Committing Sociology

April 27, 2013
Posted by Jay Livingston

“This is not a time to commit sociology,” said Canada’s prime minister Stephen Harper. 

It’s witty, especially if you don’t know that Auden made the same joke nearly seventy years ago.
Thou shalt not answer questionnaires
Or quizzes upon World-Affairs,
    Nor with compliance
Take any test. Thou shalt not sit
With statisticians nor commit
    A social science.
But it’s wit in the service of a bad idea – willful ignorance.  The less witty version is the introductory clause, “We don’t need ‘studies’ to know that . . .” with the word studies almost always in quote.   As I’ve said in earlier posts (here and here) the phrase is pretty much a guarantee that the writer has no systematic evidence or that the available evidence points in the opposite direction.

It’s not so bad when the sentiment comes from a poet few people know of. But when it comes from people with real power, it can do real damage.  Here in the US the Republicans in Congress don’t like political science research.  Understandably.  But they are not just clapping their hands over their ears and shouting, “I don’t hear you.”  They are saying, “I won’t fund you.”  And now some of them want to eliminate funding for all science that can’t wave a patriotic flag.
 the new chairman of the House of Representatives Committee on Science, Space, and Technology floated the idea of having every NSF grant application include a statement of how the research, if funded, "would directly benefit the American people." Representative Lamar Smith (R-TX) said that he was not trying to "micromanage" the $7 billion agency but that NSF needs to do a better job of deciding what to fund given the low success rates for grant applicants and a shrinking federal budget.  (More here.)
The sociology Harper was referring to consisted apparently of remarks by Justin Trudeau on the topic of terrorism.  Trudeau suggested that a strategy to prevent further terrorism should include a consideration of “root causes.”

Naive Trudeau.  Had he been more familiar with to the recent history of his neighbor to the south, he would never have used that phrase.  When crime was in the US rose drastically in the late twentieth century (when Justin’s dad Pierre was prime minister of Canada), some people suggested that to reduce crime, it might help to understand “root causes.”  Conservatives, the defenders of “law and order,” hooted with contempt.  We didn’t need to understand. We needed to punish the bad guys, the more harshly the better. 

The same reaction seems to have been taking place in Canada in the days following the Boston bombing and the discovery of a plot to blow up Canadian trains.  Apparently that is what Harper meant by “this is not the time.”  At a time like this, when people are “uncertain and afraid” (Auden again), they do not want to understand.  They want reassurance both of their safety and of their moral rightness.  They want actions and words that reinforce the boundary between Us and Them.

The trouble, especially with potential terrorists in our midst, is that we need the help of people who look like Them. Terrorism plots are foiled by information from insiders.  Do we really want to paint the boundary in bright colors and force them to choose a side?  Here is the sociology that Trudeau committed.
But we also need to make sure that as we go forward, that we don’t emphasize a culture of fear and mistrust. Because that ends up marginalizing even further those who already are feeling like they are enemies of society.

Wanted – Bad Research

April 22, 2013
Posted by Jay Livingston

I’m not a research director.  But if I were, I hope I wouldn’t write questions that are obviously designed to bias the results.*  And if I did ask such questions, I wouldn’t boast about it in the newspaper, especially if my stacking of the deck got barely a majority to give the answer I wanted. 

But then, I’m not Michael Saltsman, research director for the Employment Policies Institute, whose letter to the Record (formerly known as The Bergen Record) was published today.
Regarding "Most favor minimum wage hike" (Page L-7, April 18):

The recent Rutgers-Eagleton poll finding that 76 percent of New Jerseyans support a minimum wage increase only proves that incomplete poll questions yield misleading results.

My organization commissioned ORC International to conduct a similar poll regarding an increase in the minimum wage. When respondents were informed of the unintended consequences of minimum wage hikes — particularly how such hikes make it more difficult for the least-skilled to find work— 70 percent support flipped to 56 percent opposition. [emphasis added]

This consequence isn't a hypothetical: Fully 85 percent of the most credible economic studies from the past two decades indicate a loss of job opportunities following a wage hike.

Michael Saltsman
Washington, D.C. , April 18
As for the facts on the effects of an increase in the minimum wage, Saltsman’s literature review is on a par with his questionnaire construction.  Apparently he missed John Schmitt’s CEPR article from two months ago (here).    The title pretty much sums it up:
Why Does the Minimum Wage Have No Discernible Effect on Employment?
Schmitt includes this graph of minimum-wage effects from a meta-analysis.

Hristos Doucouliagos and T. D. Stanley (2009) conducted a meta-study of 64 minimum-wage studies published between 1972 and 2007 measuring the impact of minimum wages on teenage employment in the United States. When they graphed every employment estimate contained in these studies (over 1,000 in total), weighing each estimate by its statistical precision, they found that the most precise estimates were heavily clustered at or near zero employment effects.
Schmitt offers several guesses as to why employers don’t cut jobs when the minimum wage rises – maybe they raise prices, or accept a lower profit margin, or reduce the wages of better-paid employees; or maybe the increased minimum wage brings more customers, and so on.**

But regardless of the findings on minimum wage, Saltsman’s letter carries a more important if depressing message.  We try to teach our students to design good research.  We tell them that good research skills might help them get jobs.  Yet here is an example of a research-director job that depends on designing bad surveys and doing bad research. 
*In his methods course, my colleague Chris Donoghue uses a made-up abortion item for teaching items that introduce bias:
“Every year in the US, over a million babies are killed by abortion. Do you agree that laws should make it more difficult to get an abortion?”

** Brad Plumer at WaPo’s WonkBlog has more on this, including a fuller discussion of Schmitt’s paper (here).

Ulysses in LaLa Land

April 17, 2013
Posted by Jay Livingston

A non-sociological post. 
I’ve never been all that good at resisting the obvious.

From today’s New York Times:
The indictment also named Molly Bloom, who made headlines in 2011 for her role in arranging clandestine games for high-rollers, including Tobey Maguire, Leonardo DiCaprio, Matt Damon and Ben Affleck.
And then he asked me if he could get into the game yes and would Matt be there yes and Ben with himself so pumped up and proud yes yes and Leo too all blond and often staying in with only king seven yes and he took out his checkbook and asked with his eyes if this would be enough for the buy-in yes and it was a wondrous number with lots of zeros yes yes yes

Underground Demography

April 16, 2013
Posted by Jay Livingston
Cross-posted at Sociological Images

The magic of demographic knowledge is a memorable moment in John Sayles’s 1984 movie “Brother From Another Planet.”   On the A train, a young man shows an elaborate card trick to the title alien, who looks like an African American but seems to have no understanding of the trick.  So the magician offers another.

From 59th St. to 125th St. is one stop on the express.  But as the movie shows, that short ride covers a large demographic change, and it’s not just racial.  The New Yorker has posted interactive graphics (here) showing the median income of the census tracts surrounding subway stations.* 

Take the A train one stop  – from the southern border of Central Park to a few blocks above its northern border – and see median income drop by $100,000. 

Many other lines travel the extremes of economic inequality.  My line is the 2. 

In the early morning commute, I see blue collar workers in their hoodies or rough jackets and steel-toe boots next to well-dressed people reading The Wall Street Journal.  They didn’t get on at the same stop.  The people who live in and work in the Wall Street census tract, which includes Park Place, are not on the train.  Here’s what their housing looks like.

And here is Franklin St., Brooklyn.

The subway demographic trick is not limited to New York. Here’s a time-lapse video of the Red Line of Chicago’s CTA.
(If the video does not play, you can see it here.)
Despite the social class segregation in housing, in cities like New York and Chicago, people of vastly different economic circumstances are likely to share the same subway car, at least for a few stops. 

Yet I don’t get a sense of strong resentment or even envy among the have-nots (though I wish I had systematic data on this).  These cities are also where the rich are more likely to be liberal and in favor of redistributionist policies.  As Andrew Gelman has shown, the wealthy in rich states are far more liberal than the wealthy in poor states.  That may be partly because in rich states, the wealthy live in the large cities.  How strong would that effect be if we used Upstate New York, Downstate Illinois, Massachusetts outside Rte. 128, and so on?

Or to quote James Carville’s famous line about Pennsylvania: “Philadelphia in the east, Pittsburgh in the west, and Alabama in between.”

HT: Jenn Lena for the link.

Bitcoin - "A Currency Without an Army*

April 15, 2013
Posted by Jay Livingston

I must be missing something in Paul Krugman’s dismissal of bitcoins (today’s NYT column here).  Krugman says that unlike gold and paper money, the value of bitcoins does not rest on some intrinsic usefulness or upon the power of a state. 
Bitcoins, however, derive their value, if any, purely from self-fulfilling prophecy, the belief that other people will accept them as payment.
Then a few paragraphs later, he approvingly quotes Paul Samuelson saying that money is a “social contrivance.”  What makes paper, silver, or gold worth something is “the expectation that other people would accept them as payment.”

So bitcoin and metals and paper all depend on socially constructed definitions.  But then how is bitcoin different from more traditional kinds of money? 

Or does Krugman mean that because the free-floating bitcoin is untethered to precious metals or governments, those definitions are less stable and that the bitcoin’s value is more susceptible to the mood swings of the public? (FWIW, the price of gold has fallen 13% since Thursday.)

*The subject line of this post is a variant of what has often been said of Yiddish – a language without an army

Special Victims

April 13, 2013
Posted by Jay Livingston

An op-ed, by Glenn McGovern in the Wall Street Journal (here but behind a paywall) says that attacks on prosecutors are on the rise.  McGovern begins with the recent shootings of district attorneys in Texas.  Then he says,
Each year in this country, well over 100 police officers are killed in the line of duty.
That number is correct in a technical sense, but since McGovern is writing entirely about “attacks” on law enforcement officials, it’s misleading. Most police officers who die on the job – usually about 60% – are killed in accidents. 

 The number of lethal attacks on police never gets close to 100.

As for attacks on prosecutors, according to McGovern, the number for this decade, as of April 1, 2013, is 15.  By comparison, as of the same 44-months-into-the-decade* of the 1990s and 2000s, only six such acts of violence had been recorded in each of those periods.  He finds no “geographical logic” or other demographic patterns in these attacks.  But with a total of only 27 attacks over a 23-year period, differences would show up only if they were extreme. 

The 543 killings of police officers in the last decade do indeed show regional difference.

(Click on the graphs for a larger view.)

Both in absolute numbers and rates per population, cop-killing is most prevalent in the South.  My first guess was that this had to do with the greater prevalence of guns in the South.  It’s no surprise that guns, especially handguns, are the most frequent weapon when cops wind up dead.

But when it came to choice of weapons, differences between regions were minimal and in an unexpected direction.  In the South, about 3.5% of the weapons used in nonlethal attacks on the police were guns (not including “personal weapons,” i.e. fists and feet).  For the Northeast percentage of guns was slightly higher – 4%.  Yet the South kills far more police.  So if it’s not the choice of weapon, we are left the Southern culture-of-violence explanation: When Southern men feel they have been seriously wronged, they are more likely to use violence to defend their honor. 

It’s in the South that we are most likely to find “stand your ground laws” allowing the deadly defense against the intrusions of other people.  It’s also where we’re more likley to hear anti-gun-control arguments based on the idea that guns are necessary to defend against the intrusions of government. 

This explanation should hold for attacks on prosecutors as well.  As McGovern says, the prosecutor must almost inevitably denigrate the honor of the defendant:
For hours and hours over many days and weeks, under the glaring eyes of a defendant seething with anger, these prosecutors argue to a judge or jury that this person should be locked away for life, or even forfeit his life.
The number of incidents is too small to reveal patterns of regions of urban vs. small town.  Let us hope that it remains that way.  Sometimes a small n is just what we want.
*I’m not sure how April 1 is 44 months into the decade rather than 39. 

AKD 2013

April 10, 2013
Posted by Jay Livingston

Last week, seventeen Montclair students were officially inducted into AKD, the sociology honor society.   We were glad to have so many, especially since we raised the GPA bar this year. 

(Click on the picture for a larger view.)
Left to right.
  • Lisa Kaiser
  • Desiree Velez
  • Keri Anne Hart
  • Tom Rorke
  • Jackie Cano
  • Rachel Druker
  • Luis Bernal
  • Kristine Nemec
  • Jamie Sommer
  • Liz Sondej
  • Joseph Della Fave
  • Rachel Lyn Matthews
  • Armita Haghshenas
  • Victoria Sirianni

Not in the picture:
  • Jovo Bjelcevic
  • Kelly Orosz 
  • Sean Wilkinson

We even managed to get students and faculty together for a photo op.

Front row: Yong Wang, Chris Donoghue, Sangeeta Parashar, Yasemin Besen-Cassino, Faye Allard, George Martin, Arnie Korotkin
Second row: (faculty/staff only):  Janet Ruane, Jay Livingston, Susan O'Neil, Bob Podhurst

Our speaker was Paul Hirschfield of Rutgers who has been looking at school bullying, focusing the cases that ended in suicide (“bullicide”). He pointed out the obvious shortcomings of purely psychological explanations.  Whatever the common traits of bullies or of victims, these don’t explain why bullying increases in prevalence or intensity in some schools and not others.  He had a similar criticism of the explanation that focuses on culture.  “Lazy sociology,” was his term for it.  Instead, he looks at social structure, particularly status hierarchies.  Among kids, hierarchies can promote and sustain bullying.  Among teachers, those at the top of the status hierarchy – the more senior teachers – may take an old-school non-interventionist, non-alarmist view.  Younger teachers who might want to intervene “risk making powerful enemies” (high-status students).  In his review of 50 bullicide cases, Paul noticed that teachers were especially less likely to challenge anti-homosexual bullying.

We had a pretty good turnout.

And faculty and students could get together.

Drunk as a Lord

April 6, 2013
Posted by Jay Livingston

Like Andrew Gelman, I’m puzzled by Tyler Cowen’s assertion about alcohol:
There is an elite which has absolutely no problems handling the institution in question, but still there is the question of whether the nation really can have such bifurcated social norms, namely one set of standards for the elite and another set for everybody else.
Do the elite really have no problem handling alcohol?  I guess it comes down to definitions of problem and handling.  Despite examples like Ted Kennedy, most drunks don’t kill people.  More to the point, Kennedy’s elite status insulated him from the worst consequences of his fatal drunken driving.  It’s good to be the king.  Or a Kennedy.*  No doubt, many among the elite can’t handle liquor, but they don’t have a problem.  Even for middle-class people with less economic and social capital, drunkenness and even alcoholism need not be a problem. 

As long as the drug is the sole preserve of the elite, it’s not a problem for society either.  But what happens when a drug becomes democratized? Until the 1980s, cocaine had, thanks to its cost, been confined mostly to the elite. Then, in its inexpensive form, crack, it became widely available to the masses.  Suddenly, it was a social problem.  In typical fashion, US policy-makers defined the problem as criminality and dealt with it by enacting more and more draconian punishments. 

Those new laws amply illustrate what Tyler Cowen refers to as “bifurcated social norms, namely one set of standards for the elite and another set for everybody else,” except that these weren’t norms, they were laws. The sentence for selling  5 grams of crack was the same as for selling 500 grams of cocaine. As for that other set of more punitive norms for “everybody else,” guess who“everybody else” was.  Mostly Black people. 

The changing demographic for cocaine and the reaction to that change paralleled what happened in the “gin crisis” in England only a few hundred years before. [I am now going to recycle some paragraphs and a jpeg from a post I did five years ago.]

Up until the 1730s, only the wealthy, propertied classes could afford distilled spirits, mostly brandy. It’s not that they didn’t drink to excess – the phrase “drunk as a lord” dates back to the mid-1600s – but their drinking wasn’t a social problem.

Then came cheap gin and the democratization of drunkenness. The lower classes had the tuppence to get drunk as a lord. But they lacked the means to keep the drunkenness from becoming a problem. I suppose it didn’t really matter if the lords were too drunk to work; their wealth insulated them, their families, and the society against the drawbacks of drunkenness. Not so the inhabitants of Hogarth’s Gin Lane.

What followed were the gin laws of 1736, so discriminatory that they provoked riots. That may be the main place where the parallels between gin and crack diverge. It’s hard to imagine people taking to the streets over the 100-1 cocaine-to-crack law in the way that lower-class Londoners rioted to protest the gin laws. But then, lower-class Londoners did not have the vote; the streets may have been their only avenue for political action. In any case, the gin laws were not very effective (back to the parallels with crack), but after fifteen or twenty years, the crisis had run its course, and lower-class drinking was no longer a threat to the integrity of society.

* As Clone High viewers know, nothing bad ever happens to the Kennedys. (HT: Max)

Social Science Evidence and the Court

April 4, 2013
Posted by Jay Livingston

Doug Hartman at The Society Pages  is upset about Justice Scalia’s casual and inaccurate summary of social science data.   During the oral arguments on DOMA, Scalia said 
There’s considerable disagreement among sociologists as to what the consequences of raising a child in a single-sex family, whether that is harmful to the child or not.
This is a bit like saying that there’s considerable disagreement among climate scientists as to whether the earth climate is getting warmer. 

Doug Hartman concludes:
 For Scalia and his ilk, there is no real knowledge in the social sciences, no authority. Not even any real data or useful information. Just a lot of disagreement and differences of opinion.
The title of his post is “Scalia Takes It from ‘Bad’ to ‘Really Bad.’” That still may be understating things.  It’s not just that Scalia sees social science as mere opinion.  But even when the scientific conclusions are irrefutable, Scalia finds social science knowledge irrelevant.  At least when that knowledge is inconvenient for his argument.

The case I have in mind is McCleskey v. Kemp, decided in 1987.  Scalia is the only member of that Court still on the bench. He didn’t write the opinion, Justice Powell did, but Scalia was apparently in full agreement. 

McCleskey was a Georgia death penalty case.  McCleskey, a Black man, had killed a White man.  The defense presented the findings of a careful study by David Baldus on race and the death penalty in Georgia.  He had looked at 2500 murder cases and concluded, even after adjusting for dozens of other variables, that race made a difference in capital sentencing.  In cases with Black defendants, prosecutors were slightly more likely to seek and win the death penalty.  The race of the victim weighed even more heavily. When the murder victim was White, prosecutors were four times more likely to seek the death penalty.  Unsurprisingly, the cases most likely to bring a death sentence were those like McCleskey’s – Black defendant, White victim.

The underlying assumption of prosecutors and perhaps jurors seems to have been that White lives were more valuable than Black lives.*  The taking of a White life, whether by an individual or by the state, was a much more serious event.

Regardless of the accuracy of the Baldus findings, in the majority opinion, they were irrelevant.  The study may have shown a general bias in the system.  But that didn’t mean that the Equal Protection clause of the Fourteenth Amendment applied to this particular case. 
The statistics do not prove that race enters into any capital sentencing decisions or that race was a factor in petitioners case. [emphasis in original]
To win his case, McCleskey would have to show that the prosecutors in his particular case were acting on racial prejudice.  If the racism was unconscious, that would be an impossible task.  And even if prosecutors were aware that they valued White lives above Black and were acting on the basis of that evaluation, it’s unlikely that they would have been writing memos revealing their prejudice.

The majority did have a point.  You can’t use aggregate data to establish a connection in any single case.  That’s the ecological fallacy.  But the Court could have said that Goergia’s death penalty system was so tainted by racial prejudice that it would have to be suspended.  Instead, the court said,
At most, the Baldus study indicates a discrepancy that appears to correlate with race. Apparent disparities in sentencing are an inevitable part of our criminal justice system.
The Court had moved far from its weighing of social science evidence in Brown v. Board of Ed.  In that case too, as my colleague Jessica Henry reminds her students, the Court could have said, after reviewing the data, “Apparent disparities in education are an inevitable part of the school system.”  Instead, it said that those disparities were in violation of the Equal Protection clause and that school systems must reduce those disparities by desegregating. 

The message in McCleskey was much different, the Court tossing the data aside and saying in effect, “It’s racist, it’s unfair.  Get over it.”  I doubt that Scalia’s relation to social science data is any different today.

* There’s a quotation often attributed to an unspecified 19th centurty Southern prosecutor or judge:  “If a Black man kills a White man, that’s capital murder.  If a White man kills a Black man, that’s justifiable homicide.  And if a Black man kills a Black man, that’s just one more dead nigger.” The quote may be apocryphal.  The sentiment and cognitions it expresses were real then; they were real at the time of the Baldus study; and they may still be real today.

Plenty of Fish . . . For Now

April 2, 2013
Posted by Jay Livingston

In 1950s, the stereotype was that girls went to college not so much to get a B.A. but to get an M.R.S.  Not such a bad idea, says Susan Patton.  She’s the Princeton alumna whose letter of advice to undergraduate women was published in The Daily Princetonian:
Find a husband on campus before you graduate.
There was much reaction – pro (WSJ) and con (HufPo and just about everyone else).  (Patton has since been on TV, and the traffic to get her letter online crashed The Daily Princetonian’s Website.)
Oddly, none of these responses – at least the ones I’ve seen – addresses the basic question with actual data.  Is there no empirical evidence on this? Hasn’t anyone done a survey of women who went to elite universities?  Such a survey would surely have included a question on how you met your husband and how old you were when you got married.  And surely there would be outcome variables – satisfaction with different areas of life, including marriage.* 

We do know some facts. On the whole, college-educated women are delaying marriage.  Presumably, the longer they wait after graduation, the less likely it is that they are marrying a college sweetheart.  According to Patton, that strategy is a loser.  Yet at the same time, divorce rates among the college educated women are declining.  As for their happiness in those marriages, I can’t even guess. The GSS shows no clear trend. 

But the Ns are small, (40-60 through the 1990s, and after that, on average, about 100), and the categories (“Very happy,” “Pretty happy,” “Not too happy”) may not capture the full range of how women feel about their marriages. 
Besides “college educated” is not the same as “Princeton educated,” and Patton says explicitly that the reason for finding your husband at Princeton is that you want to be sure to marry not just any college graduate but someone who is at least as smart as you.
Simply put, there is a very limited population of men who are as smart or smarter than we are. And I say again--you will never again be surrounded by this concentration of men who are worthy of you.
(You can see why some people thought Patton was just a tad elitist.)

Haven’t we been here before?  Yes we have. Readers of a certain age may remember the 1986 Newsweek article that caused a similar stir.  I think the article focused on Yale women for its journalistic anecdotes, but the statistical conclusion – the line that went as viral as a line could go in the pre-Internet age –  was that if a college-educated woman was still single at age forty, she had a lower probability of getting married than of “being killed by a terrorist.”  And that was before anyone had heard of Al Qaeda. 

As with the Patton letter, there was much criticism of the article and its conclusions.  And Newsweek issued a retraction . . .  twenty years later.

 * Sources who know much more about this than I do (Philip Cohen) suggest that some surveys like the NFSH might have trend data on these questions.
UPDATE April 4:  Philip has culled the ACS for data on the education level of husbands and wives.  His findings with graphs are here.