• We’ve got to trust our personal experience while stretching beyond it.

• We’ve got to trust our personal experience while stretching beyond it.
A friend of mine sees everything in racial terms. The drunk guy must be Irish, he says. Mexicans hardly ever smile, and they have enormous families. Salvadoreños are rude. Connecticut has better schools than Massachusetts because it has more Jews. He has peopled the world not with individuals but with moral categories in the guise of ethnic groups. When I challenge him, he’ll recite incidents at the grocery store or the beach to justify his prejudice. However he acquired the racism, he built it on a profound misunderstanding of how his experience reflects the world. He takes examples out of context, offers the most vivid or unusual examples to represent the group, doesn't consider alternative explanations, assumes his own behavior is the norm, never asks about the vast world he hasn’t seen directly. He builds great generalizations from a very small data set.*

As noted above, most of us are much more successful dealing with “worldly” problems than with abstract logic puzzles. “We understand and 'reason' about the world through the lens of schemas that are not guaranteed to deliver infallible truths about what we see and hear. Since the mind is by nature a knowledge medium rather than a logic machine, it tends to use schemas which tell us what is typically the case, how the world usually goes, and not what is necessarily true" (Campbell 90).

For the same reason, we’re not very good at understanding probabilities and statistics, either-- how what we observe relates to everything that’s out there, whether a particular instance is typical of the world or highly unusual. After all, until our ancestors learned how to keep written records, the pool of information we had to draw on, our sample population, was very small indeed. Even in “civilized” times, the great Church Fathers, Augustine and them, grew entire cosmologies out of their own tormented sex lives. And even now we tend to regard our lives and experience as the way the world should be-- and consider the rest abnormal, even wicked.

We prefer stories over statistics even when we have good data. In one experiment, subjects read profiles of people drawn from a group of 30 lawyers and 70 engineers. The profiles described personality and personal history, but made no mention of profession; subjects were asked to guess which job each person had. And though, given the information, each profiled person had exactly a 70% chance of being an engineer, that’s not what the subjects reported. Instead, they assigned professions based on characteristics such as math ability or political leanings. We don’t often consider the actual distribution or likelihood of events and conditions.

Why do we ignore base rates? Probably because in real life, we never know what they are. The more usual task for humans at large is generalizing some aspect of a small sample to an entire population, not the other way around. That is what happens in stereotypes. . . . We take our small sample of experiences and make conclusions about entire classes of objects, creatures, or events because we can never hope to sample every member of the entire class. Because we rarely know what the characteristics of an entire population are, we have very little experience in using base rate information during decision making. This is one reason why it is so hard to combat stereotypes and prejudice with factual information. Only greater experience with members of the maligned population helps-- assuming those experiences are good! (Cummins 192-3).

The problems of norming our own experience come up when we talk about this group or that which are discriminated against. Justice groups tend to rely a lot on anecdotal evidence about unfair treatment, hoping that the stories will convince by the force of their drama. Since we’ve all been treated unfairly at one time or another, on the one hand, we might sympathize. But the very act of identifying with the experience means we also apply our own experience of frequency. I’ve run into mean cops, too; I’ve got an incompetent boss, too; but I put up with them and so should you.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Campaigning to reduce the number of Jewish students, Harvard's President Lawrence Lowell told an alumnus that half the students caught stealing books from the library were Jewish. Later it turned out that the grand total of students caught that year was two (Lipset & Riesman 146).
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


• Sometimes we exaggerate the likelihood of what comes easiest to mind. The shortcoming is not so much that we ignore the base rate, but that we confuse how often something happens in the world with how often we think about it.

I know this because I need money. I need a lot of money. If I could win the lottery, I’d pay off my debts, get a new car, go visit my family, fund my favorite do-gooders, quit my job and travel around the world. I can picture the scene when I hand the big check to the director of the soup kitchen, and all the scruffy people standing in line with their plates and cutlery love me. I know the odds are against me to win the lottery, but maybe I could win. It would change my life. I believe I’ll buy me a ticket --no, two-- that will double my chances of winning.

Governments and other gambling interests make every effort to make winning seem more vivid and substantial than losing. My state, Tennessee, recently established a lottery to avoid a state income tax; it’s always easier to tax low-income people than the rich. And politicians can bet on our penchant for

“Exaggerating the Improbable. People tend to exaggerate the probability of very rare events. That is why so many people enter lotteries and why they buy airline disaster insurance.
“People are especially likely to exaggerate the likelihood of a rare event if its consequences are catastrophic. One reason is the availability heuristic, the tendency to judge the probability of an event by how easy it is to think of examples or instances. Certain kinds of catastrophic occurences tend to stand out in our minds and are therefore more ‘available’ to be recalled than others . . .” [tornados more than allergic shock, for instance, or accidents more than disease].
“People will sometimes work themselves up into a froth about unlikely events, such as dying in an airplane crash, yet irrationally ignore real dangers that are harder to visualize, such as a growth in cancer rates due to depletion of the ozone layer in the earth’s atmosphere” (Tavris & Wade 1995, 270).


*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~* Risk Assessment *~*~*~*~*~*~*~*~*~*~*~*~*~*

"The concern over an agent that causes a one-in-a-million change in the odds of prostate cancer is is obviously going to be much higher in a country where people survive to get prostate cancer . . . . the economic logic behind dumping a load of toxic waste in the lowest-wage country is impeccable."
-- World Bank chief economist Larry Summers, January 1991 (R. Bradley 17). 18 years later Summers became Obama’s chief economist.

Note that our tendency to overestimate the occurrence of more vivid or dramatic events is not the same as the claims of “risk management” gurus, that we exaggerate the risks of corporate high technology. Often these company hacks try to obscure what’s at stake with a veil of improbability, and to pretend that a risk that I knowingly choose is equivalent to a risk imposed on me. For instance, travelers rightly worry more about airplanes than about cars, despite the statistics showing that airplanes are safer per miles traveled. For one thing, a comparison of fatalities per trip would not be nearly so reassuring. But the big difference is that with cars we can increase our safety considerably with such simple measures as driving more slowly, or avoiding bad weather. By contrast, all a passenger can do in a diving plane is put his head between his knees and kiss his ass goodbye. And of course, it’s a whole planeload of people that die at once.

Aggregate numbers can mask big impacts in small places. In one Kentucky community, heavily polluted by a tannery, government experts dismissed the high rates of cancers and miscarriages because, with only a few hundred residents in all, the rates were not “statistically significant”. Or consider the 5.6% unemployment rate. Sounds not too bad, 5.6 out of a hundred. But the actual rate is much higher for some groups, and the consequences of losing a job more extreme-- we can lose our homes, health insurance, even our families. By contrast, a black Republican newspaper columnist is not likely to lose his job, and he has plenty of cushion and alternatives if he does.

The risk managers like to blather on most about nuclear power, and again they’re wrong.
a) despite all the fail-safes, because of the complexity of the process and the number of installations, nuclear reactors are bound to break down-- and have, many times, much more often than most people hear about.
b) when such breakdowns start cascading together, the consequences are almost immeasurably dangerous. The Chernobyl disaster in Ukrainia poisoned thousands of square miles for generations.
c) we can reduce many health risks, such as smoking, by individual actions. On the contrary, once a power plant goes kaput, running away is our only option-- if we get the warning in time.

No, the real problem with exaggerating the improbable is that we overlook the daily hazards and let the pols redirect our attention to what serves their purposes. In 2001 about twice as many people died in their places of work as in the World Trade Center, but the face we put on terrorism is Osama’s, not the folks at corporate HQ-- which suits Halliburton just fine.
*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*~*


• We tend to remember the instances which support a claim and overlook contrary or missing data. Researchers call this “confirmation bias,” and it’s worse with the small samples that are typical of our experience. My stereotyping friend never thinks to consider how many drunk people he’s seen in all, Irish and otherwise; the ethnic bckgrounds of all the frowning people he’s seen (under what circumstances); whether he’s met non-Jewish people who value education; the number of gringos with large families (like his own) and the number of Mexicans without.

Here’s one way to look at it. Let’s say you pray for a sister to recover from breast cancer. She does, which confirms for you the power of prayer. Then you start collecting stories from friends. You find 13 instances where people recovered from deadly diseases after they’d been prayed for. This is surely powerful proof. But did you also keep track of the other cases? In this kind of situation there are four possibilities: recovery follows prayer; no recovery follows prayer; recovery follows no prayer; no recovery follows no prayer. Let’s say there are a hundred very sick people and you prayed for half of them and half of them got better. Offhand that sounds impressive, but you have to look at the negative instances, too. You could have a 2 X 2 grid with results like this:

PATIENTS PRAYED FOR
Yes No

Yes 37 13
GOT BETTER?
No 13 37

These are very good results for prayer: more than twice as many people recovered after prayer as did not recover, and of the folks who got no prayers, more than twice as many died.

Or you could have results like this:

PATIENTS PRAYED FOR
Yes No

Yes 13 37
GOT BETTER?
No 37 13

This is disheartening news. If we looked just at the 13 in the Yes/Yes box, we might think that prayer is a very powerful medicine. But 13 out of 50 is not that great, especially considering that most of the folks who were not prayed for got better. So when you hear Pat Robertson say, “There is a woman in Kansas City who has sinus. The Lord is drying that up right now. Thank you Jesus. There is a man with a financial need . . . “ (Nickell 143), keep in mind not just how Pat might know all this --visions? voices? a fax from Up There?-- but how many women there are in Kansas City, how many have “sinus”, and how many are feeling better today.

Confirmation bias is a core element of the psychic and astrology business. Any psychic worth her salt will have a record of all the correct calls she’s made-- whether predicting the winner of the World Series or “seeing” the fact that you have “relationship problems”. The question not often asked is, What per cent of hits did you get of all the calls you made. Even rarer are records of predictions made before the predicted event.

This is not just a true believer’s selectivity. Psychologists Peter Glick and Mark Snyder asked both believers and skeptics to test astrology by questioning a person to see if his personality matched his horoscope. “The horoscope suggested that the individual in question was very extroverted. When asked to test this hypothesis, both believers and skeptics asked a large number of questions aimed at confirming the conclusion about extroversion (e.g., ‘Do you like to go to parties?’) and few questions aimed at denial.”

The person interviewed, a confederate employed by the experimenters, had been instructed to answer positively to all questions. Since most believers and skeptics asked only questions about being extroverted, the believers were confirmed in their belief in astrology, and the skeptics had to admit astrology was accurate. “Believers showed an additional bias in their use of the information received. As a group, they said the horoscope was accurate whether their probing of the confederate produced strong or weak support for that conclusion.” Only the skeptics who asked about the alternative possibility of introversion discovered that the horoscope was not accurate (Vyse 120). Apparently we’re not very good at uncovering the entire context of an event or observation.

Recall the early years of the second U.S. invasion of Iraq. Terrorists killed U.S. soldiers and Iraqis almost every day, and we saw the explosions on the nightly news. Halliburton told us to look at the big picture, all the ways Iraqis were better off than they had been before. And that’s right. To judge the occupation, we’d have to know at least what are the main concerns of most Iraqis. I imagine that, being in the line of fire, they were even more horrified by the bombings and murders than I am; but I didn’t know if that was their main worry and who they blamed and what hopes they harbored. Interviews helped, but we didn’t know how representative the individuals were. Possibly Iraqis themselves reassessed the situation from day to day. It was criminal of Halliburton to start the war on the basis of lies or no data at all, but it calculated correctly that the war would ensure the electoral victory in ‘04.

TV images of bombings in Iraq represented reality at least to the extent that this wass what Iraqis saw as well; but by themselves did not tell us all we needed to know about most Iraqis’ day to day experience. How many lost relatives? How did most rank the war among the other risks they faced? Presumably we will find out more in the postwar period, if and when we get there.

The related “Barnum effect” is a favorite tool of the huckster’s arsenal: beguiling us in such general terms that we, with our innate skill at developing stories from very little data, can always find something that sounds like our own situation. “World leader dies in 2005!” sounds like a safe bet. Likewise, I feel pretty confident that “You will turn a problem into an opportunity.” Or how about the TV medium who throws a bunch of questions at the audience, and (some) listeners are awestruck when some of the questions elicit sensible answers. We can see how these “predictions” or “readings” apply to ourselves, but often don’t notice they apply just as well to thousands or millions of others.


* Sometimes it’s hard to tell whether prejudices are learned misunderstandings or a conscious insistence on imposing one’s preferred order on the world; in other words less a matter of belief than of style. Katherine Anne Porter describes Gertrude Stein’s host of perhaps deliberately eccentric inanities. For example, Stein “found herself drawn to Picabia because his name was Francis. She had discovered that men named Francis were always elegant, and though they might not know anything else, they always knew about themselves” (Porter 265); “the Spaniard has a natural center of ignorance . . . . Painters are always little short round men” (259); “ . . . all Western geniuses worth mentioning were Orientals: look at Picasso, look at Einstein. Russians are Tartars, Spaniards are Saracens . . . ” 268). I imagine that Stein didn’t expect others to share her convictions, and might even have been offended if they did; it would have diminished her sense of her own special genius. So what does it matter what bizarre lies this woman told herself? I don’t like the possibility that, in elevating this kind of nonsense to the level of artistic insight, and building their politics on it, smug liars like Stein offer rightists a model of and justification for truth-denying self-satisfaction and contempt for the world.