A friend pointed me to this monstrosity. As an MIT grad, I’d like to think that Technology Review could do better.
To elaborate a bit: A one-paragraph blurb would be fine to me, you can report that someone ran some regressions on the GSS and came up with an amusing hypothesis. That’s enough, then move on to the real technology news of robots playing ping-pong or whatever. I’m not saying to suppress this sort of thing, just place it in the appropriate context. It is what it is. If you want to write a full article on it, fine, but then talk to someone who studies the subject area (you’re Technology Review, you can get these people on the phone!) and move the ball forward a bit. Otherwise why bother at all?
> As an MIT grad, I’d like to think that Technology Review could do better.
So would I but I’m not holding my breath. They’ve dedicated a significant fraction of their pages to (the printed equivalent of) clickbait and hype for as long as I can remember – at least since they went glossy about 20 years ago.
Chris:
Clickbait is fine. My problem is that they’re taking the analysis so seriously. I’d think they could either (a) do the article in one paragraph and get their clickbait right there, or (b) write the longer article in a more interesting way by talking with a scholar of religion.
In this case I assume the problem is that the reporter doesn’t know any better. That’s understandable too—statistics is hard and confused a lot of people—but I’d think it’s the editor’s job to explain to the reporter how to write a science story, especially if you’re Technology Review. You’re supposed to get multiple perspectives and to talk with experts.
It looks to me in the graph that the real increase in slope occurs after 1995. I note Fox News started in 1996 and has increased its ratings in lockstep with the % of people giving up religion.
:)
Clearly, the authors saw your post of a few days ago in which you questioned whether you were becoming too negative, and realized that your moment of weakness was the perfect time to strike!
Diet Coke was introduced in 1982. I blame the proliferation of diet drinks.
Uh-oh, I have to revise my Diet Coke thesis: ESPN started in 1978. Must be the brain rotting effects of cable TV: we sit at home flipping endlessly flipping channels instead of acting churchly.
Maybe it’s watching all those athletes thank God that causes religious skepticism.
Really, if God took an interest, Little Sisters of the Poor would win the national championship every year.
I think it is a dose issue – 4 years of diet coke sounds about the right lag …
“Using the Internet can destroy your faith” in science writers to stop reporting on one guy’s silly regressions purporting to give a “causal explanation”.
The make-believe way the writer pushes this as a credible explanation surviving criticism is most annoying of all.
Don’t they teach “ecological fallacy” at MIT?
http://xkcd.com/1235/
I wonder if butter production in Bangladesh might explain religious affiliation which in turn “causes” internet prevalence to ‘unexplainably’ jump up.
http://shookrun.com/documents/stupidmining.pdf
Confidence intervals and prediction intervals be damned, I say it’s sheep population in the US and Bangladesh that explains it.
“Correlation does provide evidence in favor of causation, especially when we can eliminate alternative explanations or have reason to…”. Gelman, you’re being far too nice with this. Let the inner insult comic come out. Come to the dark side.
Any bets that the analysis presenting this dreck were all independent 2 factor tests? A five note could win you twenty. Takers?
I need to step away from my browser for a bit. When I come back I predict my rabbi will have left me an email stating how I’ve found my religion again and would I please anti up with my building fund donation….
Oops that was me… we need a subscriber login!
There’s a link to the original paper on arXiv. It was all logistic regression.
On a related note, the results are all presented in tables rather than graphs. One assumes that if Dr. Gelman had not been too tired of things like this, he would have pointed this out as a minor flaw in his blog-based criticism.
What I really wish is that I’d thought of the title, “02139 is a joke.” Maybe I’ll be able to use it for some future post.
Gotta get up get up get get down 02139 is a joke it’s your town
Sometimes, I think that the habit of talking about sciences as “hard” or “soft” imbues people in the harder sciences with the idea that they can venture into softer territory with no special need for preparation or domain-knowledge (case in point: econophysics).
Downey’s hypothesis basically points to cognitive mechanisms, but it fails to look at actual beliefs. Americans still believe in God, although there is a small increase in those who believe in a more diffuse God instead of a personal God – mirroring what we see in the rest of the West. As such, a Putnamesque thesis of people preferring to bowl alone seems all the more fitting.
I’m curious … I’m mostly studying trends of secularization and desecularization in Europe but I have skimmed through studies about the phenomena in the U.S., too. I thought there was an actual increase of Atheists or at least Non-believers in the U.S. and at the same time the amount of Evangelicals is increasing. For Europe we can see clear patterns of an increase in Non-believers over the last decades, this is especially true for Western Europe, so I’m not sure what you mean, when you say that the U.S.-trend is “mirroring what we see in the rest of the West”. Could you refer me to a source stating that the change in the religious sphere in the U.S. is mostly (only?) a change from personal to more diffuse notions of god?
Depends on what you mean by non-believers, which is the point. In a lot of countries – for example Scandinavia – beliefs haven’t really changed that much during the last 30 years. But affiliation has, which illustrates that it depends on what you mean by “non-believer”
As regards the individualisation thesis, think of some of Bellah’s work or his term “Sheilaism”. It’s in no way the only process at work, but it’s a component. Look at the GSS; the change in the view on God is almost entirely driven by more people opting for a view of “Some higher powers” instead of the traditional personal image.
My point is: Americans are becoming less churched, but they’re also less active in clubs, PTAs etcetera. Thus, the disaffiliation cannot be seen as a sign of declining religious belief, as Downey’s hypothesis appears, partly because beliefs are still present and partly because it mirrors the situation in other social organisations.
http://www.smbc-comics.com/index.php?id=3129
This reminds me of one of my favorite examples of causality, i.e., the fact that the decrease in the use of Microsoft Internet explorer has reduced the murder rate in the US:
http://gizmodo.com/5977989/internet-explorer-vs-murder-rate-will-be-your-favorite-chart-today/all
Damn! And I consulted for Microsoft. Does that make me a killer?
But he put income, and the other usual confounders in the model! So, as he explains, either A causes B, B causes A or some other stuff is causing both
I think MIT Review got the date on the article wrong.
The published date was April 4.
The correct date is April 1.
Jeyendran:
I hate to tell you this but I think these guys are dead serious. See here for example. If Allen Downey is running a hoax on us, I gotta admire him for playing it perfectly straight.
Pingback: “MIT Built a Theranos for Plants” « Statistical Modeling, Causal Inference, and Social Science