Overcoming Bias

Sleepy Fools

Regarding if we should avoid disagreeing, Eliezer once wrote:

The central argument for Modesty proposes something like a Rawlsian veil of ignorance - how can you know which of you is the honest truthseeker, and which the stubborn self-deceiver? ... the obvious reply:  "But I know perfectly well who the fool is.  It's the other guy.  It doesn't matter that he says the same thing - he's still the fool."  This reply sounds bald and unconvincing when you consider it abstractly.  But if you actually face a creationist, then it certainly feels like the correct answer. ...

Those who dream do not know they dream; but when you wake you know you are awake.  Dreaming, you may think you are awake.  You may even be convinced of it.  But right now, when you really are awake, there isn't any doubt in your mind - nor should there be. ... [you have] just an (ahem) incommunicable insight that you were awake.  ... "That we can postulate a mind of sufficiently low (dreaming) or distorted (insane) consciousness as to genuinely not know whether it's Russell or Napoleon doesn't mean ... [I'm] Napoleon. 

Are sleepers who think they are awake a good analogy for justified disagreement with fools?  Do sleepers and fools both have a broken mental state blocking them from assimilating key relevant info?  Interestingly, the merely sleepy do seem to know they are sleepy: 

Sixty-four adults participated in a study examining the accuracy of metacognitive judgments.  During 28 hr of sleep deprivation (SD) and continuous cognitive work. ... Subjective and objective measures of sleepiness confirmed the expected patterns of increasing fatigue with SD. ... Traditional indices of the confidence-accuracy relation (i.e., calibration, resolution, over- and underconfidence), as well as the accuracy of pre- and posttask estimates of performance, remained stable over the SD period. The findings suggest that people can accurately assess their own cognitive performance when deprived of 1 night of sleep.

Apparently it is also possible to know that you are dreaming: 

A Model Disagreement

We have often pondered the question: when you find that you and someone else disagree, how much weight should you give to your and their opinions in forming your new opinion?  To explore this, I've worked out a simple math model of disagreement between two "Bayesian wannabes", i.e., agents who are trying to act like Bayesians, but know that they make mistakes, and try to adjust for this fact. 

Consider two agents, A and B, having a conversation about a truth t = x₁ + x₂ + x₃ + ...  First A sees clue x₁, and reports r₁, his estimate of truth t.   Next B sees report r₁, and also clue x₂, and then reports r₂, his estimate of truth t.  A now sees report r₂, a new clue x₃ and reports r₃.  The two of them could go back and forth like this for a long time. 

If A and B were perfect Bayesians (and if each x_i were independently and normally distributed with zero mean and a known variance V_i), then we would have r_i = x_i + r_i-1.  When combining their last two expressed opinions, each agent puts zero weight on his own last report, and just adds his new clue to the other agent's last report!

OK, but what about imperfect agents?  I assume:

Kids, Parents Disagree on Spouses

Monday's Post:

Do young people and their parents really disagree about the qualities of a suitable mate? ... A study involving Dutch, American and Kurdish students ... found that the cliche is, in fact, true. Young Americans told the researchers that qualities they would find unappealing in a potential mate included low intelligence and physical unattractiveness. But they said their parents would object to a mate who was of a different ethnicity, was poor or lacked a good family background.

The responses of Dutch and Kurdish students were similar in that young people invariably considered the potential mate's attractiveness the most important quality, whereas parents uniformly paid more attention to the suitors' social background or group affiliation -- race, religious background and social class.

[The authors] said the consistency of the conflict across cultures suggests the hand of evolution: Parents and offspring ... genetic self-interests, while overlapping, are not identical. The reason young people care so much about intellectual and physical attractiveness, the scientists suggested, is that these characteristics are markers of genetic fitness. By contrast, they said, parents care about group affiliations because parents are primarily interested in whether an incoming member of the family is likely to make a good parent -- and a good all-around team player.

There should indeed be some conflict between kids and parents on suitable spouses, but the size of the conflict seems surprisingly large - do parent and kid genetic interests really diverge that much?   Here's a graphic showing huge differences:

Inhuman Rationality?

We seem comfortable celebrating those who practice complex statistical analysis, even if only one in a thousand can do so.  And the one in a billion genius celebrated for greatly improving our understanding or practice of statistics, or other rationality, seems to us an epitome of the best in humanity.  Who would call such exemplars "inhuman"?

But when rationality seems to require not rare celebrated abilities, but common despised tendencies, suddenly many complain rationality is "inhuman."  For example, many balk at my suggestion that rationality requires us to be more conformist in our beliefs, deferring more to others' opinions, even if in conversation we explore different evidence and analyzes.  Yet conformity is widespread and most people have far more diverse word-opinions than betting-opinions.  I'd guess far more than one in a thousand humans realize something close to the degree of conformity I'm suggesting. 

It seems to me that the real issue here is not humanity but social status.  We are eager to achieve aspects of rationality that get high social status, but not those that get low status.  To gain high status, we don't mind moving far out into the tails of the distribution of human features, but if low status is a prospect we suddenly express grave concern about moving out into the "inhuman" tails.

For example, when I said: 

Distinguish Info, Analysis, Belief, Action

I have argued often that rational (i.e., most truthful) beliefs are less diverse than those we commonly observe now.  Many (such as Hal Finney in recent comments) have responded by saying diversity is valuable, and so society might lose if beliefs became less diverse.  It seems to me, however, that while there is value in the diversity of info, analysis, and action, there is not much additional value in diversity of belief.

Our info is the clues we find about how things are, including what we see, what we feel, and what we observe about others.  Our analyzes are our thoughts, both conscious and unconscious, in looking for patterns and dependencies between clues and hypotheses, which suggest which clues might support which hypotheses.  Our beliefs are the conclusions we form from our analyzes about the world (including ourselves).  Finally, we take concrete actions based on our beliefs and our values.

If we are going to share our clues, we are of course better off if we each collect different clues than if we all re-collect the same clues.  Similarly, (as Scott Page emphasizes) we are better off if we share and compare the results of different styles of thinking, instead of each of us redoing the same kind of analysis.  And we must often take different actions if we are to actually find different clues and use different types of analysis.

We do not, however, require different beliefs to achieve any of these other differences.  For example, searchers in an Easter egg hunt do not need different beliefs about egg locations to spread out across the lawn - there is a strategic advantage to searching where others are not searching.  And the best strategy in many social situations is to randomly choose among many possibilities, so that people with the same beliefs end up taking different actions.

We do not need different beliefs to hold and share different info and analysis.  We can talk directly about the clues we have seen, and about the conclusions our analyzes seem to favor.  We can get into the habit of distinguishing "what I saw was ..." and "it seems to me that ..." from "I believe that ..."  Yes, to share info we must each describe what we saw and what our analyzes suggest, but we need not be so stupid as to form our beliefs only on such things.  When choosing beliefs, we are wise to weigh heavily the info, analysis, and beliefs of others, even if when talking we are wise to offer distinctive and diverse insights. 

I do not dispute that there are in fact some specific situations where a diversity of belief will lead people to sacrifice their own ends for the benefit of society.  What I question is how representative are such situations.  There are many other situations where belief diversity hurts society, such as when innovations are rejected because they were "not invented here." 

Reject Random Beliefs

A recent New Scientist mentions a 2005 American Political Science Review paper on the genetic basis of political beliefs, which includes this key table, breaking variation in opinions (among 30,000 Virginia twins) on 28 specific topics into three origin components: genetic (heritability), family (shared environment), and other (unshared environment):

Ignoring Advice

When do people listen to advice?  I teach my health econ students about studies showing no effect from randomized trials giving (or not giving) advice to teens about smoking, to heart attack victims about healthy living, and to new mothers about caring for their low birth weight babies.   Here is a new related result:

Affari Tuoi is the Italian prototype of the television show Deal or No Deal ...114 television episodes ... with large monetary stakes. When faced with a decision problem in Affari Tuoi, a contestant may seek advice from the audience, which comes in a form of the vote results. While there is a positive trend between contestants' decisions and advice, this relation is not statistically significant. ... When contestants do not have an opportunity to use advice or when the option of advice is available but not used, they make ex post "wrong" decisions in 52.9% and 54.6% of cases respectively. However, when they choose to consult the audience, the fraction of ex post "wrong" decisions decreases to 36.1%. Moreover, ... by following advice contestants increase their earnings (Table 1). Subjects make ex post "wrong" decisions in 46.2% of cases when they neglect the advice and only in 30.4% of cases when they follow the advice.

However, the literature does show that in some situations people seem to listen too much to advice: 

Schotter (2003) surveys several laboratory studies on advice when nonoverlapping “generations” of subjects play ultimatum and coordination games. In these studies (e.g. Schotter and Sopher, 2004, 2007) subjects often rely on the advice of naïve advice. ... who hardly possess more expertise or knowledge than we do.

So why do we not listen sometimes and listen other times?

Nanotech Views Value Driven

In another experiment conducted with the Washington-based Project on Emerging Nanotechnologies at the Woodrow Wilson International Center for Scholars, Kahan found that when volunteers heard about the risks of nanotechnology from different experts, they gravitated toward the views of experts who seemed to share their personal values -- individualists followed the lead of experts who appeared to be individualists, while people who believed in hierarchy were most likely to be influenced by experts who espoused similar views. Once volunteers decided which experts were most like them, it did not make a difference whether the experts said nanotechnology was risky or safe -- either way, the volunteers agreed with them. ... When people clash on hot-button issues, their disagreements may have more to do with clashing values than facts. One person may conclude nanotechnology is dangerous while another person concludes it is safe, but neither realizes their conclusions are being driven by underlying values that have nothing to do with nanotechnology.

That is from the Post.  Of course regarding policy conclusions, all else equal it does make sense to listen more to people who share your values.   But it seems a shame if your views about facts contain nothing more. 

Just Enough Expertise?

Karl Sabbagh on this year's Edge question:

I used to believe that there were experts and non-experts and that, on the whole, the judgment of experts is more accurate, more valid, and more correct than my own judgment. But over the years, thinking - and I should add, experience - has changed my mind. What experts have that I don't are knowledge and experience in some specialized area. What, as a class, they don't have any more than I do is the skills of judgment, rational thinking and wisdom. ... Most of us confuse expertise with judgment. ...

As a result of changing my mind about this, I now view the judgments of others, however distinguished or expert they are, as no more valid than my own. If someone who is a 'specialist' in the field disagrees with me about a book idea, the solution to the Middle East problems, the non-existence of the paranormal or nuclear power, I am now entirely comfortable with the disagreement because I know I'm just as likely to be right as they are.

I'm confident that, as an author and television producer, Karl Sabbagh frequently identifies others he thinks less likely than him to be right.  For example, he probably rejects most book or TV show concepts proposed by ordinary people, justifiably pointing to his superior experience and success in such ventures.  Thus Sabbagh thinks that on most topics he happens to have about as much expertize as is useful; less expertize than his hurts your accuracy, but more expertise than his doesn't help your accuracy. 

Debugging Comments

Back in October, Greg Mankiw cancelled comments at his blog:

Unfortunately, a few (usually anonymous) commenters too often crossed the line.  I just don't have the time to police comments and enforce good behavior, especially since some posts were generating more than 100 comments. And I don't want to host a party in which a small vitriolic minority consistently tries to ruin the event for everyone else. So I decided to turn the comments feature off.

Xeni Jardin acted similarly:

When our audience was small in the early days, interacting was simple. ... No moderation, no complication, come as you are, anonymity's fine. ... The audience grew. Fast. And with that, grew the number of antisocial actors, "drive-by trolls," people for whom dialogue wasn't the point. ... With much regret, we removed the comments feature entirely. ...