[Phys-l] failure is always an option

[John Denker <jsd@av8n.com>]

On 05/05/2008 06:14 AM, Michael Edmiston wrote:
> ..... The top 5% of high-school 
> students are either used to working or will rise to the occasion.  They know 
> if they don't, there is a long line of students waiting to take their 
> places.

This got me thinking along a tangent.  I wondered which colleges had the
highest drop-out rates.  I wondered how high the flunk-out rate was at
the "hard" schools compared to the more "nurturing" schools.  I had some 
hypotheses I wanted to check.  (Some of my guesses were qualitatively
correct, but most were wildly wrong.  Data that conflicts with my guesses
always piques my interest.)

It's easy to collect a bunch of numbers from
  http://www.ucan-network.org/

  Also: you can get lots of gory details from a form called "Common Data Set" 
  that most schools fill out.  Many of them post it on their web site.
  I used queries of the form 
      http://www.google.com/search?q="Report+for+the+cohort+of+full-time+first-time+bachelor's"+site:haverford.edu
  Probably cleverer queries are possible, but I didn't need to pursue it.

The first number I looked at is "4-year persistence".  This number has as 
its denominator the number of students who enrolled in a particular year, 
seeking a four-year degree.  The numerator is the subset of those 
students who actually completed the program in four years (or less).  This 
is a little hard to interpret, because AFAICT it treats transferring-out 
the same as dropping-out ... but it is what it is.  I'm not sure what it 
means, but presumably it means "something".

I collected numbers for:
 Arizona State
 Bates
 Caltech
 Cornell
 Harvard
 Haverford
 MIT
 Monmouth
 Ripon
 Rochester
 Stanford
 UC Santa Cruz
 UMASS (Amherst)
 University of North Carolina, Charlotte
 Wellesley

I tried to look at schools with a "party school" reputation or not,
state-run or not, with lots of graduate research or not, highly selective 
or not, large size or not, specialized in science+engineering or not.

I emphasize that I don't really know what these numbers /mean/ in any
deep sense.  And by some accounts, the people who collect such numbers
don't always know what they mean, either.  An overview of the situation
can be found here:
  http://www.compact.org/advancedtoolkit/pdf/framing_compact.pdf

OTOH as the saying goes, imperfect data is better than no data.  And
as a general rule, this is what sets research apart from the rest of
science and engineering:  research involves collecting lots of data 
/before/ anybody know what it means.

  On the third hand, a couple of years ago I looked at some _high school_
  graduation data, and decided it had been cooked to the point of being
  worse than useless.  I hope the college data is more honest.

So, before you look at the numbers, would you like to guess?
 -- What are typical 4-year persistence numbers?
 -- What is the overall range, from lowest to highest?
 -- Which schools are highest?
 -- Which are lowest?
 -- Where does your school fit in?









A few collected numbers:
  http://www.av8n.com/physics/cds-combined.html