On Sun, 16 May 2004 13:25:07 GMT, Art Kendall <[EMAIL PROTECTED]>
wrote:

> I concur with Rich.
> IQ is a norm-referenced test.  It is restandardized periodically to make 
>   the comparison to the current average.  Over time, if the 
> standardization is not too out of date, roughly half the kids are over 
> 100 in each period. To be in the middle now requires getting more 
> answers correct than it used to.
> 
> I guess it comes back to the original question on data transformation.
> The raw number of items correct doesn't have much meaning at a given 
> time except in comparison to the current cohort of kids.

I'm not sure about that "current cohort" part.  -- Isn't it a 
legitimate, serious puzzle, why the number is improving?  

On the one hand:  Even though the change is probably not
owed to the schools, it is fortuitous information for 
defending the public schools.  Can you imagine the demagogs
if the trends and most of the available data seemed to *support*  
the claims of poor performance of the schools, instead of 
the opposite?

Further:  I have to wonder, Are today's kids really smarter in ways 
that matter?   I suspect that they are.  Now, how do you measure it?

When a TV network put on a "national IQ test" in a 2-hour 
special last year, they concluded that the "professors" in their
studio sample "scored higher" than the "students" -- but the 
professors were older, and were given six or ten points extra
on their scores, just for being older.  Separate age norms.
Half-way towards conceding the question, I thought.
 - I thought that I remembered posting about that test last year, 
shortly after it, but I did not find anything in a quick search.

-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to