I liked Tom L’s take-down of Bobo’s anti-data screed so much that I decided to disregard my analyst’s warnings and read the Bobo piece in question.
Like David Brooks, I am sick of hearing about big data and am deeply suspicious of over-reliance on data. Yes, data is all we have to measure the world, in some sense, but the presence of complicated confounding factors means that it’s very difficult to connect causal dots, generally speaking.
But Brooks’ example is stupid: a CEO decides not to pull his company out of Italy because he thinks doing so would make people trust the company less. Bobo implies that he’s doing this to get into heaven or make Burke happy or whatever, but the guy could have very well have decided that losing trust has some monetary value and that when you throw that in, it’s worth it to stay in Italy. If his analysts had told him that staying in Italy would destroy his company, I bet he would have pulled out. He used data, presumably, to make the decision that staying in Italy would not destroy his company. (Or maybe he doesn’t care, he just wants a big bonus…I don’t know but I bet he used data to make the decision in any case.)
Bobo always wants to have his cake and eat it too. He wants to cite sociological studies that use data to show we should live our lives in a more conservative way. But then he wants to say numbers, shmumbers when a study uses data to argue something he doesn’t agree with. I’d argue that’s the problem with neoconservatives in general, that the Straussian tradition of esoteric readings that says “sure Plato said that but he doesn’t mean it” (when Plato seems to have said something they don’t like) leads naturally to a tradition of finding “philosophical” objections to the use of data when it points the other way, and embracing it as the best kind of hard-nosed Burkeanism when it points your own way.
But here’s what gets me:
Your brain is pretty bad at math (quick, what’s the square root of 437),
437 is just a little less than 441 which is 21-squared, so I immediately estimated that the square root of 437 was 20.9 (which turns out to be accurate within 5 one-thousandths). Couldn’t he have picked a better example? The human brain *is* very bad at estimating square roots, but only when the numbers are reasonably large.
End of rant.
The Red Pen
What’s the square root of 69?
8-something.
(That never gets old)
Forum Transmitted Disease
The CEO data points have been winnowed to two: “how large is my bonus and how much do I get if I leave?”
This is why we have an economy that’s making the old Soviet system look pretty good. At least they had full employment and made things.
ThatLeftTurnInABQ
There, fixed. That wasn’t so hard, now, was it? The man just needs a competent editor, that’s all. For want of a nail the shoe was lost…
Comrade Jake
I’m no fan of Brooks, but in his defense, it is clear in context that what he is saying is that our brains are bad at math compared to modern computers. Putting aside the rest of his argument, that basic concept seems fairly sound to me.
BGinCHI
David Brooks:
James Gary
Your brain is pretty bad at math (quick, what’s the square root of 437)
AAAARGGGGHHHH. Metaphor FAIL. I mean, I don’t have DougJ’s “familiarity with the integers” (nor his gift at picking song-lyric titles) but I could still–by using successive approximation–find the square root of 437 in a minute or two. Not being able to compute it instantly is just about the stupidest possible illustration of the “analysis/intuition” dichotomy that I can imagine.
Cris (without an H)
Suggesting that having difficulty doing a quick calculation means that “your brain is pretty bad at math” shows that he has no idea what “math” even means.
The fact that your brain can follow a logical chain of applied rules and algorithms to arrive at a conclusion that isn’t immediately intuitive demonstrates that your brain is fucking great at math.
Comrade Jake
Incidentally, code page 437 is the character set of the original IBM PC.
Alex
As Nelson Muntz famously said, “It’s like the square root of a million. No one will ever know.”
c u n d gulag
I find it amazing that in 800 words, twice a week, Bobo always proves that Bobo’s right about things – no matter how much data there is disproving it beyond even a shadow of a shadow of a doubt.
No wonder he’s teaching a course called “Humility.”
jl
But Doug mathy-math egghead Doug GaltJ couldn’t get the exact square root of 437, could he?
So, Dave Brooks, humbility master is smarter than Doug GaltJ afterall!
Ha ha.
Edit: probably can’t even do the square roots of 19 and 23…
Doug Galt
@Cris (without an H):
That’s a very good point too.
ricky
Do worm farms enhance or endanger square roots?
ThatLeftTurnInABQ
@c u n d gulag:
Actually he’s teaching a course called “The Death of Irony: Suicide or Murder? An interpretive re-enactment”. The fact that not only is the audience deceived with regard to the real title of the course, but that neither the university administration nor the instructor have any idea either, merely adds to the postmodern deliciousness of it all. Of couse they might have gotten a clue if they’d paid attention to what Chekhov famously said about introducing a loaded gun in the first act, but then that would spoil the joke a bit, wouldn’t it?
errg
@Alex:
ah, you were too fast for me…
the funny thing there, is that if you even know a little bit about math, a million happens to be one of the few large numbers that you can instantly say the square root of…
schrodinger's cat
@Doug Galt: Isn’t it sad that Punditubbies like Bobo, Sully and McMegan can’t do grade school arithmetic. Like calculating percentages and taking square roots. A 10 year old probably has better math skills than those 3. But then they turn around and pontificate about economics and statistics.
Baud
Doug, I think the problem is that you see math as a rigorous and principled discipline, whereas Brooks sees it, as so many things, as an allegory.
ranchandsyrup
Data sets and all sorts of facts are irrelevant in these hypertribal days. How a fact can be used is more important that what it actually is. The filibuster is the most important tool ever today for the GOP. If they ever escape from being the tyrannical minority and become the majority, the filibuster will become the worst thing ever foisted upon this nation.
Bill Arnold
You know, you’re not the target audience.
Betcha fewer than 1/10000 in the general Brooks-reading population can do that quick accurate estimate.
Re the Italy example, half the calculation was an intuitive quantification of the value of trust. The other half was probably pretty shaky too, involving prediction.
(I haven’t been able to shed the intuition that DB at least occasionally reads posts here and maybe even reads comments.)
Arclite
Whoa! Check out the big brain on Doug!
Anthony
Quick, say a number that’s literally impossible to say in a finite amount of time!
hitchhiker
Oh fer fuck’s sake. The first sentence is all you need:
I so do not care what comes after that.
Wag
Facts are simple and facts are straight,
Facts are lazy and facts are late
Facts just twist the truth around
Facts are nothing but a big mistake.
Omnes Omnibus
Quick do a textual analysis of the Queen Mab speech. That’s hard too. I suppose our brains are bad at literature.
Bobo is an idiot.
RSA
Oh, for God’s sake.
You know what I hate? Essays holding forth about some topic that the writer knows very little about. This is David Brooks’s oeuvre, but still… You can guess that Brooks has never actually talked to someone who works with big data. He does say that “big data [is] a great tool,” though data is just data–it’s raw material for inference. But he implies throughout the article that people are making decisions purely based on what their inherently limited analysis of big data tells them to do, and that values are being neglected. No one does this except in Brooks’s imagination.
befuggled
@RSA:
I wouldn’t be surprised if that bank exec Bobo had dinner with did.
SiubhanDuinne
@hitchhiker:
I’m wondering, did the bank executive (the large bank executive!) surreptitiously place his hand on Bobo’s thigh under cover of the damask tablecloth?
RSA
@befuggled:
Well, I take the view that the bank exec has to make a decision in a given situation to trust the analysis of the data or not. Unless he’s completely delegated an entire category of decisions to some algorithm, in which case he’s not really behaving like a human being (not even a bank executive version of a human being).
So, maybe, but I’m still skeptical.
cat
Ohh, haha stop it, thats so funny. I always feeling like your trolling me, but I’m going to go for it anyways.
We, Humans, hardly use data at all to analyze our problems. We aren’t even near approaching a sensible use of data as a whole. Sure you could trot out some example where data was misinterpreted, big whoop. I can trot out examples of cutting edge use of data teasing out answers and solutions we couldn’t have done in our lifetimes.
How I judge the if we use data is based on several, still anecdotal, things.
1) Most companies don’t have advanced analysis teams and software solutions. If its in Excel they don’t use data to guide their decisions. The number of major failures have been tracked back to excel is legion.
2) Any company that just sits down and figures out how to measures their output and their customers satisfaction with their output has a competitive advantage. See six sigma and design for six sigma.
Nylund
That bugs the hell out of me too. It’s especially frustrating with the likes of Brooks, McArdle, etc. who constantly cite academic work based on statistical and econometric methods they don’t understand in the slightest (or why that methodology may be incorrect for that particular problem).
Omnes Omnibus
@Nylund: It happens anytime someone with some expertise in a field has to read a know-it-all amateur’s take on things. The problem isn’t the amateur status, but rather the know-it-all bit.
RSA
@Omnes Omnibus:
It’s definitely this for me. There are commenters on this thread alone who know more about big data and data analysis than Brooks does, and yet he has this enormous megaphone to communicate stuff that may not be horribly wrong (in this case) but is pretty misleading.
AHH onna Droid
The @schrodinger’s cat: The IB program allows you to skip out of math with a bs, 6th grade level arithmetic class called math methods. These writing geniuses can earn advanced degrees without knowing the basic level of math to evaluate public policy. Our chattering class is as ignorant as Doctor Aquinas and proud of the fact.
mainmati
I was raised mostly in the pre-ubiquitous calculator age and learned mathematical strategies in my head that yielded pretty fast results. These were very common in the 1960s. I’m always amazing my high school daughter with using these to quickly answer simple arithmetic examples and I was always basically a dummy in higher math.
SFAW
@jl:
Nice try with the trick question ..
They’re both prime numbers, so they can’t have square roots.
Hell, even Bobo “knows” that.
OmerosPeanut
David Brooks’ brain is pretty bad at sociology. Quick, David, correctly use a single modern sociological theory.
Grammar Overlord
@Comrade Jake: Sadly, I knew this fact. I appreciate that I’m not the only one out there. I’ll bet your brain is good at math, too.
MaxxLange
Computers are in fact terrible at math. This is one of the many surprising things I learned when I studied them. First of all, how many integers are there? And how many can you represent on a computer?
the Conster
Irony locked itself in David Brooks’ bathroom but teh stoopid shot its way in.
cat
@MaxxLange:
CPUs you find in desktops are terrible at math. The IEEE floating point standard is terrible at math.
Computers are perfect at math if you use the proper solutions. They may not be as fast as we wish they were, but they are still many orders of magnitude faster then humans.
Nancy
How we read David Brooks at my house.
1. Note the title; perhaps look at a few sentences of the first paragraph.
2. Scroll to comments.
3. Select “Readers’ Picks”
4. Read intellectually stimulating remarks pointing out the flaws in his latest profundity.
Shinobi
As someone who works as a Data Scientist (the new cool sexy name for people who work with Big Data, because apparently Statistician is too academic or something) I agree that Big Data is Problematic.
Just because you have a ton of data and can crunch it doesn’t mean you’re going to get good answers.
No amount of data can save businesses from stupidity. Stupidity comes in many forms, it comes in the form of analysts who don’t think through all aspects of their analysis, it comes in the form of executives who want data to justify the answer they already have.
Interrobang
So in other words, Brooks just rediscovered GIGO? Why does this guy have a job? I mean, I’m admittedly one of these liberal arts majors everyone loves to hate on, and generally about as good at math as a cognitively-impaired sea cucumber, but I do know a few things about data analysis that Brooks apparently missed somewhere along the line.* Holy cow. “Be suspicious of data.” If that isn’t pandering massively to every residual anti-intellectual fibre in his readership’s psyches, and I don’t know what is. Not only that, but it seems like a blatant attempt to cosy up with the teabaggers and wingnuts, all of whom make a virtue out of being anti-data and anti-intellectual.
____________
* I’m not from the US and I had a European-style liberal arts education (the institution where I did my undergrad had pretensions of being Oxford), and I’ve written peer-reviewed educational material on formal logic (which I took in university) and inferential statistics. (A well-taught degree in English also teaches you to look at primary sources, apply multiple types of analysis to see what shakes out, and to evaluate and shred bad arguments.)