Abusing the Numbers

Glenn Reynolds — probably inadvertently — puts a little spin on the topic of priestly sex abuse as he passes it along to his readers:

Bottom line: It's huge: "even though the report entirely discards incidents involving a further 3,300 priests who had died, and only deals with incidents in which a victim of abuse has come forward, the number still represents over 4 percent of all priests who served in that period."

At the very least, it would seem a significant oversight for Reynolds not to include a consideration that Patrick Belton, whom he quotes, mentions in passing: that 4% represents accusations, not confirmed incidents. Closer inspection shows, however, that while the CNN report from which Belton gets his information certainly shows that sex abuse was a problem, Belton misstated the perspective, leading Prof. Reynolds to convey some outright errors.

For one thing, that 3,300 number isn't priests; it's accusations. And more importantly, they apparently are included in the 4% figure. Here are the relevant paragraphs:

The survey, to be released February 27, found that children made more than 11,000 allegations of sexual abuse by priests. The 4,450 accused priests represent about 4 percent of the 110,000 priests who served during the 52 years covered by the study. ...

More than half of the accused priests had only one allegation against them. Nearly 25 percent, or 1,112 priests, had two or three allegations, and almost 13 percent, or 578 priests, had four to nine allegations, according to the draft report. Nearly 3 percent, or 133 of the priests, had 10 or more allegations.

The report said that 6,700 of the 11,000 allegations were investigated and substantiated, and another 1,000 were unsubstantiated. The remaining 3,300 were not investigated because the priests involved had died by the time the allegation was made.

Objectively speaking, I imagine that the substantiation rate for the deceased priests would be lower than the average (since there's that much less disincentive to make a false accusation), but for this post, let's assume that the 87% of accusations that held up in the 7,700 cases that were investigated is valid for the total. That makes a total 9,570 substantiated accusations.

Unfortunately, we don't know how many priests are exonerated by those 1,430 unsubstantiated claims. We can figure out that 59%, or 2,627, of the 4,450 priests were only accused once. If all of the unsubstantiated claims were made against this group, that would leave the total number of priests facing substantiated charges at 3,020, or 2.7% of priests who served over the 52-year period.

Obviously, if all of the unsubstantiated claims were made against priests with other, substantiated, accusations against them, the number would remain at 4%. But that's really not very likely. I'd guess that one-accusation priests would have a higher rate of false claims than the average, but let's not assume that. Instead, let's just apply that 87% substantiation rate directly to the number of priests. That leads to 3,872, or 3.5% of the 110,000 serving priests.

Of course, Reynolds, following Belton, conveyed that 3,300 priests weren't investigated, which would have left 1,150 priests, or 1% of all priests serving. Furthermore, Reynolds didn't mention that we're dealing with "accusations," so the number that fits the circumstances he conveys to his readers (cutting out the 13% of unsubstantiated claims) would drop another 150 priests out of the total, leaving 1,000, or 0.9% of the pool of priests. As in:

It's huge: "even though the report entirely discards incidents involving a further 3,300 priests who had died, and only deals with incidents in which a victim of abuse has come forward, the number still represents 0.9 percent of all priests who served in that period.

Any abusing priests are too many, of course, but for purposes of "perspective," we do well not to over report the problem by 444%. Now that's huge.