Is ISIS’s popularity low? Hard to Tell

The Washington Post reported on a Pew survey of countries with a significant Muslim population, asking whether people there have a positive or negative view of ISIS:

FT_15.11.17_isis_views
Pew survey of countries with a significant Muslim population, spring 2015.

As the graph shows, in every country, ISIS is unpopular, by huge margins (and I find it interesting that according to the same survey, attitudes vary more from country to country and between religions; but I’m not sure that’s important right now).

But there’s low, and then there’s low. I wouldn’t drink water that was “only” 1% arsenic (the EPA limit is a million times smaller than that), and if there were “only” a 1% chance of crashing every time I got on the Beltway, I’d likely be dead within a year. So there’s small and then there’s small.

If the KKK enjoyed the same level of popularity in Alabama as ISIS does in Turkey, above, it wouldn’t be reported as “KKK only has minority support”. It would be reported as “One in 13 Alabamans still supports KKK”.

So I’d like to see some figures for comparison: how popular is ISIS compared to, say, the Lord’s Resistance Army, or Nazis, or the Tutsi army? Maybe they have comparable levels of popularity, and the double-digit favorability numbers are statistical artifacts. And in fact, it seems that the fact that its popularity varies more from country to country, than from religion to religion within the same country, points at this explanation.

But my fear is that ISIS really does enjoy insufficiently-low popularity.

Catholic Church 99 44/100% Pure

BillDo has a post in which he plays down the Catholic priesthood’s image problem:

Catholic League president Bill Donohue comments on the findings of the 2011 Annual Report on priestly sexual abuse that was released by the bishops’ conference; the survey was done by a Georgetown institute:

The headlines should read, “Abuse Problem Near Zero Among Priests,” but that is not what is being reported.

According to the 2011 Official Catholic Directory, there are 40,271 priests in the U.S. The report says there were 23 credible accusations of the sexual abuse of a minor made against priests for incidences last year. Of that number, 9 were deemed credible by law enforcement. Which means that 99.98% of priests nationwide had no such accusation made against them last year. Nowhere is this being reported.

If that’s his standard of purity, then I’m sure Bill would have no problem drinking a glass of 99.98% water and only 0.02% urine, right?

The thing is that very few men in general are child abusers. The question (or one question) is, does the Catholic clergy contain more child abusers than the population at large?

I wasn’t able to quickly find child-abuse statistics for the United States, but I did find the FBI’s Uniform Crime Reporting statistics for violent crime in 2010, which shows an aggregate of 27.8 forcible rapes per 100,000 victims. The FBI defines “forcible rape” as:

The carnal knowledge of a female forcibly and against her will. Rapes by force and attempts or assaults to rape, regardless of the age of the victim, are included. Statutory offenses (no force used—victim under age of consent) are excluded.

So the numbers are not directly comparable: the report Donohue is quoting concerns itself only with sexual abuse of minors, while the FBI’s number covers all rape. The FBI’s 2010 number excludes sexual abuse of males, while BillDo emphasizes that in the report he’s quoting, “almost all the offenses involve homosexuality“. And BillDo calculates the rate per offender while the FBI counts the rate per victim, which means that BillDo’s number tends to undercount priests who abused multiple victims, compared to what the FBI counts.

Having said that, BillDo’s figure of 9 credible accusations and 40,271 priests works out to 22.3 per 100,000, compared to the FBI’s 27.8 per 100,000. So the number of pedophile priests seems to be in the same ballpark as the number of rapists in the US as a whole. That seems pretty bad, especially for a group that presents itself as the guardians of morality.

BillDo also ignores, as usual, that the Catholic church’s problem is not so much one of having rapists in its ranks — any large organization is bound to have some — but of covering up its members’ crimes. The abuse itself can be blamed on individual priests, sure. But the coverup is a problem for the organization.

What’s Three Orders of Magnitude Among Friends?

(Alternate title: “Numbers Mean Things”.)

The increasingly-irrelevant Uncommon Descent blag had a post today, commenting on an article in Science News.

Right now, UD’s post is entitled “Timing of human use of fire pushed back by 300,000 years”, but when it showed up in my RSS reader, it was “Timing of human use of fire pushed back by 300 million years“. This mistake survives in the post’s URL:
http://www.uncommondescent.com/human-evolution/timing-of-human-use-of-fire-pushed-back-by-300-million-years/

From skimming the Science News article, it looks as though a new study found evidence of fire being used one million years ago, pushing back the earliest-known use of fire by 300,000 years. So presumably the previous record-holder was 700,000 years ago.

The author at Uncommon Descent reported the 300,000-year difference as “300 million years”. But hey, what’s a factor of 1000 between friends?

To illustrate, imagine a student in school in 2012, writing a report about, say, e-commerce. At first, she dates the origin of e-commerce to 1994, when Amazon.com was founded. But upon further investigation, she finds an example of a company selling stuff on the Internet in 1987 and revises her report to say that e-commerce is 25 years old, not 18. That’s about the magnitude of what the scientists found.

Now, along comes UD and reports this as “Origin of e-commerce pushed back to 22,000 BC.” That’s the size of their mistake.

It’s easy to make fun of primitive people whose counting system goes “one, two, three, many”. But the truth is, we all do this to some extent. Imagine a newspaper headline that says, “Federal budget increases by $600 billion, including $300 million increase in NASA funding.” Did you think, “holy cow! NASA got half of that extra money!”? If so, I’m talking to you: you’re not counting “one, two, three, many”, but you are counting “ten, hundred, thousand, illion”.

At any rate, I still question the numeracy of whoever wrote that UD headline. If you’re going to spell out “million” in letters, it should trigger a reality-check mechanism in your brain that makes you ask, “Wait a sec. 300 million years ago. That’s the age of dinosaurs or earlier.”

Numbers Mean Things

So I saw this headline in The Washington Post:

UN envoy says $5 billion malaria fight has saved several thousand lives in recent years

My first thought was, “$5 billion divided by, let’s say 5000 people, that comes out to a million bucks per person saved. A noble result, to be sure, but isn’t there a more cost-effective way of achieving the same result?”

Then I read the first paragraph:

UNITED NATIONS — The U.N. chief’s envoy for malaria says a $5 billion campaign has saved several hundred thousand lives in recent years, keeping international efforts on track to virtually end deaths from the mosquito-borne disease by 2015.

(emphasis added) and now the cost per person drops two orders of magnitude, from $1 million to $10,000. Much more reasonable (though it’d still be nice if it were even cheaper).

But I suspect that either the reporter, or someone at AP or WaPo decided that the word “hundred” didn’t change the meaning enough to make it worth taking up valuable headline space. I’m sorry, but it does.

Then again, what’s two orders of magnitude among friends?

(Cross-posted at UMDSI.)

Fact-Checking the BillDo

Recently, BillDo farted the following onto the intertubes:

Moreover, Jenkins wrote that “Out of 100,000 priests active in the U.S. in this half-century, a cadre of just 149 individuals—one priest out of every 750—accounted for over a quarter of all allegations of clergy abuse.” In other words, almost all priests have never had anything to do with sexual molestation.

(italics in the original).

Just for comparison, the Wikipedia page for Crime in Detroit, Michigan, says that the murder rate there was 40.1 per 100,000 people in 2009.

Assuming that each murder was committed by a different person, this means that about one Detroiter out of every 2500 accounted for all of the murder in 2009. In other words, almost all Detroiters are not murderers.

So by BillDo’s reasoning, Detroit does not have a murder problem. Good to know. Presumably if I gave him a glass of water with only one part of arsenic in 750, he’d drink it.