^ My desired ratio of type one to type two errors is 1:50
for crimes not including murder, major violence, major
fraud.  For those crimes, a jury must ponder the evidence.
AGSC ^

On 2006-Aug 24, at 05:43hrs AM, Wouter van Ooijen wrote:

>> This is SO wrong!
>>
>> http://www.thenewspaper.com/news/12/1296.asp
>
> 1.    It's almost certain that there was a drug crime involved.

That depends on your value of 'almost certain'. I would describe it as
'not proven beyond reasonable doubt'.

The real question here is how you want to balance between type one
(guilty, not convicted) and type two (not guilty, convicted) errors of
the system. If you want to reduce the occurence of one type of error you
will (when all other parameters are kept equal) increase the occurence
of the other type.

In my country I have the bad feeling that pressure to decrease type one
errors, increasing workload, and budget reductions are pushing type two
errors up, way above what I feel reasonable.

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: www.voti.nl/hvu


--  
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AGSC  Augustus Gustavius Salvatore Calabrese
4337 Raleigh Street
Denver, CO
720 222 1309     303 908 7716 cell
adding " spam2006 " bypasses my spam blocker.  Please place in the  
text or at the END of the subject line.
( i am hard to reach by phone )
All ideas, text, drawings and audio , that are originated by me,  and  
included with this signature are to be deemed to be released to the  
public domain as of the date of this communication .  AGSC
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^


-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist