when NPR gets unlistenably stupid:

During elections, arguably during pledge drive season and, apparently, after school shootings.

I consider myself a pretty liberal left-leaning NPR-friendly sortof guy (describing what I actually am involves a lot more -isms), but I find myself turning off NPR in disgust a lot more since sandy hook. This story, for example, was so mindblowingly dumb I had to turn off the radio, because I was afraid I might plow my car into a telephone pole:

Yang is from China. She says that in college there, she studied math, and then suddenly — totally without prompting — I find myself in another conversation about possibilities and probabilities. Yang, it turns out, specialized in statistics, and since the shooting has been thinking a lot about possibilities and probabilities, reconsidering her original feelings about them.

Yang tells me that she had always assumed that she was safe because the chance of a shooting happening to her specifically was very small. But since the shooting she's been focused on this one rule of statistics she learned in college, which she calls the "large number certainty theorem."

"If the base is big enough," she explains, "even though the probability is small, things will happen with certainty."

By Yang's reckoning, this is how the large number certainty theorem applies.

...

"So, you know, mathematically, something somewhere will happen with certainty," she says.

And so though Yang previously depended on the idea that school shootings were so rare they would probably happen to someone else, the shooting has taught her that "we should not wait until it actually happens to us to take action."

Yang has decided to get more involved with fighting for gun control. This, to her, seems like the logical thing to do.

I .. I just .. What do you say to this? First, I love the hilariously awkward and blatant appeal to authority in the way that they present her as some sort of statistics expert because of a "rule of statistics she learned in college". See? It's a theorem! That sounds very sciencey! You can't argue with FACTS like that!

A fun mental exercise is to substitute literally anything into this line of thought:

Yang tells me that she had always assumed that she was safe because the chance of slipping on a banana peel and splitting her skull open specifically was very small. But since some other guy slipped on a banana peel and split his skull open she's been focused on this one rule of statistics she learned in college, which she calls the "large number certainty theorem."

...

"So, you know, mathematically, something somewhere will happen with certainty," she says.

And so though Yang previously depended on the idea that slipping on banana peels open was so rare it would probably happen to someone else, the dude that slipped on a banana peel has taught her that "we should not wait until it actually happens to us to take action."

Yang has decided to get more involved with fighting for banana control. This, to her, seems like the logical thing to do.

SCIENCED!!!

  • Pasthour

    Chris,  as a scientist, I was initially attracted to the substance of the story.  For similar reasons you mention, I became more interested in the “large number certainty theorem”.  Given my remote history of taking no less than 8 courses in stats on the way to a Ph.D., I did not recall this specific theorem and gave it a second thought tonight. Yang was apparently remembering the “Law of Averages” or “Law of Large Numbers for Discrete Random Variables”, specifically the “Weak Law of Averages”; not the Strong Law of Averages.  “‘So, you know, mathematically, something somewhere will happen with certainty,’ she says.”  Taken literally, this statement means one can mathematically prove occurance of a random event in the future.  Ok that is relatively sound logic, albeit grossly generalized and without specification of the variable.  This type of statement is characteristic of individuals who think in very abstract concepts, but what underlys the generation of the thought is likely what the reporter was attempting to get at using a very clever approach (trying to pry at normal anxiety responses through the backdoor of statistics). 
    The Law of Large Numbers for Discrete Random Variables does not directly address the comment “So, you know, mathematically, something somewhere will happen with certainty,”.   This proof merely demonstrates that given enough correctly repeated trials of a SINGLE event, eventually the distribution of scores will demonstrate consistency and symmetry about the mean or average score respresenting that random variable.  This is a far cry from what the reporter implied in her commentary and probably what Yang ment in her statement as quoted above.  Unfortunately,  we have media negatively influencing the public almost continuously.  This sells ads!  Best with your Blog.
    Scientist from NC

  • Jon

    I’m reminded of this essay by Bruce Schneier http://www.schneier.com/essay-401.html where he argues that when we witness some sort of rare event that is none the less horrific, we’re hardwired to give it much more weight in our mental model of probabilities. One example is that post a movie theatre shooting, we now think movie theatres are more dangerous than the car ride to the theatre which, he states, is safer than it has ever been.

    He ends the essay with what I found to be an interesting statement: “But wear a seat belt all the same.”

    So while the article you quote is indeed pretty lame, there’s something to be said for preparing for a rare, yet inevitable, event.