Part 2—Big newspaper tries and fails: We do love our phony statistics!
Let’s put that a slightly different way. Sometimes, if it weren’t for the phony statistics, we’d have no statistics at all!
This Sunday, we offered these bromides to the analysts after a very rough week. Consider what happened on Saturday morning, when we let them enjoy their weekly political cartoons, courtesy of the Washington Post.
Right at the top of the Post’s op-ed page sat the latest four-panel effort by the Boston Globe’s hilarious Dan Wasserman.
At the Boston Globe site, the hilarious effort bears this headline: “Woman on $20 bill?”
Could a woman be pictured on the twenty? Wasserman ends with a thoughtful observation: You could put a woman on the $20 bill. “But it would only be worth $15.40!”
Hilariously, Wasserman had worked with a treasured statistic. According to this treasured statistic, women are paid 77 cents on the dollar, compared to men, for doing the same or equal work.
Can we talk? No one who works in the field thinks that’s a valid statistic. But we pseudo-liberals love the claim. We just love to repeat it.
Good old Wasserman! He applied the bogus statistic, then ran his cartoon in the Globe. The Washington Post liked it so much they put it at the top of their weekly political cartoons, where our young analysts saw it.
What gives our nation's bogus statistics so much juice? We’ll offer two explanations:
For obvious reasons, nitwits within the various tribes love to build their tribal claims around embellished statistics. Meanwhile, “journalists” at our leading newspapers can’t work with statistics at all.
Our “journalists” can’t handle statistics! Consider the terrible, horrible news report which appeared in Wednesday’s New York Times.
The analysts were hopeful as they read the hard-copy headline:
“Longtime Nursing Pay Gap Hasn’t Budged, Study Says”
Interesting, they loudly exclaimed. They pictured themselves getting the dope on some aspect of the gender pay gap.
Momentarily, they forgot! They forgot what frequently happens when reporters at the Times try to deal with the latest study concerning some important aspect of American life!
In this case, it happened again! Reporter Catherine Saint Louis and her unnamed editor produced a thoroughly bungled analysis of a new study.
It would be hard to be more clueless than Saint Louis and her unnamed editor were. But as we’ll be noting all week, this is what happens when the Times attempts to report key statistics.
Uh-oh! By the time they’d read just two paragraphs, the analysts were worried. Already, Saint Louis and her editor had employed a suspiciously fuzzy term:
SAINT LOUIS (3/25/15): Longtime Nursing Pay Gap Hasn’t Budged, Study Says“In similar positions?” What did that formulation mean? Already, the analysts thought they saw doom approaching.
Male nurses make $5,100 more on average per year than female colleagues in similar positions, researchers reported on Tuesday.
The new analysis, which included data on more than 290,000 registered nurses, also found that the pay gap had not narrowed within workplace settings and specialties from 1988 to 2013. The new study is the first to have measured gender disparities in pay among nurses over time.
Let’s understand—everyone knows that women, on average, earn smaller annual incomes than men in virtually every field. The famous “77 cents on the dollar” statistic—it’s already outdated, but still in wide use—involved a comparison of annual incomes for all full-time workers, defined as people working 35 hours per week or more.
Among those two groups, women were earning 77 percent as much as men, on average, over the course of a year. But that was before the statistics were adjusted for such factors as total hours worked per week (on average, men worked more hours), years of seniority and type of position.
When such adjustments were made, the difference in pay was much smaller. Every expert understands this fact, whatever their political orientation. You just don’t hear it much on cable! Or in other types of cartoons!
As she started her report, Saint Louis reported that male nurses “in similar positions” earn more on an annual basis. Instant problems:
She presented a dollar amount without converting it to a percentage. She didn’t specify the statistical adjustments which may or may not have been made.
Concerning statistical adjustments, was some part of that pay differential due to different hours of work per week? Way far down in paragraphs 9-11, she finally offered this:
SAINT LOUIS: The study did not address reasons underpinning the persistent gap. There could be several reasons, Dr. Muench said: Men may be better negotiators, for instance, or perhaps women more often leave the work force to raise children. Women may have a tougher time getting promoted, she said.The study did not address reasons underpinning the gap, Saint Louis now wrote. And not only that:
“A workplace may offer a bit more to the men in order to diversify,” said Diana Mason, a professor of nursing at Hunter College of The City University of New York and former editor of The American Journal of Nursing.
Still, it is possible that women earn less because of a “lingering bias that a man is more of an expert because he’s a man,” she said.
“Women more often leave the work force to raise children?” That suggested that the study was comparing the annual incomes of men and women with different levels of seniority.
“Women may have a tougher time getting promoted?” That suggested that the study was comparing annual income of men and women who were actually holding different types of positions.
In the end, the journalistic incompetence here is hard to believe. Let’s consider those two basic problems:
Who needs percentages: The gender wage gap is normally discussed in terms of percentages. All over these United States, we the people have had that “77 percent” statistic drummed into our heads. That said, Saint Louis never converted the difference in income among nurses into a percentage.
That $5,100 difference in income sounds like a lot—and it is! That said, it doesn’t come anywhere near the statistical difference commonly cited on a percentage basis.
According to Wasserman’s bogus statistic, women are paid 23 percent less than men. But based on income data which appear in the study, that $5,100 seems to represent something like seven percent less income for women in this field. Though we can’t be entirely sure about that, due to the Times’ weak reporting.
Saint Louis did a long report. Given the conventional way this topic is discussed, she and her unnamed editor should have thrown in a percentage.
Who needs adjustments: That $5,100 difference in income is still a large amount. To what extent does it represent less pay for the same work?
Amazingly, Saint Louis and her editor never really attempt to settle this question. At first, her use of the term “in similar positions” suggested to many readers (you can see them in comments) that she was talking about amounts of pay “for the same or equal work.”
Later, though, she seemed to say that the study didn’t make such adjustments. Here’s where this gets really sad:
In comments, many commenters said they had clicked to the study. Most of the study lies behind a substantial pay wall. But in one portion which is visible, the study seems to suggest that it did adjust for some standard factors at some point.
Or something! Who can tell?
JAMA STUDY (3/24/15): Using ordinary least-squares regression and employment information in the NSSRN, we assessed how much of the annual salary differences could be accounted for by demographic factors, work hours, experience, work setting, clinical specialty, job position, survey year, state of residence, and other factors...The researchers say they assessed “how much of the annual salary differences” could be accounted for by such factors as “work hours, experience [and] job position.”
Does that mean that the $5,100 was the raw difference in average income? Does that mean that adjustments were later made on the basis of such factors as “hours worked?”
We don’t have the slightest idea. In even a slightly rational world, that’s the sort of thing the New York Times would have explained in its news report!
The gender wage gap has become a battle cry for us on the pseudo-left. We love to repeat our bogus statistic, just as pseudo-conservatives have always loved to advance their own tribe’s bogus claims.
Into the fray stepped the New York Times. The analysts were gnashing their teeth as they struggled with the paper's report.
“Can’t anyone here play this game?” one analyst even asked.
Last Wednesday, this news report drove the analysts wild. The very next day, the youngsters would come to despair over the way the Times handled a striking statistic about another high-profile topic—the rate of police shootings by race.
How many shootings is just about right? We will admit that we were struck by what the New York Times said.
Tomorrow: Shootings by race. Still coming:
Doctored statistics about math achievement. Also, those peculiar dueling statistics concerning campus rape! With brief side trips to health care spending! Also, Mitt Romney’s tax rate!
“Can anyone here play this game?” We’d call the answer surprising but obvious.