[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Just a single Question for the Candidates



On Fri, Mar 05, 2004 at 06:26:44PM -0600, Manoj Srivastava wrote:
> On Fri, 5 Mar 2004 19:58:03 +0000, Andrew Suffield <asuffield@debian.org> said: 
> 
> > On Fri, Mar 05, 2004 at 01:16:43PM -0500, David Nusinow wrote:
> >> On Fri, Mar 05, 2004 at 03:35:03PM +0000, Andrew Suffield wrote:
> >> > On Fri, Mar 05, 2004 at 08:21:08AM -0600, Manoj Srivastava wrote:
> >> > > 	You have an alternate theory explaining the low incidence of
> >> > >  women in male dominated activities like Debian, free software coding,
> >> > >  coding in general, and CS overall?
> >> >
> >> > Sunspots. It's at least as convincing.
> >>
> >> Way to completely ignore the problem, as well as testimonials by
> >> those involved. What a productive attitude.
> 
> > The plural of "anecdote" is not "data".
> 
> 	Yes, very clever. And also very silly. When collated in large
>  numbers, anecdotes _do_ become data -- ask any psychologist or
>  sociologist.

No, I refuse to accept this. Psychology and sociology are fuzzy
"sciences" for the most part, where very little is proven. That does
not mean that the standards for proof should be lowered, it means that
their conclusions should be treated with the usual skepticism and not
as things which have been conclusively proven.

If somebody were to demonstrate that the majority of people with
orange hair liked to eat babies, then it might be reasonable to
allocate more resources to watch them more closely. It would not be
reasonable to assume that because a given person had orange hair, they
liked to eat babies. Most things that come out of sociology and
psychology take this form - they can give you "probably", or "N% of
this group will do X", but they can't usually give you "true in this
particular case". This is the difference between proof and
circumstancial evidence.

>  And there have indeed been documented studies of the
>  barriers women face breaking into male dominated institutions and
>  workplaces -- and debian certainly qualifies as the former.

That doesn't prove anything. It's not even particularly
convincing. "Debian is like another system where this happened,
therefore it will behave in the same way, because most other ones
did". That just indicates there is a plausible argument with a
not-insignificant probability of being accurate, it does not
intrinsically indicate that the argument is accurate.

Correlation across a large number of systems does *not* demonstrate
that the same thing will happen in any individual system. What about
all the systems where it didn't happen?

-- 
  .''`.  ** Debian GNU/Linux ** | Andrew Suffield
 : :' :  http://www.debian.org/ |
 `. `'                          |
   `-             -><-          |

Attachment: signature.asc
Description: Digital signature


Reply to: