[discuss] Survey

Avri Doria avri at acm.org
Tue Mar 4 19:12:27 UTC 2014


If you are looking for, and believe in, statistical truths, yes, that is 
true - and all the statistical work I have done, and the course i took 
as a social sciences grad student, tells me so.  But this is survey 
monkey, not statistical social science.

If, however, you consider it anecdotal evidence of a self selected group 
that gives a range of possible ways of looking at something, it can give 
a clue - especially if there is broad range of opinion.

Before this we had only the Montevideo statement which was our Dear 
Leaders' statement that had little vetting with all the folks at the 
bottom.  Now we are adding a few of the bottom voices to the mix.  As 
one of the voices towards the middle of the heap, I am glad to have it 
recorded somewhere.


On 04-Mar-14 18:59, Andrew Sullivan wrote:
> On Tue, Mar 04, 2014 at 06:42:15PM +0000, Avri Doria wrote:
>> I recommend people do it.  And say No and why if that is the way you
>> feel.  By giving us a 3 value answer yes, yes/but and no, they have
>> neutralized their positions.
> I'm sorry, but I could not disagree more strongly.  Junk surveys are
> an effective way to trump up mistakes, confusion, nonsense, and even
> lies as truth.  This survey was either designed by people who don't
> know what they are doing, or else it is an attempt to misrepresent the
> opinions of a group.  Bad surveys are _worse_ than no knowledge at
> all, because they mislead people into thinking that they know
> something.
> And the three-way answer is almost impossible to evaluate and is
> certainly impossible to analyse numerically, because the middle value
> there is effectively uncountable.
> Best regards,
> A

More information about the discuss mailing list