Death, Freedom, and Cold Winters

July 13th, 2009 by Christian Rudder

First of all, thanks everybody for your comments and emails over the last couple weeks! Just to know that so many people have taken the time to read our writing and question our intelligence is an honor beyond measure.

Today, we’re going to revisit the mapping program from our first post two weeks ago and discuss some new questions we’ve plotted, starting with the below plots.

For anybody who didn’t see our previous map post, Rape-Fantasies and Hygiene By State, these show the responses of OkCupid users to selected user-submitted match questions. States answering “Yes” more often than the national average are greenish; states more often saying “No” are reddish. Yellow states are near the mean yes/no proportion.

Are some human lives worth more than others?

268,864 people have answered

This graph struck me right off because our map-making program is supposed to color the states from solid green to solid red, and there’s no true red on this map. This had Chris and I confused for a while until we realized: the true red is Washington D.C.; you can barely see the little dot there by Virginia. We’d forgotten that our Google Maps API plots D.C. as a separate data set. It’s the most ‘brotherhood of man’ place in America. Weird, huh?

I looked at this graph for a while and realized that the areas more likely to value some lives over others can be generally summarized as follows: it’s the Mountain and Pacific time zones plus the former Confederacy…

OkCupid Rules

Anyhow, let’s compare that first map with this guy:

If you knew for sure you would not get caught,
would you commit murder for any reason?

359,761 people have answered

Despite the fact that it’s located in Minnesota, Minnesota actually seems like a nice place to live. It’s the only state to come out much more humane than average on both charts. On the other hand, North Dakotans are strange: they’re apparently more for the equality of life, but also more for killing. These men are nihilists.

Overall, the Rocky Mountain states are the most into “getting away with murder.” This shouldn’t surprise us, given the results of map #1 and the heavy shit that went down in Cliffhanger:

OkCupid Rules

Now let’s look at a map with broad implications. It’s one of the highest-quality questions in OkCupid’s database, meaning that our users have determined that it’s very important to them in finding the right match:

Rate Your Self-Confidence

581,443 people have answered

Generally speaking, the colder it is, the more likely you are to hate yourself. It’s interesting that every U.S. President since Kennedy has come from a green-tinted state, except for Gerald Ford (Michigan), who was never actually elected anyhow. I’d love to hear any of your theories about this map.

Chris grew up in New England and points out plenty of Mainers are in fact self-confident. But most of them move to New York or die snowmobiling.

I feel like the redness of economically depressed states like Michigan and Pennsylvania is self-explanatory. But why the extreme redness of Vermont? And why is a rich and otherwise successful place like Massachusetts skewing red?

OkCupid Rules

The following was the single most asked-for map, and we’ll publish it, though there aren’t many surprises:

Is homosexuality a sin?

346,925 people have answered

A state’s skew on this question very closely reflects how it voted in the 2008 election. With the exception of Arizona for obvious reasons, the relatively “No” states voted for Obama and the “Yes” ones voted for McCain. The maps for the many abortion questions in our database look very similar to this one, so I won’t post those, but we thought the below question map was probing enough to publish. Again, we see North Dakota’s peculiar take on life and death:

Is it logically inconsistent to support the death penalty but oppose abortion?

115,459 people have answered

Finally, I’ll leave you guys with this map.

Which would you rather lose?

283,859 people have answered

I put this up because the question was interesting and also implies a paradox. If the people who most love guns were offered this choice, the rest of us could pass real gun control. Voila.

As always, Chris and I are interested in your comments and ideas for other maps and comparisons. We’d like to do the U.K., Australia, Canada, etc., but Google’s chart API doesn’t map regions inside those countries. The whole country would be one color. If anyone can point us to a good solution, please comment below.

167 Responses to “Death, Freedom, and Cold Winters”

  1. logic says:

    “Is it logically inconsistent to support the death penalty but oppose abortion?”

    It’s the other way round really. Opposing executing criminals but supporting aborting innocent babies.

  2. factorytour says:

    re: self confidence
    maybe in states like Vermont there is a self-selecting sample problem. If online dating is for losers in Vermont, then only losers with low self-confidence would be on OkCupid. I’ve never been to Vemont, but I can only assume the dating scene there is brutal.

  3. Craig says:

    Wow, fascinating. Makes me proud to be a Minnesotan!

  4. Shaughn says:

    As a USA (and indeed the world), watcher myself, this stuff is fascinating and I have forwarded it to a number of friends all over.
    The ethical and social / political science issues raised are extremely discussion worthy. It reminds me of a less personal version of the classic “Book Of questions” that gets trotted out on social occasions when someone wants to stir things up. Trawling for answers with barbed ethically probing hooks is fraught with possibilities of relationships dissolving before your eyes as partners doscover new things about each other.

    this is important stuff


  5. Shaughn says:

    and funny too

  6. Sam says:

    On the “Which would you rather lose?” question it seems like both choices are lose-lose situation. Good question though, and I really enjoyed these blog posts. I’m glad you guys decided to do something interesting with the vast amounts of data you have collected. It’s pretty sweet.

  7. Sova says:

    “We’d like to do the U.K., Australia, Canada, etc., but Google’s chart API doesn’t map regions inside those countries. The whole country would be one color. If anyone can point us to a good solution, please comment below.”

    Well, a map of the EU-27 should be possible with Google (assuming you have the necessary data), and definitely very interesting.

  8. LithAmoniel says:

    I love these maps, and the random, weird ass explanaitons you guys give as to why they present themselves the way they do. It would be fantastic if you could somehow find a way to, after a person answers a question, see the map for that question.

  9. Dustin says:

    “If the people who most love guns were offered this choice, the rest of us could pass real gun control. Voila.”

    But as has been proven recently in Iran, the right to vote, without a well armed populace, is worthless. Your vote means nothing if it cannot be backed up with true political power (as defined by chairman Mao).

  10. R.G. says:

    I’m surprised Idaho loves guns more than my home state of TEXAS.

    Then again Texas isn’t like the Texas that you’ll see in Western movies or King Of The Hill anymore…

  11. capt_carl says:

    I chuckled a bit when I saw North Dakota the only red surrounded by mostly greens for is it logically inconsistent to support the death penalty but oppose abortion.

  12. Valeroth says:

    I find these state graphs fascinating. An interesting look into how we as a nation view things.

  13. Valeroth says:

    If you can’t split up other countries… I would like to see how the US compares to other countries. Perhaps a World map for some of these questions would be very interesting as well.

  14. cynical_nerd says:

    If you lose your right to bear arms, then your right to vote is lost as well.

  15. Kutaly says:

    I’m waiting for the world map

  16. John says:


    I would like to see some of these maps batch’d out by county rather than state . . .

    . . . And . . .

    I would like to see maps which plot out [by county] various pairs of answers and _their_ _correlations_ instead of straight answers on a single variable.

    Political leaning vs. ___ would provide [probably] obvious results, where you look as ___ as gun control, or vote vs. gun rights.

    By county, Generating the graphs of 2 variable pairings should be relatively easy, and picking the pairs of variables to turn into graphs could easily come from the internal statistics you no doubt already have.

    Could easily do additional graphs where you cut the population of answers from 300k to , say, 100k, by using a particular answer to a 3rd question….

    Running graphs by county might be more interesting if the images were for single states with large enough populations in your database, and with large differences in “upstate” and “downstate” make ups. Say….New York, Massachussets, Louisiana, California all come to mind as states with divided outlook.

  17. Sing_le says:

    Hmmm…I’m a New Yorker who supports gun control and abortion rights and opposes the death penalty,and refuses to answer the question “Is homosexuality a sin?” because answering yes implies I think a feeling amounts to an act,and answering no (since feelings CAN’T be acts) implies I think it’s OK to engage in same-sex sex (which I don’t!)

  18. carlyzx says:

    It’s so washington, or at least the western washington to be like “OH MY LIFE IS BETTER THAN YOURS” LOL.

  19. ed says:

    How much data do you have on where people in the UK live? If you’ve got postcodes (or at least, the first half of a postcode), then it shouldn’t be too hard to do maps for the UK, even if it might require hitting things with a stick a few times…

    Is there any chance you could make available some sets of data for the UK, Canada and Australia – like just the answers to one or two questions, along with a location, and then let us have a go at finding a solution to charting those places.

  20. Cavalary says:

    You could just make world maps, so each country being of one color wouldn’t be so bad? Alternatively, you could use that city mapping thing you used for the flag burning question, though I believe that’ll be very hard to read.
    Hm, I see no state giving anywhere close to the right answers from my point of view…

  21. Julie says:

    I think it is interesting that the chart from Rape Fantasies and Hygiene By State – The Hygiene graph, is VERY close to the graph from Sweet-ass American Trends for Self-Confidence level.
    It makes sense that a depressed or self-loathing person would care less about their looks and hygiene than someone who is self-confident and likes themselves would.

  22. Benjamin says:

    Be nice to the colorblind! This is almost cruel…

  23. Ara Pacis says:

    I think the self confidence chart needs some additional interpretation. It’s asked in a dating context. Those in the south can be more active year ’round and use online dating as an adjunct to other activities for dating. However, cold climes have lower self esteem in dating, as you mention. Perhaps this is because being stuck indoors in a region of small towns and low populations makes online dating the only decent method of meeting someone new.

    Of course it could also mean that happier people flee the frozen north for the sunbelt and leave their depressed cousins behind.

  24. Onik says:

    Woot! My question was used!

  25. UniqueMaterial says:

    You guys are the bomb. The data you have and what you’ve done with it impressed me to no end. Please – don’t sell your method or the data to the government unless they pay you big time. LOL. Keep up the good work.

  26. M says:

    Is there a way to post what the mean and the high and low numbers are? I think it would add to the graphs. A national average of 75% high self-confidence is very different from a national average of 25%.

  27. Mike Caprio says:

    Could you also add a map of which states have death penalites, like how you’ve done with the Confederacy map?

  28. baconman says:

    I think it could be an interesting experiment to chart the world by country, much like you do the USA by state. Or a side-by-side, because apparently every state here acts almost like an independent nation of their own, it seems!

  29. tglenn993nn says:

    >I put this up because the question was interesting
    > and also implies a paradox. If the people who most
    > love guns were offered this choice, the rest of us
    > could pass real gun control. Voila.

    Passing gun control like this would violate the premise of your question. If this was an actual choice, and the subsequent vote was for gun control (to the level of effectively losing the right to bear arms), then the actual choice was not what it was presented as. The actual choice was “would you like to lose your right to bear arms and your right to vote, or just your right to bear arms?”

    I wonder how much y’all reflect on that you are subjective viewers who appear to be very much within the norm for where you live (according to your own data).

    I’d be interest in seeing this data mapped by county, not state, and knowing the standard deviations for the data in each area. The first would be evidence (or not) towards my theory that population density has a higher correlation with some of the results. The second would be a graphical representation of how unified an area is in holding a particular belief.

  30. bludude says:

    I’d explain the self confidence question with the average yearly sunlight in those states. More northern states get less sunlight and more southern states get more sunlight.

  31. NS says:

    The paradox of the final map compared to the first map, is that Washington, DC residents don’t have the right to fully bear arms or have the right to vote. As a federal colony, residents of the District of Columbia pay taxes and fight in wars, but are not allowed to vote for their own Senators or Representatives. Instead they have a mayor, city council, and a non-voting delegate in the House of Representatives, but Congress dictates what laws they can pass and how they can spend local money. So when it comes to matters like war & peace, which involve the use of arms, DC residents don’t have the right to vote.

  32. jeremy says:

    “John” up above had good ideas. The data is fascinating, but some of the abrupt color changes indicate there’s something not being captured.

    He suggested showing the correlation between pairs of questions. That would be very interesting and compensate for some bias.

    For instance, try the rape fantasy question vs political affilitiation. I bet that conservative repression would lead to more fantasies.

    I also like the idea of plotting by country, which would help solve population skew. But it sounds like google maps doesn’t have that?

  33. jeremy says:

    You might also want to correlate it with census data.

    This isn’t measuring the general population, but rather the population of uses in a particular state. Okcupid might ‘catch on’ in certain demographics that only exist in certain places or situations.

    For example, if it became wildly popular at a North Dakota university, those users could outnumber the members from maine, even though maine is a larger state.

  34. Justin says:

    Is it any surprise the state with legalized gambling and prostitution is the most confident?

  35. Sample size problem? says:

    Is there a sample size problem with the smaller states? I could very easily be mistaken, but it looks to my eye like the big empty states — the Dakotas, Montana, Idaho, Wyoming — are more frequently at the extremes than most states. If there are far fewer respondents in those states, then they’d be more likely to have unusually lopsided results — no?

  36. mbobo1185 says:

    This is a really interesting set of maps. My question is…What do the scales mean?? It looks like some of the maps go from solid red to solid green. Others look mostly yellow and orange. Does that mean 1% of people in Massachusetts and 99% of people in Alabama think homosexuality is a sin. If a huge percentage of people in Idaho would rather give up the right to vote than the right to bear arms, (is that 5%? 50%? 95%? etc.). In fact, it would be interesting to have that kind of information for a lot of questions.

  37. mofotronica says:

    According to the wikipedia article on population by state (, states like Vermont and North Dakota have tiny populations. It’d be useful to see the proportional spread of OKC membership per state, normalized against state populations. Maybe the twelve North Dakotans on here are absolutely insane? And what’s the reason for not including Alaska and Hawaii?

    You guys are hilarious, ingenious, and of course, brilliant. If that isn’t enough toadying praise to get my resume in the queue for a developer position, I can go into more detail about why I like your blend of statistics and Sylvester Stallone.

  38. LeAnna says:

    Oh my gosh, this was HILAAAARIOUS. I was raised in ND and now live in MN, so I really got a kick out of this post. And it is very accurate. Many an editorial has been written in ND papers about how people are all about making sure every baby conceived touches ground alive, but after that, screw up one too many times and you’re toast.

    And yes, we humble Minnesotans are pretty live and let live. We may not think we’re the shit, but at least we don’t want to kill you.

  39. ombrainfumo says:

    Interesting how Massachusetts skewed so far green on that last map, especially since the first battles of the the revolutionary war were fought in Massachusetts in defense of the colonists right to bear arms. I guess times have changed; maybe it has something to do with that self confidence thing…

    Anyway, I’m glad you guys are starting to look at the data by place. Maybe you could crunch the numbers to give people a match percent with places, as well as people. It’d be helpful for those of us with wanderlust… and to keep those poor germaphobes away from the unbathed hippies Vermont and Oregon. Another idea, you could make the maps into a game, like the virgin game or the gaydar test. Show a map and have people guess which match question it is.

  40. Maize says:

    I wonder to what degree the north/south distribution of self-esteem is related to light levels — that it’s not so much “the colder your climate, the more you hate yourself,” but rather, “The less overall sunlight you get, the more susceptible you are to depression.”

  41. Maize says:

    “If you lose your right to bear arms, then your right to vote is lost as well.”

    This doesn’t stand up to even the most trivial real-world analysis. Many countries don’t guarantee the right to bear arms in the sense that the U.S. does but are perfectly functional democracies or republics that have had voting citizenry for ages: Canada, the U.K., etc.

    On the other hand, if the idea is that if the U.S. government with military backing decided to rescind the right to vote, you could use force to defend yourselves, do you really realistically think that the citizenry could and would overthrow the modern U.S. military? The gap between the weaponry available to the average citizen and the weaponry available to the government is too vast. You have the “right to bear arms,” in the same sense that someone facing a soldier in body armour and carrying an automatic firearm has the right to harsh language. It’s an illusion meant to keep you complacent.

  42. Eric says:

    I’m pretty liberal (and pro-choice), but it is NOT logically inconsistent to oppose abortion while supporting the death penalty.

    Implied in the “right-to-life” position is pro-INNOCENT-life. I don’t think the pro-life movement really seeks to protect all life, no matter how destructively it is used.

    I’ve always thought the pro-choice movement’s charges of hypocrisy on this front are intellectually dishonest. I also think this unnecessary mess could be avoided if we called the right-to-life movement what it really is: not “pro-life” but “anti-choice.” (They are no more “pro-life” than the pro-choice movement is “pro-abortion.”)

  43. ryanthepirate82 says:

    As a California resident who I believe said yes to the question of “are some human lives worth more than others?” I’d like to say that I never applied any racial connotation to my answer, I was thinking more of the value of the life of a serial rapist versus that of a nobel laureate. I think some of these questions are too ambiguous and open to several interpretations to be reading too much into what the data may imply about a population.

  44. Joe says:

    I agree with Benjamin. I love these charts, but they are very difficult to read for a red-green colorblind individual. Can you experiment with using a different color scale?

  45. Peter says:

    Its worth noting that the Right to Bear Arms was drafted with voting in mind, and less so for hunting/sporting. The founding fathers of America wanted to be sure the citizens could revolt against their government. This becomes particularly important in a country where voting could be considered useless.

    Pulling a lever doesn’t have the impact that a .50 caliber bullet does.

    Giving up your right to bear arms to retain your right to vote is like choosing to encourage posting on forums on the internet over publishing books. Yes I just made an analogy where books are guns. Deal with it.

  46. wing says:

    I’ll offer some theory into Vermont’s low self-confidence…

    From what I’ve noticed, not too many folks in VT use OKC. Hell, not too many folks in VT use the Internet. My guess is that most of the confident folks are too busy pulling tractors or whatever on their organic farms while us less-confident folk are hanging out on the webs.

  47. johnny says:

    lol, what’s wrong with minnesota? i live in minnesota. it’s a good state :p

  48. Sleeplongtime says:

    Hahaha! I agree with LeAnna, only I’ll add in Wisconsin to what she said. Just don’t mess with our cows, or our cheese, and we’re a friendly bunch.

    Again, a great job on the maps. But there have have been some intersting questions raised about the population sizes and the way it relates to the where states fall in the codes. A lot of these color coded questions results could change depending on the number of users living in urban and rural areas. Are the cities driving the states averages towards one extreme or the other?

  49. DrShaffopolis says:

    I generally skip questions where the possible answers are “very”, “somewhat”, “not very”, etc.

    The reason is that I think these say as much about a person’s perception of other people as they say about the person.

  50. DrShaffopolis says:

    Darn, hit submit by accident –

    The way the above is relevant is that I’m wondering whether the “very high” to “very low” self-confidence is really a reflection of people’s self-confidence, or their perception of what a normal level of confidence is.

    It’d be interesting to plot some match questions against things with actual statistics available. I bet that if you plotted “are you overweight” or “do you drink a lot” against actual weight and alcohol consumption maps, there would be some interesting differences in the actual vs. the perceived maps. Or, maybe not, but it’d be interesting to know my guess is wrong.