. The states that switched from red to blue form high to mid income and the states that switched from blue to red from low to mid income.
Those are your new battlegrounds and I would bet both parties will focus on these states.
A map for relgion would have helped with your religion post. Use non-religious, non evangelical and evangelical for the three breakdowns.
Also the racial background would be interesting too. I want to see what states Obama won the white vote in if any.
id have to agree with a previous post. new mexico (esp northern new mexico) attracts the working and retired rich more than creating homegrown wealth. i seem to have learned this on fivethiertyeight.com actually, though it was back when richardson was still in the running for the nomination. anyways, that might explain that state.
I would like to see these maps broken down by regular vs decaf coffee drinkers.
Just kidding. This is a great follow-up to your previous post on income & voting. Thanks, Andrew!
Rich people want to stay rich and keep other people down. Sometimes they don't WANT to keep them down, but they just don't want to lift them up.
If rich people were smart, they'd know that the best thing in the long run would be to lift everyone up that they can, because if they don't there would be no one left able to afford the crap they peddle. Alas, most rich people have tunnel vision and only see things as "me want more money, me want lower taxes." Fortunately, there's a whole heap of us middles and lowers around to choose our leaders.
Andrew
Your comment about New Mexico suggests that you think of it as a conservative state. It is not. It is a liberal state with large hispanic and Indian populations. There is, moreover, such a thing as an hispanic upper class in New Mexico comprised in many cases of families who trace their ancestry to before 1848.
Results based on pre-election polls aren't very meaningful, are they? Based on exit polls, where Obama carried the "rich" by several percentage points, surely he should be doing better than the map above shows.
Ahem . . . not to be nitpicky, but aren't we forgetting a couple of states?
Well, anyway, very interesting analysis Andrew; thanks for your diligence. One suggestion to complement the analysis would be electoral tabulations alongside the respective graphs. By my count (which admittedly was hastily done), Obama/Biden were at 207 EV for Rich-Only, but made it to 277 on the strength of Middle-Only.
@ David desJardins
You must have played hooky for the Nate Silver Schooling on Exit Polls.
Very interesting. So, adding a layer of interpretation to the results -- Rush Limbaugh's stategy of hoping that Obama fails will actually promote continuing dominance of the Democratic party. (I know this is counter-intuitive, but just hear it out). If Obama is not able to get the economy on a positive track, people's financial stability will continue to decline, driving more people to vote Democratic. (This seems to be evidenced from your charts. Also, if the economy continues to decline, people will want to vote for the party that is more concerned about providing an econimic safety net.)
If, on the other hand, the economy revives and people's economic outlook begins to look rosy again, people will be more inclined to vote Republican. (Again, this is evidenced from your charts. Also, as people do better economically, they become more focused on retaining their good fortune via lower taxes).
So, bottom line -- the best longer term outcome for Republicans is that Obama is successful in turning around the economy.
I'd be interested in hearing people's thoughts on this.
harold and Statler, I understand your indignation, but spewing the same kind of invective back is not the way to move this country forward. Please either stick to facts or don't feed the trolls.
nova_middle_man, I've got a map with demographic information on it, including racial breakdowns at
http://maps.webfoot.com/demos/election2008/
There are some nice maps of religion at
http://www.valpo.edu/geomet/geo/courses/geo200/religion.html
(I would like to add those to my demographic map, but I am awaiting permission from the Glenmary Research Center to redistribute that data.)
Re New Mexico: There is a layer on my map which shows the combined parties of governor, senators, and representative mapped that by House District. (4 Dems is dark blue, 3Dems, 1R is light blue, 2-2 is white, 3R1D is light red, 4R is dark red.) I was surprised to see how few places had all Dem or all GOP, but northern NM was one of them. Southern NM was 3 Dem/1 GOP. It looks like NM is more liberal than I would have guessed.
Ruth -
So, bottom line -- the best longer term outcome for Republicans is that Obama is successful in turning around the economy.
I'm sure this would be true, if the Republican party were still able to function as a party that turns the middle class against the poor.
However, that's unlikely to happen, because in order to kowtow to bigots and nutjobs, they need to insult increasingly large segments of the middle class.
At this point, they certainly lose if people get poorer (especially since they've been crowing their hopes for that).
But they also lose if people get richer, because they have to offend a majority to retain their funders.
Looks pretty grim.
I'd compare them to the Whigs, but that's an insult to the Whigs.
I find it interesting that the map based on rich Americans only corresponds so closely to the image of the US presented by pundits TV talking heads etc (not just David Brooks). The middle class map maybe maybe too.
It is as if poor people didn't exist for US opinion leaders (you know the ones who call a plan to increase the taxes of 2% and reduce the taxes of 95% a "tax increase").
By my count, the middle-income map yields 291 electoral votes for the Dems and 247 for the GOP. Not bad.
I like the technical detail. The comments section is always an awful mess, so just ignore the flames from the peanut gallery. I've enjoyed all your posts so far, and I don't want you to dumb them down because of some vocal ignorant blowhards.
As one who retired to Santa Fe, I agree with the comments that others have made about New Mexico, and would add a couple of points. Santa Fe and Taos were mentioned, but Los Alamos, one of the wealthiest and best educated counties in the nation, also contributes to liberal strength in northern New Mexico. And somehow or other, in 2008 even the conservative south-eastern part of the state, often called "little Texas", managed to elect Democrat Harry Teague to Congress.
Wow, thats unbelievable. Pennsylvania is a true blue state...more blue than NJ, MN, IA, and WA. Amazing...lol.
Rich, middle and poor, all vote blue in PA...for the most part I guess.
It'd be interesting to see the results in PA sans the 5 county Philadelphia area (MontCo, DelCo, ChesCo, BucksCo, PhilaCo).
I like being able to look at the technical detail if I like. So to keep both sides happy I would suggest doing what you did in this article where you have the synopsis of the results on the main page and a link to even further details if the reader would like to look at them.
Thanks for the links duckysherwoood
I find this fascinating, but wasn't keyed into this level of analysis for the 2000 or 2004 elections. I'm writing an essay at the moment for which these figures would be extremely helpful, so if anyone could link me to them I would be hugely grateful.
Bravo, Statler! Enjoyed your entertaining encyclopedic erudition emmensely [sic]. :-)
To be fair, you did not direct your insults only at Steele: you also mentioned something about emasculating Peter Kent being a waste of time. I would be happier if people didn't feed the trolls because I don't want the comments here to degenerate into flame wars. (And this is the absolute last I will say on this topic.)
You asked "When they do it, it's entertainment-and when I do it it isn't?", basically: how come they can do it and I can't?
a) I'd like us liberals to be better than them conservatives. I would like people to think "reasonable" when they think of "liberal" and "wacko nutcase" when they think of "conservative".
b) The more we liberals toss around ad hominem insults, the more conservatives can justify their ad hominem attacks. We could try to say, "But they started it!", but that would be futile. Insults have been flying around since the founding of the country, what's important is what people are doing NOW.
I recognize that these are personal opinions about strategy. The request for civility that I made was a request. I'd like to hope that you would at least consider it.
Following up with Matt Hare, I too would be interested in getting the data. I'd like to put them into my maps...
Do you have numbers on Alaska and Hawaii?
Rich are smarter than the poor? How much you have in your pocket or what degree you have on your wall doesnt denote intelligence. There are plenty of uneducated rich people and some very intelligent and educated poor people. Inherited money happens even in this country. Didnt the GOp try to claim that socialism is an elitist policy? Then how can the socialist agenda cater to the poor when its claimed that Obama and his buddies are elitists. Also according to exit polls, Obama won the $250,000 plus category. The same category he wants to tax. Talk about interesting. If socialism is for poor people, than why are the wealthiest and stable nations like Denmark, Swedan, Finland, Norway, etc all socialist? BTW Obama is not a Socialist, he's a Communitarian. Get schooled in political culture and terminology.Both parties have Socialism in their beliefs. Any time you feel the government has to regulate. DEMs want to regulate things like businesses , economy, schools, while GOp wants to regulate families, sexuality, religion, patriotism, etc.
Thanks for this. Nifty stuff! (and incidentally, I just purchased your books on Data Analysis: Using Regression and Multilevel/Hierarchical Models as well as Red State Blue State - thanks!)
I'm wondering - do things look any different if you were to use relative income? That is, does the idea that being wealthy in NY actually means something different (greater income required) than being wealthy in Mississippi matter? Presumably if relative income were used (say local income deciles by state), the income patterns might show up even more strongly. The truly rich (top local decile) in NY might vote republican. The truly middle class in Mississippi (mid-decile) might be more on the fence. Cheers.
Andrew, thanks so much for this great post! I really enjoy your contributions to this site.
Andrew,
Thanks, this is very interesting. I am curious where you got the data on income of voters across states. Are you just using the average income of a state? Or are you using data from polls?
I'm startled by the rich/Washington State result. Stereotypically, Washington State politics have an east/west Republican/Democratic split, and I'd have said that by far more of the rich people were in the west (the coastal cities--Seattle, Bellevue, Redmond).
Maybe ethnicity is a confounding factor.
Anyway, cool maps--thanks! I second the request for a similar breakdown by religion, if the data are available.
New Mexico's center is more to the left than in most states - we're small, population wise, and so the combination of Taos/Santa Fe is nice for making what would otherwise be a small fringe a viable constituency. Los Alamos, the per-capita wealthiest part of the state (and per-capita the most doctorates of anywhere) is interesting - it's partly an energy lab, but it's mostly a straight-up nuke lab, making it favor big Defense, but also fairly reliant on forward-thinking administrations of any party. So they get to occupy the NM center, along with the center-left that comes from a hispanic and catholic population.
It's generally a low income state, so the poor are a strong electorate, and that motivates all politics to the center-left.
And then there is Albuquerque. With a vocal libertarian anti-Bush right existing in the state and being present in the city itself, the right has less room to maneuver. The left-left candidates for mayoral races all come from the hippier enclaves/progressive districts within the city, and the open election system means that the Mayor of Albuquerque for 16 of my twenty years in the city has been a Clintonian centrist. His constituency is developers, in the personal-politics government + big business collusion kind of way you see in some democratic systems. There's big money in the center left in NM, and the right itself tends to be a bit more scattered.
At least, that's my best guess for explaining why Wealthy NM went blue. I mean, besides the fact that we're awesome.
Make one for voters by education level and by sex please! This is fascinating.
Fascinating yes but factually sound??
You say "The numbers are noisy in some of the states (for example, New Hampshire had 15 respondents in the lowest income category and only 11 in the highest category), so we piped the data through a hierarchical model to get a more stable estimate for the voting patterns." My take is that whether you pipe it through coaxial cable, the Holland tunnel,a bong,or any "model" when you have 11 or 15 responses you don't have "data' you have entrails to examine. Statistics doesn't clean up the GIGO problem. The best you could do I'd think would be aggregate those number with same income block in a state that seemed culturally & ethnically similar (e.g. Maine, Vt.) -- but then you'd just have a regional (not state) class estimate. Please enlighten on how model makes wildly inadequate sample meaningful. --Wayne Rosalie
Post a Comment