Archive for the ‘Reality has a well-known liberal bias’ Category

The term “black swan” refers to an event that has enormous impact and is rare and difficult to predict. The prime example used by Taleb is 9/11 — an event of outsize importance that typical risk analyses would have been unable to predict or identify. There’s a fair amount of focus now on black swans, how to predict and control for them.

Sometimes, though, we are looking for black swans when we should be looking for white ones — events that have enormous impact, but that are neither rare nor particularly difficult to predict. An event like the Newtown murders may feel like a black swan — who could have predicted? But the preconditions for the event make it more of a white swan than a black one. For one, mass gun violence is relatively common in America; about 80 people die every day from gun violence. Although it is tragic that 27 murders and 1 suicide occurred in one place, and that many of the dead are children, it is not extraordinary in a world where gun violence occurs with regularity.

As long as it is easier for a mentally disturbed young man to get a handgun than mental health treatment, President Obama’s exhortations that we will do more to protect children strike me as hollow. Gun control laws have been eviscerated by the Supreme Court, and the gun lobby’s loud voice in the public conversation make movement on that front almost unimaginable. Mental health treatment is only slightly more likely, and my guess is that such laws would be targeted at committing people to quasi-incarceration rather than actually providing therapeutic treatment.

We cannot predict mass murder with precision, of course, but we can say with some probability that murders with guns will occur regularly through the day, week, month, year, etc. Without concerted efforts to either reduce the availability of firearms or increase the availability and reduce the stigma of mental health services, mass murder will continue to be a white swan rather than a black one.

(On the gun control topic for a second, I get all the 2nd Amendment stuff — we need to have firearms in case we need to overthrow the government. Sure. But a state monopoly on violence goes a long way to reducing violence among the populace. If this is an explicit trade-off being made, then fine, but I don’t think we have properly costed in the price of lost lives and mental/physical trauma.)

Read Full Post »

Politifact reminds me of Michael Scott when he’s trying to roast; they are so eager to roast and find “lies” that they don’t do any deeper introspection on the substance of the statements they are evaluating.

Consider, for instance, their recent 4-Pinocchio rating for the old chestnut that “women make 77 cents for the same job as men.” Take this, exaggerating Obama:

The Obama campaign took a legitimate statistic and described it in a way that makes it sound much more dramatic than it actually is. The 77-cent figure is real, but it does not factor in occupations held, hours worked or length of tenure.


Well, except look at those caveats–(1) occupations held; (2) hours worked; or (3) length of tenure. If we look at BLS occupation figures, (warning: PDF!) we see that women are still pulling down less than men in a variety of categories, so how might those three factors play into this equation?

  • Occupations held: Well, this is a problem, except even in the most specific BLS figures the disparity exists. Female chief executives make 69% what male chief executives make. Unless there is some massive difference in the types of “first-line supervisors of retail sales workers,” women are still making 79% of what their male counterparts make. That’s a pretty specific job type, and yet, the disparity still exists. Now, there’s no way of knowing exactly which job they hold; (PDF) research suggests that within the same establishment, wage gaps are smaller, but the overall pattern still holds that women are paid less than men for similar occupations regardless of establishment. So is Politifact right that the 77-cent figure for “the same work” is false? I mean, maybe, but the real figure might be something like 80-cents. How much better is that?
  • Hours worked: Well, it’s hard to say for hours worked; those pesky women are always taking so much more time off! Or not. BLS figures show that there’s a small difference in aggregate hours worked, but the difference between 8.2 hours and 7.8 hours, even considering overtime, doesn’t make up for the 23-cent gap, and as part-time employees, women actually work more than men. Plus, in “white-collar” occupations where hourly wages don’t matter (managers, supervisors, chief executives, elementary school teachers, accountants, social workers), women still earn less than men. Maybe they’re also working fewer hours there, which is why their companies reward them less. Or maybe something else. But they definitely earn less for similar work, if not the same.
  • Length of tenure: This one pisses me off the most. Maybe those lower-paid women have not worked there as long. True! But is that “not the same job”? A seventh-year teacher and a third-year teacher/cashier/clerk/nurse are doing the “same work.” They do, however, have different levels of seniority. And it turns out that men get promoted at a much higher rate than women (10.6% of men get promoted, as opposed to 7.6% of women), even though their wage growth at each level of promotion is similar. Again, this could be because men are just much better, hard-working, committed, etc. to their jobs than women. Or, perhaps, it could be that there continue to be discriminatory hiring, firing, and promotion practices at these establishments.

The point of this whole exercise is to illustrate that Politifact’s urge to get Obama in a hits-generating BOOM! ROASTED! moment has actually obscured the truth behind the statement. Just because it is difficult to get an exact comparison of apples-to-apples, doesn’t mean that women aren’t working for less pay doing essentially the same job. The bottom line is: women are promoted less often than men, earn less money than men in similar occupations (or “the same work”), and are subject to discriminatory hiring, promotion, and pay. Even if the 77-cent figure did take into account those three missing factors, it would still hit pretty close to the mark.

Read Full Post »

William Saletan atSlate always has these big pseudo-scientific questions that he thinks are deeply thought-provoking but are actually pretty schmeh.

For example, he has a long-standing (and probably wrong) hypothesis about race-linked intelligence. (I have previously noted the goofiness of this “scientific” discovery here.)

Now Slate has two stories about a study finding that children of gay parents at a time when being gay and having children was/is maligned can be stressful and difficult. That’s pretty understandable. Saletan’s take is probably more line with mine in that he thinks it still proves gay marriage is a good outcome (two parents, loving household, financial support, etc.). Still, he takes the study as methodologically sound (some criticisms here).

That said, let’s presume, for the sake of argument, that the study is right and two same-sex parents are actually in fact worse for the child than two opposite-parent biological parents. So what? Lots of children are raised in households without two opposite-sex biological parents; couldn’t a two-parent same-sex household still be better than, say, single parents? Couldn’t some alternate arrangement (let’s say, oh I don’t know, three parents in a household, or a two-parent biological household with grandparents in the home to provide childcare) provide even better results than the two-parent biological household? Should the government or society encourage such behaviors? Maybe, but maybe not. Attacking the “worst” child-rearing environments probably yields the most returns for society; certainly two-parent same-sex households are better than, say, institutional housing or constantly shifting foster care. Since there is high demand for same-sex households to have children, maybe we should be encouraging lots of adoption by any combination of two-parent households.

My point is that much like any presumed difference in intelligence between races (which, as I’ve noted, is probably wrong on its face anyways), the difference between a two-parent same-sex financially-supported household and a two-parent opposite-sex financially-supported household is probably so marginal that the policy implications are nil compared to the differences between a two-parent household and a no-parent household, or a two-parent household and an institutional care facility, or the difference between a poor family and a rich family.

Read Full Post »

The most popular TED talks tend to be uncontroversial and, to use TED’s own terminology, “jaw-dropping.” People want to stare in wonderment or be blown away by the advances we have made in the sciences or hear about a new discovery.

People don’t like to hear about race, poverty, and justice, so Bryan Stevenson’s TEDTalk is a bit outside of TED’s usual purview of Technology, Entertainment, and Design.

But to me, this fits perfectly into the question of design. When we design something, we must think about its purpose; design is more than attractive chairs. Institutional design determines how we administer justice in our society. What are the values we uphold? What does the design of our systems tell us about the answer to that question?

Stevenson’s talk doesn’t quite answer those questions, but it digs at the heart of what we believe to be a justice system and the shrugging apathy we afford it.

Read Full Post »

OK, it was a bad joke about how women should just keep their legs closed if they don’t want to get pregnant. And it’s easy for the Santorum campaign to shrug it off as a joke, just as it’s easy for us to laugh about it.

But the problem here is that it’s actually not a joke at all. Rick Santorum believes that birth control is a dangerous sin, and that abstinence is the only acceptable form of birth control.

That is to say, the joke about “squeezing an aspiring between your knees” as birth control is not “off-color” or “crossing the line”: It’s accurately depicting the Santorum campaign’s position on birth control and consensual male-female sex.

Santorum may say now that he thinks birth control should be available (one wonders why his argument doesn’t apply to abortion, but we’ll skip that), but his stance on birth control affordability/availability is clear: he believes birth control pills are bad, and that people should not use them and that abstinence is the only method.

“Aspirin between the knees” is the reality of the Santorum campaign, not just a one-off line by a surrogate.

Read Full Post »

Ta-Nehisi Coates has a great reply to Andrew Sullivan’s rather tone-deaf regurgitation of an unsourced story on a “p.c. egalitarianism” stranglehold on so-called “intelligence research.” He writes a takedown as it should be written — with respect for the author and no tolerance for bullshit.

The problem with Sullivan’s suggestion that researchers continue pursuing race-based intelligence is that it ignores what it is that “intelligence research” is supposed to find in the first place. Context matters, and the search for the elusive g — a measure of “general intelligence” — has largely led to skewed results in favor of the dominant ruling class of well-educated white males.

Let me give you an example. Here’s a common IQ test type logic puzzle that has a clear “right” answer (the Wason selection test, for those interested).

You are told to check if the following statement is false: “If a card has a vowel on one side, the number on the other side will be even.”

You are then given the following cards: A, G, 7, 4.

Which cards do you have to flip over to check if the statement is false?

The answer is to flip over card A (obvious) and card 7 (not as obvious). It doesn’t matter what’s on the other side of G, and if card 4 had a vowel or consonant on the other side, the statement could still be true.

But let’s frame the question a different way:

You are told to check if the following rule is false: “If a person is drinking alcohol, they are over 21.” You either know the beverage they are drinking or their age.

The four people you see are: age 16, age 22, drinking a beer, and drinking a Coke.

Who do you need to check?

Most people get this one correct. (See the paper here.) Why? Because we have experienced events like these and adapted our brains to understand them. I would bet that the 10 percent of people who get the vowel/number version correct have taken formal logic courses or have wrestled with such logic puzzles before.

When I was teaching, one of the standardized tests that the book included had an extended reading sample about a family’s experiences with skiing and snow. For my students, many of whom had never left the city of Chicago, such a reading sample was totally foreign. The test wasn’t testing their reading comprehension; it was, in many ways, simply testing their familiarity with skiing and its terms.

The search for g is a search for something that doesn’t exist — no intelligence is “general”; it shifts with the context of the activity and the world we live in. I’m not a big believer in the woo-woo Gardner multiple intelligences, but I think we should be honest when we discuss intelligence. We were not born with intelligence; our culture created it. Performance on cognitive tests is as artificial as anything else we can cook up.

Read Full Post »

Despite my schadenfreude at the hilarious Perry implosion at the “debate” last night, I want to point out that Perry is being lambasted for forgetting which three agencies he wanted to destroy, but he should have been disqualified for wanting to simply eliminate three federal agencies at all!

Instead, the whole Republican field is so far to the right that Perry’s position to eliminate federal agencies was downright moderate.

Rather than knocking Perry (or Cain) out for his awful policies and crazy ideas, it looks like the Republicans will probably just oust him for being a boob.

Read Full Post »

Older Posts »