[EM] Extremely simple voting for committee [Was Re: PR vs. Geographic Representation]

[Gervase Lam]

> Date: Wed, 28 Jan 2004 10:20:09 -0500
> From: Adam Tarr
> Subject: RE: PR vs. Geographic Representation [WAS: RE: [EM] Bill
>   Lewis, never re-district]

> I also suggest you check out proportional approval voting (PAV).  Here's
> the initial thread about it:
>
> http://groups.yahoo.com/group/election-methods-list/message/6367
>
> and here's some commentary I had before:
>
> http://groups.yahoo.com/group/election-methods-list/message/8744

I've read these threads a few times in the past and had particular 
interest in doing so again yesterday, together with related posts that 
Yahoo thought were outside the threads.

As a result of the phasing in of legislation, the small-to-medium sized 
company I work for needs to show that the committee that represents the 
employees has been ratified by the employees.  The committee, which 
already exists, is and must be made up of 1 high ranked manager and 8 
others (which I assume means less senior than the manager, I THINK).

In order to do this, employees working for the Human Resources department 
have decided to issue a vote on this.  I think this vote is on whether to 
accept the committee as it currently stands with its current members or 
not.  The reason why I say I think here is because most if not everbody 
could not access the computerised ballot (the corporate software is Lotus 
Notes, if anyone really wants to know).

Now, this form of voting is obviously not the best way to determine who 
should be in the committee.  Though I would willingly accept the committee 
as is, I think some form of Approval Voting here would be good.

Though the ballot is computerised, it is really just an extremely basic 
card file database.  Therefore, anything much above Approval is definitely 
out.  I would say Borda, Cumulative Voting and even Majority Choice 
Approval are too complicated.  PAV and even Sequential PAV (SPAV) are also 
out of the question because of their reliance on quotas/weightings.

Sure, the database could be programmed to do sophisticated stuff.  
However, this means that (1) Human Resources need to get a handle on what 
is going on, which I think would be too much for them (i.e. keep it simple 
silly!) and (2) the System Administration guys who administer the database 
would need to program it and also, more importantly, test it.

The System Administration guys and Human Resources have much more 
important things to deal with.  The same goes for all of the employees of 
the company!

So the question is, considering these limitations, can anybody suggest any 
ideas?  What follows is a few of the ideas I thought of.

One of the most of obvious things to do is to have two ballots.  One would 
just be normal Approval voting for the manager position.  The other ballot 
would be just plain Approval voting open only to those below the high 
ranking manager in seniority.  The top 8 vote getters from this would make 
up the remainder of the committee.

Another option, which I think is not as good the more I think about it, is 
to have just one Approval ballot.  The manager position would be taken 
by the manager with the most votes.  The remainder of the committee would 
be made of the remaining top 8 vote getters.

This has the advantage that if the employees feel that the managers are 
good, then there could be several managers in the committee.  On the other 
hand, the employees may not, which is a disadvantage.

Another disadvantage with this is that with there is only one ballot, the 
priority of the race would be on getting the right manager.  This would be 
at the expense of putting the "correct" people into the other 8 places.

The final option I thought of is for there to be two ballots, as mentioned 
previously.  The first one would just be normal Approval voting for the 
manager.  The second one would be for the remainder of the committee.  

However, in the second ballot, you are only allowed to vote for a maximum 
of 4 (i.e. half of the 8 places) candidates.  The idea is to compensate 
for the fact that the weighting of ballots do not decrease as soon as a 
set of ballots make a candidate fill a place on the committee, as in PAV 
and SPAV.  If your vote fills half of the 8 places, then "other ballots" 
will determine who will fill in the remaining half of the 8 places.

There is also an obvious disadvantage that even a person with no knowledge 
of voting methods can spot.  Only allowing a maximum of 4 votes is not 
consistent with the Approval vote for the manager.  Therefore, shouldn't 
the vote for the manager only allow a maximum of 1 vote (i.e. half of the 
1 place available)?  Therefore, I think of the three options I've thought 
of, I think the first is the best.

Is what I have mentioned above the only best ways available of doing this? 
 Or could any improvements be made to any of the above?  I must admit, I 
don't think so.

In the interest of keeping this post short, I'll probably post later with 
regards to top-N Approval strategy for this, which I need to read up on, 
and probably giving zero-info guidelines to the voters on how to vote, 
without using the words cardinal ratings, utility or synonyms of those 
words.  So, unless anybody wants to make any comments on these two 
issues...

Thanks,
Gervase.