Group Voting – Watch the Bias


Melbourne Street Art 2011

Recently I have been noticing a heavy group biasing effect with internal teams. This seems to be occurring a great deal around any visible voting system such as Dot Voting.

This has been to the extent that people are voting down the real intent and privately desired outcomes in favour of a conservative group safe outcome. Which is usually not what is wanted.

Dot Voting

Currently dot voting is achieved by handing out a nominated number of dots (votes) and these are placed in full view of  group members on the nominated items, with the previous votes visible at the same time.

This type of voting process usually occurs when we have a limited number of selections, like after a brainstorming session, affinity or any decision making process.  The selection at this point is usually just to allow for filtering of ideas that move forward.

As you can see this means that the complete group dynamics of dominate personalities, internal relationships and the like will come to into effect and change or direct outcomes.

For example, if a participant doesn’t like the way a manager voted they may vote the other way just to counter their vote.

This means that we are not going to get a really true indicator of what people are thinking and how they are evaluating the items without running into the groupthink bias.

Not something we really want, sure it can’t be helped, but we can try and minimise the manipulation of the results.

 What We Need

What we need is a process that can just simply help to remove the bias.

Ideally we want to allow people to:

  • Vote quickly
  • Vote without scrutiny
  • Vote in private
  • Vote without reference of the current results

This means that we may have to consider taking the process off the wall.

Even so we still we need the process to be:

  • Logistically easy to do.
  • Easy to implement.
  • Easy to tally results.

Ideally why not just take the post-its for the items concerned off the wall and vote privately.  This would be good, but you know there would be biasing on some people recognising the handwriting and voting accordingly.  Also the major negative would be in  managing the volume of large research discovery processes with hundreds of items.

This leaves us with a separate quickly transcribing a handwritten list showing the relevant items.   However that doesn’t allow people to see the context of the information that makes up a category for instance.  This would also slow the process down as the list is compiled.

During the voting process you basically want to hide the votes and yet also not allow any hint that there is a preference for one item over another.

Alternative System

Now if we considering that we just want to use the resources we have on hand, dots, postits and sharpies, then any process change has to be simple.

Let’s consider this:.

  1. Hand out the dots as normal.
  2. Hand out postits.
  3. Allocated a unique number to each participant.
  4. Each participant writes their unique number on the front of their postits.
  5. Approach the wall, for the first item.
  6. Vote on the back (the sticky side) of the postit with the dots.
  7. Each participant then places the voted postit on the wall, number visible.
  8. No vote, then just place a numbered postit on the wall.
  9. Move onto the next item.

Write Number, Vote, Place on wall.

This way the votes are hidden, but can be amended, as a participant knows which are their votes.  Tallying is easy.    You can’t tell really who voted for what.  A downside maybe the number of extra postits on the wall.

Okay it’s a bit of a process, but not much really.   Now if you suspect you are getting groupthink happening this maybe a way around it.

Tags: , , , , , ,

Looks like there is no conversation here yet, why not start one.