Monday, August 17, 2015

the NYT's shoddy Upshot column: bullshit and statistics

Another Sunday, another Monday, another idiotic Upshot article in the NYT. Upshot has become the home for the NYT’s consi derable rightwing cheering section, with Cohn, Barro, and Cowen providing the juice.  Barro, the scion of one of the plutocracy’s big defenders at the University of Chicago,  Robert Barro, has settled into the role of “reasonable conservative” that the NYT editors just love love love – it’s the David Brooks gig. Although, to be fair, Barro sometimes is worth reading – which I don’t think one can ever say about David Brooks. This Sunday, though, Nate Cohn was up at the bat to tell us two things: Bernie Sanders is a mere pimple on the vast system so ably managed by our elites – his surge is just exaggerated because, as Cohn puts it in the incomparable jazz style preferred by the Times:

“Mr. Sanders has become the favorite of one of the Democratic Party’s mostimportant factions: the overwhelmingly white, progressive left. These voters are plentiful in the well-educated, more secular enclaves where journalists roam. This voting support is enough for him to compete in Iowa; New Hampshire and elsewhere in New England; the Northwest; and many Western caucuses. But it is not a viable electoral coalition in a Democratic Party that is far more moderate and diverse than his supporters seem to recognize.”

So those supporters are in for a surprise, because, of course, they have never read a newspaper or a magazine telling them that the moderate center is the largest voting bloc in the country. Of course, newspapers and magazines are in the habit of presenting this as an apriori truth, instead of like going to any independent source that empirically checks the statement. Rather, they sometimes turn around in their desks and ask their neighbor, that white guy, usually, who is pulling down more than 250 thou a year – are you a moderate or centrist?

After tut tutting away the Sanders campaign, Cohn then sticks his thumb in his mouth and reflects with all the bogosity of a crooked statistician about Hilary Clinton’s favorability ratings. Here’s our poobah:

“Mrs. Clinton may be a primary juggernaut, but she could surely lose to a Republican in November 2016. President Obama’s approval ratings are in the mid-40s, so Mrs. Clinton may not benefit from the party’s incumbency. On paper, the race is more or less a tossup. In such a close contest, it might seem reasonable to argue that Mrs. Clinton’s unfavorable ratings are hugely important.

Now that is a good question, and a good answer might be founded on looking at polls putting Clinton against all of her possible Republican opponents. But, oddly enough, Cohn, who is writing in something called the Upshot, seems sadly unaware that these polls exist. He – like the NYT in general, where article after article tells us that Clinton is mired in scandal and flailing generally – leaves discretely unmentioned that in those polls, which are easily accessible on Real Clear Politics, Clinton beats all her opponents by 3 to 12 percent. RCP amalgamates all the current polls, but it shows those polling results. One can see that the reason Clinton doesn’t do better is that the Fox News polls consistently show Clinton doing 4 to 10 points worse than the rest of the polls. Pull the Fox News polls from the mix, and Clinton is beating all GOP rivals by unheard of numbers – 6 to 7 percentage points.

So much for the standard shoddy Sunday Upshot. Today, we get a retread of the GOPvoters are happier meme, which has been assiduously promoted by the head of theAmerican Enterprise Institute,  ArthurBrooks.  Our purveyor of nonsense thistime is  David Leonardt. Now I will give Leonardt some credit – he is lesssophistical than other Upshot columnists. But he is prone to publish thingsthat require a little critical thinking. The headline today is that Republicanssay they are happier with their marriages. This is, of course, the old ArthurBrooks trick – publish surveys based solely on self-reporting. No sociologist with any credibility believes that what people self-report is a perfect guide to how they really act. In fact, it is easy to show that the very fact of asking about a self-report can lead to changes in the responses one receives. So, of course, you need some other anchors to clarify the meaning of these self-reported responses. In the case of marriage, the anchors are pretty clear. If Party membership was a significant factor in happy marriages, then those states with a dominant party should, pari passu, show lower divorce rates.
It is well known that, in fact, those states that do have boost larger Republican majorities are also states with higher divorce rates. If Leonhardt was not lazy, he would have at least gotten the name Jennifer Glass from his rolodex and called her. She’s a professor at the University of Texas and has published a pretty well publicized article about the subject in the American Sociological Review (with coauthor Philip Levchak). Let me quotean abstract of the thesis from family studies.org:

“Authors Jennifer Glass and Philip Levchak are more nuanced in their own telling of the story, but their findings are provocative. The authors conclude, “The results here show that communities with large concentrations of conservative Protestants actually produce higher divorce rates than others, both because conservative Protestants themselves exhibit higher divorce risk and because individuals in communities dominated by conservative Protestants face higher divorce risks.”
As for exactly how conservative Protestants are increasing divorce risks for themselves and their neighbors, Glass and Levchak point to evidence that conservative Protestants and their communities encourage young people to marry and have children earlier, sometimes before their educations are completed. These early-marrying couples face a double dilemma of learning to live together (and perhaps raise children together) while also struggling to get by in an economy that is increasingly tough on those who don’t finish college. Then, speculating beyond their data, the authors suggest that conservative Protestant norms against premarital sex and abortion (which might encourage earlier marriage and childbearing) and disdain for religiously “mixed” marriages, along with public policies that fail to support quality public education (enacted in communities dominated by conservative Protestants) combine to create a brew which, paradoxically for divorce-disdaining conservative Protestants, undermines stable marriages.”

Notice that the speculation of these researchers consists of inferences from their data, which tell a plausible sociological tale linking the results of conservative social policies with divorce. It could well be wrong, but at least it is not a mere juxtaposition taken from dubious self-reporting statistics and lathered with speculation that has no empirical anchor whatsoever.

I could probably become a more popular blogger just by fisking the generally shoddy upshot column, and call it something like upchuckshot. But then, I do have a life.  


No comments:

Asking

Yesterday, I watched a very sparkly Biden official, who looked like he had just come from the Ken-at-High-School-UN box, answer questions fr...