16 ms·
Small payments sharply diminish gap in responses to partisan factual questions
- sharemywin 10y agowonder if there's a way to tease out what I believe versus what I think you want to hear.
- madenine 10y agoWhich is scary, given that the survey was about factual information. Welcome to the post-truth world I guess. "I think want me to say the unemployment rate is X, so I'll select that, but I believe the unemployment rate is actually Y"
- tomjen3 10y agoWhen you write it that way it sounds crazy, but if you believe the unemployment rate should be the U6, then we will have real disagreements over that fact. When the media is just the media, no longer the mainstream media, but just the media with just another (biased) agenda and world-view that is no better than any others, they end up acting like spoiled children. I hope HNners are smart enough to realize that all medias are biased in some ways (some more than others) and that fake news/post truth are just another political term (unless you are talking about the spam that floated on Facebook with completely fake headlines, not justified by the content).
- Yen 10y agoWhether the unemployment rate discussed and optimized for should be U-6, or any other particular measure, is a disagreement about opinion, not about fact. Claims that the U-1 is falsified/manipulated for political purposes, or that the "real unemployment rate" is actually much higher than even the U-6, are much more clearly false.
- sharemywin 10y agoIf your paying me to give you the answer than I'm going to go with what I think you want me to say. If fox news approached me and offered to pay for whether I thought climate change was a hoax it would be a different answer than if the question was from Nasa.
- bduerst 10y agoThat's why the survey is blinded - the respondents have no idea who is giving it and are forced to assume the responses are supposed to be whatever is true. It's like a random organization asking you if Climate Change is a hoax, for which you know nothing about their bias.
- sharemywin 10y agoHere's another problem. apparently there are 6 different numbers... http://www.cnbc.com/2016/10/07/unemployment-rate-shows-at-5-but-more-realistic-rate-is-higher.html http://www.cnbc.com/2016/10/07/unemployment-rate-shows-at-5-... So, when they ask which one are they talking about?
- bduerst 10y agoDoesn't matter. It's a red herring because the question itself was a rate of change over time, for which the answer is given in a 1.0% range. Irregardless of whether you focus on U-5 (the reported statistic) or U-6 (the denier's statistic), it's still only about a -0.2% and +0.1% change, respectively. Even if you did the math on U-5, you would still get the answer right for every of the 6 numbers.
- sharemywin 10y agothe denier's statistic is closer to use U-5 for a starting point then switch to U-6 as an end point.
- bduerst 10y agoWherein lay the bias that they're trying to capture. Of course, for this question even that nuance isn't caught.
- sharemywin 10y agosorry missed the rate of change over time.
- Semiapies 10y ago"Welcome to" implies it hasn't been that way for longer than any of us here have been alive.
- hx87 10y agoUse a variety of research groups with different political leanings & reputations?
- ianai 10y ago"The experiments show that small payments for correct and "don't know" responses sharply diminish the gap between Democrats and Republicans in responses to "partisan" factual question"
- narrator 10y agoThis study fails to impress me as anyone with some intelligence would always pick "don't know" since it's a guaranteed payment.
- alexbock 10y agoThey performed this with and without that payment. "This paper reports results from two novel experiments designed to distinguish sincere from expressive partisan differences in responses to factual questions. In both experiments, all subjects were asked factual questions, but some were given financial incentives to answer correctly. ... In our second experiment, we therefore implement a treatment in which subjects were offered incentives both for correct responses and for admitting that they did not know the correct response."
- throwaway729 10y agoBut if there's no payout then there's no incentive to say don't know. Whereas if there is a payout then the optimal strategy is clearly "dont know", assuming equal payouts for correct and dont know. Why would you take the risk of being wrong?
- hx87 10y agoQuote from paper: > The amount paid for “don’t know” responses was also assigned randomly, and was a fraction of the amount offered for a correct response: 20% of the payment for a correct response with probability .33, 25% with probability .33, and 33% with probability .33.
- 10y ago
- Fr0ntBack 10y agoInteresting finding. One of the problems with voting is that there is no incentive or a very small incentive to choose the candidate with the best policies, because the chance of your vote making a difference is so small. Therefore people vote for a candidate that makes them feel good, rather than a competent one. This is arguably the cause of many bad policies. Caplan discusses this issue in his book 'The Myth of the Rational Voter' https://en.wikipedia.org/wiki/The_Myth_of_the_Rational_Voter https://en.wikipedia.org/wiki/The_Myth_of_the_Rational_Voter
- philh 10y agoCounterpoint: http://www.stat.columbia.edu/~gelman/research/published/rational_final7.pdf http://www.stat.columbia.edu/~gelman/research/published/rati... This paper argues that voting for social reasons ("this result will benefit society") is sensible even in large elections. The value depends on the percent turnout, but not on population size.
- throwaway729 10y agoVoting theory models rational actors, not actual people. Interestingly, fontback's hypothesis is exactly how they modeled the choice, with a financial incentive plus some "psychological incentive"
- philh 10y ago"there is no incentive or a very small incentive to choose the candidate with the best policies, because the chance of your vote making a difference is so small" is a theoretical argument about rational actors, not an observation about actual people. The paper I linked argues that it doesn't hold. It also argues that there's some real-world evidence that people vote for social reasons.
- scarmig 10y agoRegardless of how a person votes, they're voting for a feel-good reason. Some people's identity gives them pleasure in voting for the candidate that rationally maximizes the Good, for some subjective measure of the good. These are the 'social preferences' Gellman talks about. Some people's identity gives them pleasure in the tribe, and they want to vote for the people like them and around them to fit in, especially so they won't be confused with the Enemy. Pretty much everyone has both tendencies, which are inherently in tension. This is cognitive dissonance. People don't like experiencing this, so they end up choosing to believe both in tribal politics and that their tribe is the rational one, which is the most convenient way to cut that Gordian knot. Now, we can debate which tribe is more rational. I think currently in the USA the blue tribe has more people motivated by the desire to be seen as rational. Then again, we'd also predict that's what everyone thinks. So probably that debate wouldn't be productive. This experiment is useful because it tells us that money is a very useful way to cut through the bullshit of tribal beliefs. It also might give us hints as to how to design our electoral system to tend toward more rational options.
- qwrusz 10y agoThe small payments ranged between 17 cents and $1. But how the payment calculations were described, based on weirdly complicated raffle entry odds for gift certificates or with different questions randomly worth 25 cents or 50 cents or whatever with various probabilities. Frankly, I doubt most participants understood the payments guidance about what amount correct answers would earn them. This could imply just putting a paragraph of confusing words with dollar signs mixed in before asking questions might also diminish political partisan bias in responses.
- gist 10y agoCould also be a version of the Hawthorne effect. [1] In this particular case the money was a proxy which was more powerful than words meant to say 'your answer is important' or 'your vote counts'. It's actually hard for me to put in words this concept but intuitively it is what seems is happening. Ok so let's take as an example upvote/downvotes on HN. Many of us give little thought to the impact and do it impulsively at least in some cases. Now let's pretend we know there is an impact, even minor, financially. We will pay more attention and I feel at least be more likely to carefully consider voting in either direction. Remember with money also people tend to not want to be wasteful it doesn't feel good. Nobody takes even pennies and throws them in the trashcan. [1] https://en.wikipedia.org/wiki/Hawthorne_effect https://en.wikipedia.org/wiki/Hawthorne_effect
- Swizec 10y ago> Nobody takes even pennies and throws them in the trashcan. And yet they should. https://what-if.xkcd.com/22/ https://what-if.xkcd.com/22/ If you spend more than 3.6 seconds actively thinking about a penny, you value your time at less than $10/h. This comment is unfortunately self-referrential. Oops
- gist 10y agoWell not to dispute some of your points but let me introduce another concept here. That is the act of 'forced savings'. And example of this is when you build equity in something that doesn't appear to be a good idea in the long run financially when compared to another alternative. Ok here is an example. Let's say that your home that you are paying a mortgage on offers a 5% increase in value per year (totally made that up) vs. putting the money in a money market that pays 6% per year (definitely totally made that up). On the surface the money market is a much better return let's stipulate and observe. After all you have liquidity and I will even further assume that the 6% is guaranteed. But wait a second. If the money is tied up in a house you can't spend it. In the money market you can. So the house is forced savings and in the end you might very well end up with more, all else equal. I say this as someone who just took cash that I had and bought a sort of expensive car by the way. Realizing that if I had the money in more real estate (which I also own) I most likely wouldn't have bought the car (since I don't like to finance or lease even though in theory it can make more sense financially).
- droithomme 10y agoIf someone asks me my opinion about something I'm much more likely to say what I know they want to hear if they offer to pay me to say that, and I'm less likely to debate with them instead. If they aren't paying me, I'm more likely to say what I really think.
- ForHackernews 10y agoAlternatively: If I'm getting paid for factually-correct answers, I'm less likely to throw away money by using my answers as an opportunity to signal my ideology.
- jstanley 10y agoIf your ideology is at odds with factual correctness... maybe you should rethink your ideology?
- ForHackernews 10y agoYou can say that, but [almost] everyone has ideological beliefs that conflict with the best available evidence. Maybe you rationalize it to yourself that "well, the jury's still out...those studies aren't conclusive" or "yes, but in principle my belief is correct, it's just that X, Y and Z are confounding variables so..."
- twoquestions 10y agoHardly! Professing belief in something blatantly untrue is a wonderful social signal, it says you're willing to stick with the in-group no matter what! Example: Someone close to me thinks the Detroit Lions are the best football team.
- bduerst 10y agoWhich seems like harmless, lightheared tribalism, sure, but there are many forms of tribalism that have quantifiable negative impact. Even for Midwest football: http://qje.oxfordjournals.org/content/early/2011/03/21/qje.qjr001.full http://qje.oxfordjournals.org/content/early/2011/03/21/qje.q...
- praptak 10y agoHaving skin in the game makes one thoughtful. Sometimes it works in debates too. "Let's do A, it's much better design." "Will you do the debugging and support for the feature?" "Ummm, maybe let's rethink this A, I see some flaws in it."
- idiot74 10y agoCould be. With no financial reward there's not much incentive to spend energy thinking about things: so answers default to supporting political views. Suddenly there's a financial incentive, and so people take a few more steps in their thought processes. (maybe, idk)
- sambe 10y agoEasy to say yes and then shirk it in future though. Unfortunately common in practice.
- BurningFrog 10y agoSo I guess this moves the mindset from "argument" towards "truth finding". Assuming that's true, how could we use that effect for, say, good?
- AlexCoventry 10y agoReducing the tendency of people to lie on survey polls is an obvious applications. You could ask people to predict how popular a candidate is, with a payoff tied to prediction accuracy.
- emmelaich 10y agoRelated: https://en.wikipedia.org/wiki/Narcissism_of_small_differences https://en.wikipedia.org/wiki/Narcissism_of_small_difference... and Ecclesiastes 1:2 - All is vanity
- dbroockman 10y agoGood thing voters are incentivized to vote correctly like they were incentivized to answer correctly in this survey. Oh, wait.
- Fr0ntBack 10y agoFor all of you reading whose political knowledge is a little rusty, here is a quiz for you: https://www.buzzfeed.com/bennyjohnson/how-well-do-you-know-basic-us-politics?utm_term=.luy4rbRzY#.nlB7odpBM https://www.buzzfeed.com/bennyjohnson/how-well-do-you-know-b... Unfortunately there is no reward for correct answers!
- jsight 10y agoThat doesn't strike me as being a very useful quiz. Would doing well on that really imply a stronger knowledge of policy?
- yellowstuff 10y agoThis is more trivia than the kind of knowledge that implies understanding. Also needs an update: > 14. How many Supreme Court justices are there? :(
- ZeroGravitas 10y agoA related finding was that the more scientifically educated an American conservative is, the more likely they are to deny various elements of climate change. The conclusion of the researchers was that the smarter respondents knew what they were supposed to say better as well as knowing the real answer better.
- dnautics 10y ago"For example, Republicans are more likely than Democrats to say that the deficit rose during the Clinton administration; Democrats are more likely to say that inflation rose under Reagan." These are both really tricky questions, because the questions described are derivatives of derivatives! The deficit is the time derivative of the debt; so it's possible that someone is hearing "did the debt rise during clinton" (which would be true). Likewise, inflation is the derivative of the value of money. So someone might be hearing "did things get more expensive over reagan" (which would also be true). The deficit/debt one is really tricky since they both start with the letter D and even though I know the difference occasionally in conversation I switch the two.
- blowski 10y agoThe question itself is not important. The basis of the research is that one's political affiliations affect how you answer the question.
- convolvatron 10y agoI think it is* somewhat important as some of these political questions are very complicated and potentially very subjective. I think it makes a clearer case if they involved resistance to arguably objective facts.
- chc 10y agoThe point is that people tend to answer the question in a way that favors their party and disfavors the other party if you just ask them, but they're more likely to get it right if you offer an incentive to do so. This can't be explained just by "The questions are tricky," though that might be contributing factor that helps people fool themselves when they want to.
- smallnamespace 10y agoI'd offer that one issue with representative democracy is that the direct external payoff for voting well (by 'well', I mean taking the time to do honest research and keeping an open mind) is pretty much zero. That leads to a contradiction -- we expect our elected politicians to do whatever the public wants, but we also want our politicians to be wise, forward thinking, selfless, and moral leaders. Those two desires can only be fully reconciled if the public itself has those qualities, but the public doesn't have much direct, quick feedback as to whether it's making good choices. We also don't want to pay our politicians very much (politicians make orders of magnitude less than, say, business leaders) and still expect them to be fully independent, unbribeable, and only beholden to the public. You get what you pay for.
- xname2 10y agoOK. So Trump is the coming president, and GW is real and Trump will not handle it properly. So it is very likely GW bad predictions are going to happen. My question is: why there is no massive sell of beach houses? a) Most beach house owners are GW deniers. b) Most beach house owners have strong confidence of the future government. c) Most beach house owners are extremely rich so they don't care. How about those who are GW believers, and seriously think it is critical timing for action rightnow, and have uncertainty about future government, and are not extremely rich. Are they selling houses and moving to inland? I think it is a pretty good test of how many people seriously believe GW. Do not tell me the incentive of home value is not strong enough.
- philipkglass 10y agoIndividuals are pretty poor at making personal decisions on the basis of risks that are well-proven numerically but not tangible from personal experience. Why do millions of Americans still smoke cigarettes? Do most smokers have awesome health plans and optimism about medical research so they assume that cancer treatment will be pleasant and affordable in the future? I think it's a pretty good test of how many people seriously believe cigarettes are carcinogenic. Do not tell me the incentive of a miserable premature death is not strong enough.
- xname2 10y agoThere are people keep smoking, but there are also people quit smoking because of knowing it's unhealthy. My question is why don't we see massive sell of beach houses. My question is NOT why there are still people not selling their beach houses.
- jeffdavis 10y agoIt could also be that people respond based on what they think the "official" answer is when paid; otherwise they answer what they actually believe. For instance, there are official measurements of debt, deficit, and inflation. But some might believe that unfunded liabilities are debt, or that inflation is higher than the official number.
- dhbradshaw 10y agoIt's worth noting that, unless you're very careful, tying incentives to answers doesn't filter for "what are the facts" as much as "what do you think I think are the facts." These can be similar but are not the same thing.
- insertion 10y agoI've wondered why partisanship doesn't exist to nearly the same extent in business and consumer behavior. This explains a lot.
- padseeker 10y agoSo what you are saying is that partisan hacks are aware they are full of crap, and will openly profess the those very things that they know are bullshit UNLESS there is a direct financial incentive for them to give the factual answer? That is depressing.
- technotarek 10y agoLet's start taxing candidates for mis-truths spoken in presidential debates!
- arikrak 10y agoIt would be interesting if people had to put their money where their mouth was more often. E.g. if a politician makes a claim about the past or the future, there should be a way to hold them accountable to it.
- k_lander 10y agoThis is the basis for prediction markets like Gnosis and Augur. Would be interesting to see how they play out in practice.
- brownbat 10y agoSome issues dismissed as partisan anti-factualism actually mask layers of nuance.* Political discussion is almost entirely about stressing different layers of nuance and dismissing or eliding the nuance of your opponents as irrelevant. So if you were to ask someone in a survey for their opinion, they might respond according to their preferred level of nuance on the politically charged topic, throwing a monkey wrench if the underlying presumption of the question is overly simplistic. They might provide a contrarian response because they see the question as loaded, as in "Did you stop beating your wife?" An incentive though turns the game into a Keynesian beauty contest, and makes people fall back on the perceived popular answer, even if they don't think it's right. * All the examples of this are incredibly fraught and risk sidetracking the entire discussion. Kindly only read these if you can charitably look for the meta-rhetorical point. But imagine: Did Trump support the war in Iraq? Depends on whether a half-hearted 'yes' on Howard Stern really counts. Did Bill Clinton lower the deficit? Depends on how much you believe the Republican House forced his hand and was the real actor. Did Reagan end the cold war? Depends on how much you want to credit internal political factors or the role of Gorbachev. Were WMDs found in Iraq? Depends on whether you count sarin-tipped warheads found in 2005, seed material found at Tuwaitha, or if you're referring exclusively to the Hans Blix UN investigations or assembled nuclear munitions. The study designers tried to use careful phrasing to avoid some of these problems. Even so, there are some questionable ones. "Did inflation rise under Reagan?" Of course it did! At least once every year! Just not overall...