XPost: alt.comp.linux   
   From: bowman@montana.com   
      
   On Tue, 24 Feb 2026 21:14:17 +1100, Daniel70 wrote:   
      
   > In my situation (1983'ish) I was running the Examination Cell for an   
   > Australian Army TaFE Electronics Wing, so Stats were kept on each   
   > question used in End of Subject Exams to determine how HARD each   
   > Question was i.e. how many got each question Correct and then a second   
   > figure compared how many of the Smarties (Top 30%, I think) got it   
   > correct compared to how many of "Dummies" group got the question   
   > correct.   
   >   
   > Those two figures, Overall Correct v Difference Correct, where then used   
   > to determine how DIFFICULT each Exam was .... and, if need by, the Marks   
   > attained could be adjusted.   
   >   
   > STATICS .... made for computers .... or COMPUTERS ... made for statics!!   
      
   When I was in college I worked summers at the New York State Dept. of   
   Education. NY has statewide Regents examinations for most high school   
   subjects and the exams are created from a pool of questions using   
   historical data on the difficulty to get the expected curve. Most of the   
   number crunching was done with Friden SRWs.   
      
   http://computermuseum.informatik.uni-stuttgart.de/dev_en/frd_srw/   
      
   Usually they got the desired results but a couple of times they had to   
   make adjustments after failing half the state's students.   
      
   The SRW did not have divide by zero protection and would diligently try to   
   get a result until you unplugged it. This was the mid-'60s and the   
   department had the usual bureaucratic lag in adopting those new fangled   
   computer things.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)