Location of the 5th f2f

WOWG,

There have been some problems regarding the decision to hold the 5th f2f 
in Manchester.  In sum, the web vote clearly showed a consensus for NY as 
the location, however the results were not correctly analyzed due to the 
existence of non-numeric votes.  The chairs believed, when deciding on 
Manchester, that the web vote was split, and believed a Manchester 
decision reflected as much of a consensus of the group as a NY decision. 
While no one disputes the chairs' authority here, the simple fact is that 
they intended to follow the consensus of the group, but had erroneous 
information as to that consensus.

I propose we revisit this decision, however in doing so we should also 
take into account new information (that was not considered in the previous 
vote) regarding the location of the 6th f2f.  I will propose a solution in 
my next message.  The rest of this message contains a recap of the 
decision, evidence showing how the decision was based on erroneous 
information, and evidence that the group CLEARLY EXPRESSED A CONSENSUS FOR 
NY as the location of the 5th f2f.

Recap:  When the Manchester location was proposed, several people pointed 
out that all Europeans would basically vote for Manchester, while all 
North Americans would split their votes between the four US locations, 
creating a skewed view of the group's consensus on Europe vs. US.  Rather 
than voting in two rounds, with the first round simply choosing between 
Europe or the US, a web page was set up which allowed members to express 
their preferences numerically, 1-5, with "1" expressing a strong 
preference and "5" expressing a strong negative preference.  In addition, 
members were allowed to vote with a "*", which was supposed to express a 
"can't live with this choice" vote - even stronger negative than a 5. 
Again, the idea was that someone prefering  e.g. a US location, could vote 
"1" for NY and "2" for Boston and "5" for Manchester, and thus in the 
average this would be reflected as support for both Boston and NY, with a 
slight preference for NY.

When the votes were analysed, the sums of the tally's for each location 
were computed, with the intention that the lowest sums would indicate the 
locations the group preferred.   In these sums, the vote seemed very 
close, 61 for NY and 66 for Manchester.  The chairs concluded there was a 
split, and took a straw poll on August 15th (the height of vacation 
season), which showed a split of those present.  Those present agreed the 
chairs would make a decision.  The chairs, believing the group to be split 
on the issue, decided for Manchester based on new information that the 6th 
f2f will have to be in Boston.

The problem here is that, while the chairs made this decision believing 
the group was split, this belief was based on incorrect analysis of the 
votes.  There were 24 numerical votes for NY, and only 22 for Manchester.  
 Two people did not submit numeric votes for Manchester, resulting in an 
overall lower sum for Manchester.   In fact, these two non-numerical 
negative votes ended up being counted as 0s in the sum, in effect 
completely reversing their sense from strongly negative to strongly 
positive!

I have converted the "." (Lynne Thompson) and "*" (Deb McGuinness) votes 
to 5s, and this resulted in the following sums:

NY: 61
Boston: 70
Pens: 75
Man: 76
Wash: 78 (taking Bernhard's "." vote for Wash as a 3 based on his 
comments)

There really can be no doubt that this result shows a clear consensus of 
the entire group to have the meeting in the eastern US, with NY preferred 
over Boston.  There is no split, the result is not close by any measure.


Dr. Christopher A. Welty, Knowledge Structures Group
IBM Watson Research Center, 19 Skyline Dr.
Hawthorne, NY  10532     USA 
Voice: +1 914.784.7055,  IBM T/L: 863.7055
Fax: +1 914.784.6078, Email: welty@us.ibm.com

Received on Wednesday, 28 August 2002 09:44:30 UTC