Is Overcoming Bias Important?

Arnold Kling says he "does not buy this argument" of mine:

If you have a cause, then other people probably disagree with you …  When other people disagree with you, they are usually more right than you think they are. … Before you go and attach yourself to this cause, shouldn’t you try to reduce the chances that you are wrong? Ergo, shouldn’t you work on trying to overcome bias?

Tyler Cowen riffs:

[Such] views are tautologically true and they simply boil down to saying that any complaint can be expressed as a concern about error of some kind or another.  … draw an analogy with statistics.  Biased estimators are one problem but not the only problem.  There is also insufficient data, lazy researchers, inefficient estimators, and so on.  Then I don’t see why we should be justified in holding a strong preference for overcoming bias, relative to other ends.

We have never claimed bias is the only problem.  But to let Arnold and Tyler more clearly identify where they disagree, let me outline an argument for the importance of overcoming bias:

  1. Our beliefs have many errors, i.e., deviations from truth.
  2. Reducing error is important goal, for which we are willing to pay substantial costs.
  3. The causes of our errors can be seen as ranging from context specific to general trends.
  4. We in fact have many identifiable stable general error trends, in addition to legion context specific causes.
  5. By reflecting on error causes, we can seek ways to adjust our pattens of thought and social institutions to reduce error.
  6. For a substantial fraction of error causes, we can in fact find feasible adjustments. 
  7. It is often more cost-effective to seek and implement adjustments for general trends, than for context specific errors.

Together these points suggest we should be willing to pay substantial costs to reflect on and seek adjustments for general error trends.  Our being here suggests we draw a similar conclusion.  It seems a confident proposer bias to suggest we claim more.

Added: Bryan Caplan weighs in.

GD Star Rating
loading...
Tagged as:
Trackback URL:
  • Stuart Armstrong

    Our beliefs have many errors, i.e., deviations from truth.
    I’d add (personally, though I suspect that this feeling is shared by others on this blog) that these errors are a lot more prevalent and important than I first thought. This may be just a déformation professionnelle, but it suggests that Arnold Kling and Tyler Cowen should explore the topic themselves, at least superficially, before dismissing it. There is a lot of knowledge to be picked up from this blog that can’t easily be described.

  • Constant

    Stuart – in his defense, Tyler does describe this blog as a must-read. So where you way, “Tyler Cowen should explore the topic [himself], at least superficially, before dismissing it”, I don’t think he can correctly be said to be dismissing it if he thinks this blog is a “must read”. He’s raising some sort of objection to some claim about it, but I think calling that “dismissal” can’t be correct.

  • Stuart Armstrong

    Stuart – in his defense, Tyler does describe this blog as a must-read. So where you way, “Tyler Cowen should explore the topic [himself], at least superficially, before dismissing it”, I don’t think he can correctly be said to be dismissing it if he thinks this blog is a “must read”. He’s raising some sort of objection to some claim about it, but I think calling that “dismissal” can’t be correct.

    I stand corrected. And Tyler’s point now seems a lot more convincing now that I know he has a broaded overview. Though his comment “say by considering a wide variety of views and methods” does give me pause. The truth is not the midpoint between the two talking faces.

  • Tobbic

    It seems to me that the best way to reduce error is to channel efforts to that eliminate that cause of error in which expected marginal gains are the biggest. So, there are two factors: 1) ease of error reduction and 2) expected gains. For OB, I think the it’s fairly easy in the margin, at least for those who are initially unfamiliar with this subject. More, I think as, Robin pointed out, expected gains towards truth can be very large because biases are general causes of error.

    I suspect many people rationally choose to be lazy. “Overcoming laziness” is an old thing (sloth was one of the 7 deadly sins), I think people have adjusted on this one. Thus, the expected marginal gains are IMO low.

  • Jeff Borack

    I think a better question would be “Is overcoming bias ethical”.

    I believe that behavior is just as much a product of evolution as our ears and eyes. From variations in temper to belief in god, these personality traits help us, as a species, survive. I find it hard to imagine that these seemingly random and illogical biases are actually random. It seems more likely that biases serve some sort of purpose related to teamwork or group cohesiveness.

    If this is the case, is it moral to deliberately try and overcome these adaptations that help us work as a team? Is there any reason to believe that overcoming bias actually would help us as individuals economically, socially, and what would the cost be to ourselves and society?

  • Jeff Borack

    I think a better question would be “Is overcoming bias ethical”.

    I believe that behavior is just as much a product of evolution as our ears and eyes. From variations in temper to belief in god, these personality traits help us, as a species, survive. I find it hard to imagine that these seemingly random and illogical biases are actually random. It seems more likely that biases serve some sort of purpose related to teamwork or group cohesiveness.

    If this is the case, is it moral to deliberately try and overcome these adaptations that help us work as a team? Is there any reason to believe that overcoming bias actually would help us as individuals economically, socially, and what would the cost be to ourselves and society?

  • Carl Shulman

    Jeff,

    What is Overcoming Bias good for individually? Many things:
    http://www.overcomingbias.com/2007/04/overcoming_bias.html

    How good would overcoming bias be for society?

    Parachuters striving to raise money for the NHS instead cost it a 13.75 pounds for every one raised:
    http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6T78-3Y8WK2S-1M&_coverDate=05/31/1999&_alid=392341591&_rdoc=1&_fmt=&_orig=search&_qd=1&_cdi=5052&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=5a0b3e917ed37a5b5971fd5fa0e21315

    Compare this to salt iodization, oral rehydration therapy, iron supplementation, and other relatively cost-effective interventions from the field of ‘normal’ charities, that populate the top $50 billion in projects selected by the Copenhagen Consensus:
    http://en.wikipedia.org/wiki/Copenhagen_Consensus

    Americans gave $295 billion in charitable contributions in 2006. If reducing bias could reallocate 5% of those donations (mostly by affecting relatively rational marginal donors), the Copenhagen Consensus could be funded in 4 years. Because of the huge variance in charitable efficacy, even if total giving fell by 90% or more, this could be offset by an increase in the most efficient giving.
    http://64.233.169.104/search?q=cache:kLrsbbfcZzoJ:www.terradaily.com/reports/US_Charitable_Giving_Sets_New_Record_Topping_Katrina_Effort_999.html+total+charitable+giving+united+states+2006&hl=en&ct=clnk&cd=1&gl=us&client=firefox-a

    As an individual, you could plausibly increase the amount of good you do for others by a factor of hundreds (or shift from causing net harm to benefit), and perhaps by billions (www.nickbostrom.com/astronomical/waste.html).

  • Constant

    Maybe someone will set up a Copenhagen Consensus-based charity.

  • http://cumulativemodel.blogspot.com aaron

    Sounds like a quest for better biases rather than eliminating biases (which sounds like a much more practical and reasonable goal).

  • http://cumulativemodel.blogspot.com aaron

    I was thinking about a charitable futures market. Contracts that payout to charities and organizations for reaching goals and showing their actions contribute to results. Contracts that payout for certain events (types of dysasters/locations).

    Could be used to promote development. Also could be hedged in other futures markets (put up X relief dollars to A organization if category 5 storm hits Z region, and buy a futures contract that pays out if the storm his).

    Companies could put up dollars to encourage conditions that would be favorable for their industry (disease control, economic developement)

    And a charitable betting clearinghouse ($10,000 to my charity if X team wins / X politician wins)

  • Jeff Borack

    Carl,

    Thank you for your response

    In your first link: ( http://www.overcomingbias.com/2007/04/overcoming_bias.html ) poor Mr. Armstrong felt that his attempts at overcoming bias have made him “better at academia, and worse at business”. Could this observation possibly be a result of Mr. Armstrong’s decreased ability to function as part of a team or his coworker’s perception of his ability to function as part of a team. He certainly indicated that it might effect his ability to rally the support of others to join his team. While I realize this is just one small example, and I’m currently plagued with the bias to see only supporting evidence, I believe it supports my hypothesis that overcoming bias makes individuals less adept at functioning in a group setting. What kind of effect might this have on society?

    I couldn’t access your second link.

    I don’t know much about the Copenhagen Consensus, and I don’t speak Danish, so I have some questions for you.
    1) It seems like many of the staff members are not part of an independent research council, but rather ‘project managers’. What conflicts of interest are present between the committee that rates projects and the people who manage various projects?
    2) How are members of the consensus committee selected?
    3) How are funds transfered to projects?

    Call me biased, but I would have more faith in the Copenhagen Consensus if they simply rated charitable organizations or projects, published a report, and didn’t deal with funding at all. In any case, it seems like a step in the right direction… but I’m not sure if we can make that evaluation with any amount of rational thinking.

    Millions of years of evolution made humans with two arms. I feel like 4 arms might be more useful. I could throw more spears at saber tooth tigers, and my wife could pick twice as many wild berries. But evolution gave me and everyone I know only 2 arms. All the logic in the world indicates that I do not need nipples, but yet I have two. Could I build a better man? It would be fun to try, but I’d probably mess it up.

    On the other hand, maybe overcoming bias is equivalent to exercise. Maybe it’s less like trying to fly and more like trying to run faster.

  • Carl Shulman

    Jeff,

    1. I agree that overcoming bias can offend some people who hold strong beliefs related to that area if one has imperfect deceptive ability. This has to be traded off against various benefits.
    2. The Copenhagen Consensus was a cost-benefit prioritization exercise to evaluate different solutions to global problems undertaken by a set of mostly prominent academic economists, initially in 2004.
    3. It only made recommendations and published its cost-benefit analyses, it doesn’t deal with funding. The exercise was to attempt to figure out how development agencies *could* effectively spend $50 billion.
    4. Its work is not in Danish, but was conducted and published in English:
    http://www.amazon.com/Global-Crises-Solutions-Bjorn-Lomborg/dp/0521606144
    5. I mentioned it because it collects voluminous evidence from a variety of literatures and directly compares the costs and benefits of different approaches, concluding that there are variations of several orders of magnitude in the cost-effectiveness of various causes. When such variance, or substantial uncertainty about it, exists the gains from reallocating effort can also be very great, and are likely to exceed the costs of analysis.