Saturday, January 25, 2014

Stanines (or, Statistics is silly)

As an undergrad, I took a statistics class in the math department, and bunch of my friends took statistics for psychology or statistics for engineering.  Whenever we'd try to compare notes, nothing made sense, because everything was labeled differently in the different fields of statistics.  There were some concepts that had several different names (sometimes even within the same field), which was confusing until we realized that we were actually talking about the same things.  Likewise, there were some names that labeled several different concepts, which was even more confusing, because we'd think we were talking about the same thing and then later realize that we were actually talking about different things.  We decided that the average number of names per item is one; the names are just distributed unevenly.  (We were too frustrated to figure out the standard deviation.)

So I had already come to the conclusion that statistics is a silly field.  Very useful, but very silly, especially in the names department.

And then this week in my Assessment class, I learned about "stanines", or "standard nines".  That's right, statisticians made up that name by smushing two words together, or blending, as linguists like to say.  Like "Brangelina" or "spork" or "linner".  (I have a good friend who wrote a rather long paper on the rules for blending in English, so I find the concept fascinating.)  This is hilarious!  Apparently it was first used during WWII by the US Army Air Force.  I can just imagine the statisticians sitting up late at night trying to decide how to divide up the normal curve:

Statistician 1:  Should we divide it into tenths?  That makes sense.
Statistician 2:  Too cliché.
Statistician 3:  Also, if we use an odd number, the mean will be contained within a category.
Statistician 1:  Fine, then let's try.... um, nine?
Statistician 3:  Oh, that's great!  Then all of our scores can be recorded with single digits!  I hate double-digit numbers, don't you guys?
Statisticians 2:  ...I guess....
(scribbling)
Statistician 3:  Hey, if we fudge these edge categories, all the middle ones could be a half of a Standard Deviation.  That's a nice unit!
Statistician 2:  ...Okey.... What should we call it?
Statistician 1:  How about Sta-nine?
Statisticians 2 and 3:  (blank stares)
Statisician 1:  Get it?  Standard nine—stanine?  Because they're standard, and there's nine of them!  Well, they're mostly standard, anyway.  Ignore the edge ones.
Statisticians 2 and 3:  Whatever.

In case you're wondering, a stanine is a unit of a half a standard deviation (except the two at the very edges), and there are nine of them in the normal curve (including the weird edge ones).  According to wikipedia, stanines are only used by the US education system and also the New Zealand education system.  And get this—there's another similar system called the "sten" system.  (I bet you can guess what two words are being blended there.)  The sten system seems unduly complicated to me; maybe I'll stick to stanines if ever I have the choice.

In conclusion:  Statisticians are bad at naming things.

3 comments:

  1. As a mathematician by training, statistics has always been and continues to be the field of mathematics most abused by the world at large, in my opinion. The theories of statistics are logically solid, just as they are in all branches of mathematics, but often researchers misuse techniques or misinterpret the results because they do not fully understand statistical concepts that can be admittedly complicated or confusing. At times, it is a heartbreaking tragedy to watch, being so fond of mathematics; at other times, I find myself saying, "You deserve what you get, statistics! You refuse to yield an absolute answer, and instead leave humanity with only partial certainty! It is no wonder people manipulate you - you lend yourself to it!"

    I try to free myself of this emotional, academic roller coaster by avoiding statistics as much as possible, but it has unfortunately been involved in the majority of my graduate studies (in applied and computational mathematics).

    ReplyDelete
  2. That's good that you are learning some statistics in your assessment class! Yeah, the stanines are pretty silly if you already know standard deviations and z-scores, but a lot of published assessments include stanines, so you should be aware of them anyways.

    ReplyDelete
  3. The silliness and abuse that can happen with statistics reminds me of the recent "fight" between Facebook and Princeton: http://www.cnn.com/2014/01/24/tech/social-media/facebook-princeton-smackdown/

    ReplyDelete

What would you like to say to the Wug Tamer?