fall 2005 - wisdom of crowds

Wisdom of the Crowds by James S
Reading notes

How to tap into the Wisdom of the Crowds, no method but need to satisfy certain conditions = diversity, independence, decentralization, three kinds of problems that groups face – cognition, coordination, cooperation

Diversity –meaningful differences among ideas not minor variations upon same concept, adds perspectives that would otherwise be absent and weakens some of destructive characteristics of group decision making (even more important for small groups, large groups normally have more diversity), development of knowledge depends on influx of the naïve, groups that are too much like each other find it harder to keep on learning because each member brings less and less information to the table, bringing in new members into an organization makes it smarter (even if the people are less experienced and capable, usually less than the person they replaced), diversity makes it easier for an individual to say what they really think, this all is about cognitive diversity

Independence – important for intelligent decision making for two reasons 1) keeps mistakes from being correlated 2) more likely to have new information than old data that others have, you can be biased and irrational but your independence does not make the group dumber, but are we really independent or able to act with only self-interest or are we embedded in social contexts, we are influenced by imitation and risk aversion = herding, information cascades = we supplement our information by looking at what others are doing and decisions are not made all at once but in sequence, at some point stop paying attention to our own knowledge and rely on the flow of others, depend more on public knowledge than private knowledge, then cascade stops becoming informative, imitation can be a rational response, no one knows everything and we can specialize, to improve organization’s decision making have decisions be made simultaneously, each person has to pay more attention to private information which can be wrong but collectively the group makes a better decision, more likely to be collectively right, encouraging people to make incorrect guesses makes the group smarter, reduce sequential element in the way people make decisions and get people to pay less attention to what everyone else is saying

Decentralization – what is = flocks of birds, free-market economies, p2p networks, public schools, modern corporations, decentralization fosters specialization (making people more productive and efficient) and increases scope of diversity of information (while the individual may become more narrow focused) and allows for tacit knowledge and the idea that the closer a person is to the problem the better to solve it, allows for individualization and specialization while being coordinated but there is no guarantee that valuable information will find its way into the system, the challenge is to find balance between making individual knowledge global and collectively useful while keeping it local and specific, need to aggregate information effectively

Coordination – it is possible for us to be coordinated even without talking to each other, people’s experiences of the world are often similar, this creates norms and conventions that regulate behavior but new rules can be formed quickly, conventions also reduce amount of cognitive work, first-come first-served seating in public places (subway, bus, movie theater, on the beach), not the best way to distribute seating but it is easy and internalized

Cooperation – need to adopt a larger view of self-interest than maximizing short term gain, have to trust those around them, societies need cooperation, laws alone wouldn’t work, but it is not rational to cooperate, “shadow of the future” by Robert Axelrod The Evolution of Cooperation, not about trust but about durability of relationship, the promise of continued interaction, but then we also cooperate with strangers such as donating to charities, buying on ebay, tipping, we see over time that trade and exchange are games that everyone gains rather than zero-sum games with only a winner and loser, reciprocity = key idea, we will pay our fair share of taxes (although we stand to benefit for services even if we don’t) if everyone does and there is a chance that those who don’t will get caught and punished, most of us are conditional consenters who cooperate if that is why the game works, it is important that we believe the system works, this creates a positive feedback loop and this is what creates societies rather than an random collection of people

Importance of small group work – small groups are ubiquitous to American life = juries, board of directors, small groups are different from large groups such as betters and stock investors, are inescapably influenced by others in the group with two consequences 1) making bad decisions tending to be more volatile and extreme 2) or can be greater than a sum of its parts by making people working harder and think smarter, two types of juries – evidence-based where do not take vote right away but sift through evidence and contemplate alternative explanations while there is also verdict-based juries where they see their mission to reach a decision as quickly and decisively as possible and usually take a vote before any discussion and debate concentrates on those who do not agree to agree, to help groups work better find a good way to aggravate their results and to include them in the actual decision-making process such as getting a vote (rather than in advisory function), group polarization = deliberation does not moderate but radicalize people’s points of view, social comparison is a factor in that if we are in the middle of group we shift to the side the group is going to keep ourselves within the middle relatively and when we move we are also moving the group to that place as in a self-fulfilling prophecy, in addition, talkativeness matters although there is no relationship between expertise and talkativeness and no relation between liking and talkativeness, talkativeness is influential (more men as jury foreman than women), sequence also matters and we tend to defer to higher status people although leaders tend to have higher opinion of themselves than justified, also extremists tend to be more sure and righteous of their ideas than moderates and will pull groups away from the middle, risky shift doesn’t always play out – sometimes groups will get less risky (conservative) as a more extreme response, if most of group already supports a position

Large group decision making – collective decision making is often confused with consensus, search for consensus encourages tepid, lowest-common denominator solutions that do not offend anyone rather than exciting everyone, does democracy (of the workplace) mean endless discussions or wider distribution of decision-making power, be careful not to think that intelligence is fungible (effective in every context), baiting crowds for suicide jumpers is how riots work, in the middle of riot they operate as one organism, some people will never riot and some will riot all the time but most people are in the middle, willingness to riot depends on what other people in crowd are doing, not if one person riots but if there is a mix of sizable people involved

Democracy – National Issues Convention Deliberative Poll, what is democracy for – do we have it because it gives people a sense of involvement and control over their lives and thus contribute to political stability or is it for individuals to have the right to rule themselves even if used in ridiculous ways, is democracy a vehicle for making intelligent decision and uncovering truths, studies of self-interest and politics, if it was solely about self-interest then government would increasingly get bigger and would also be ruled to serve the interest of powerful groups rather than the public as a whole, self-interest is important to voters but not the only factor (in fact, why would anyone vote at all, despite low turnout), ideology makes a difference, having all the knowledge is not important in a representative democracy, politicians have the information they need for good decisions while voters monitor the outcomes based on local knowledge, we need both “experts” and non experts to function well, How do we live together?, How can living together work to our mutual benefit?, the democratic experience is not getting everything that you want, seeing your opponents win and accepting it because you believe that they will not destroy things that you value and you know that you will get another chance to get what you want, compromise and change

fall 2005 - Blink

Nov. 23 2005
Notes on Blink by Malcolm Gladwell

Intro Adaptive unconscious – our brain quickly and quietly processing information we need to function as human beings, rapid cognition, which we tend to be suspicious of because we like to trust our conscious decision making abilities
Our unconscious is powerful but fallible – how then to trust our instincts?

Three premises 1) decisions made quickly can be every bit as good as decisions made cautiously and deliberately 2) what are the conditions that cause our rapid cognition to be wrong 3) snap judgments can be educated and controlled

I. Thin slices – Gottman’s studies of couples on (one hour of) videotape can predict with 95% accuracy if couple will be married in 15 years, some can it as quickly as three minutes (and not while fighting), nonexperts can make this same prediction (same tapes) as accurately as 80%
Two states of marriage 1) positive sentimental override (a buffer) 2) negative sentiment override (four horsemen of problems = defensiveness, stonewalling, criticism, contempt with contempt by and far being the worst) criticism is global condemnation of person’s character but with contempt it comes from a “superior” plane, hierarchical, this also predicts how many colds someone gets
Morse code analogy – “fists” the personal signature of someone’s sending, distinctive qualities appear naturally and automatically
Job interviews – personality workup – the Big Five Inventory
1) Extraversion. Are you sociable or retiring? Fun-loving or reserved?
2) Agreeableness. Are you trusting or suspicious? Helpful or uncooperative?
3) Conscientiousness. Are you organized or disorganized? Self-disciplined or weak willed?
4) Emotional Stability. Are you worried or calm? Insecure or secure?
5) Openness to new experiences. Are you imaginative or down-to-earth? Independent or conforming?
Compare thick slice – how a friend would rate you with someone who didn’t know you visiting your dorm room (gives us three clues to someone’s personality 1) identity claims or deliberative expression of how we would like to be seen by the world 2) behavioral residue such as dirty laudry or organized CDs 3) thoughts and feelings regulators or changes we make to our personal spaces to affect the way we feel when we inhabit them
What people tell us about themselves can be confusing, we are often unaware of how we are, don’t ask point blank a couple about their relationship ask them about their pets or such
Doctors – training vs. conversation – studies of transcripts
Not sued doctors spend more than 3 minutes longer with patients (18.3 mins. Vs. 15), also more orienting comments and active listening but no difference in amount or quality of information and not more details about meds or patient condition but how they spoke, treated patients as human beings


II. Snap judgements are extremely fast but unconscious – behind a locked door
Priming experiments – suggest certain ideas using words (polite, rude, aged, think like professor, or to identify race on test, cooperativeness) and tendency of people to act in these ways, not brainwashing because you are not revealing deep personal information or act in drastically inappropriate ways, priming doesn’t work once you become aware of it but we are not aware of it at the time and do not credit it’s effects
Unconscious works like a mental valet taking care of the minor mental details of you life leaving you free to concentrate on the main problem at hand
Explaining yourself – we do not explain our snap judgements accurately, what we say and what we do are different, often provide different answers at different times or meaningless answers – a storytelling problem – explanations for things we don’t have explanations for, we need to be cautious in how we interpret other’s explanations for their instinctual behavior (we rarely say we are ignorant, that we “don’t know”)
Led astray by snap judgments – take rapid cognition seriously – acknowledge it power for better or worse – is a step toward controlling them

Implicit Association Test (IAT)

https://implicit.harvard.edu/implicit/

we make connections more quickly between pairs of ideas that are already associated in our minds than between pairs of ideas that are less related in our minds

do it as quickly as possible

don’t skip over words

don’t worry about making mistakes

strong prior association = difference in milliseconds

predictor of how we will act in certain types of spontaneous situations

not aware of bias but will, nonverbally act slightly differently and in such ways that others can perceive this (such as job interview)

study of half of CEOs of Fortune 500 are white males but also in height, generally average almost 3 inches higher than general male population (58% compared to 14.5 as over 6 feet) – it may be that our selection process is less rationale than we think

III. controlling rapid cognition – spontaneity isn’t random, good decision making under fast-moving high-stress situations is a function of training, rules, and rehearsal
as in comic improve – the concept of agreement, have characters accept everything that is happening to them, develop action rather than block it
create conditions for successful spontaneity
watch out for “verbal overshadowing” – use of words and introspection can separate us from our instincts and impede problems that can be solved with a flash of insight. Problems that could be solved with a flash of insight have a different set of rules than other problems. Language can get in the way, sometimes stopping to talk things over gets in the way
Less is more – Lee Goldman’s algorithm for heart patients – simplified what was important to look for. Too much information (which we often intuitively try to seek out especially when challenged or threatened or stressed about what to do). But how to ignore perfectly valid information? Particularly doctors who do not want to follow mundane guidelines, more gratifying to come up with decision on your own. With more information we become more certain of our answers but our answers do not necessarily get better or more consistently correct.
Two lessons here 1) successful decision making relies on balance of deliberative and instinctive thinking 2) in good decision making frugality matters – distill to simplest elements, clear patterns
Snap decisions can be made in a snap but not if you have too many choices (jam experiment – six choices 30% sales, twenty four choices 3% sales)
IV. Louis Cheskin (marketing guru) sensation transference – transferring the impressions or sensations from the packaging to product itself, the product is the package and product combined
Need to thin slice by also looking at the context, hard to do market research on something really different, first impressions don’t really count in these cases
First impressions of experts are different from others, our tastes become more complex and can explain their understanding better, have a vocabulary and rubrics/matrix for assessing things. We cannot look behind the locked door of unconscious but with experience we become expert at using our behavior and training to interpret and decode what is behind snap judgments and first impressions. This changes the nature of first impressions. Outside of our areas of expertise our impression may not be wrong but more likely shallow, hard to explain and easily disrupted. Just because we do a lot of something (drink cola) doesn’t mean we think about it a lot.
V. FAC facial action coding system can teach us how to better read each others faces – the face is not just a signal of what is going on inside our mind but it is what is going on inside our mind, emotions can start with the face, not just be a residue or secondary billboard for internal feelings, it is an equal partner
Micro expressions not made voluntarily, and inevitably no matter how hard we try to suppress them emotions do leak out and can be read, we don’t actually know what is on our face, in fact, this would make us better at concealing things which may not be a good things at least evolutionary wise. Most of time we read facial expressions easily and accurately so what happens when this breaks down.
We have different parts of brain for facial recognition and object recognition. Autism looks at faces as objects. What if this was also a temporary condition instead of a chronic one? Such as under extreme stress – extreme visual clarity, tunnel vision, diminished sound, sense that time is slowing down – police that don’t hear gun going off
An optimal state of arousal – Larry Bird seeing the whole court, we can exceed the optimal state and be too aroused, too much feeling of pressure and lose souces of information were we become useless, motor skills slow down as does cognitive processing and vision, blood withdrawn and we become clumsy, arousal can leave us mind-blind
Time is a factor – need “white space” time to react, without time you are subject to the lowest-quality intuitive reaction, can cause temporary autism when we run out of time, we can be guided by stereotypes and predijuces, when police officers by themselves (rather than in squad car with a partner) slow things down, less bravado, will not charge in as rapidly or be ambushedand have fewer problems, with partner they speed things up split second syndrome – rush into situation with no time for thought, can go through stress inoculation training with exposure to similar stress situations and over time heart rate goes down
The key to reading faces is practice, deliberate practice – the gift of training and expertise to extract an enormous amount of meaningful information from the very thinnest slice of experience, not a blur but a blink
VI. we can be careless about our powers of rapid cognition, need to take these powers seriously and acknowledge the subtle influences that can alter or undermine or bias the products of our unconscious
If we control the environment in which rapid cognition takes place then we can control it better