Testing consensus using Roman voting

Testing consensus using Roman voting

Here’s part 2 of The Right (Way) To Vote series – posts on voting at events. Feel free to read Part 1, on hand and stand voting.

How should groups make decisions? Common answers to this question include “by voting” or “by discussion and then the boss gets to decide.” But what if we want a decision method that provides consensus, or as near consensus as we can get?

An individual making a decision may agonize over it. But when more than one person is involved, the decision-making process can turn into an argument. So groups need a way to test their agreement, discuss concerns, and arrive at a decision that all can support.

Forcing a yes/no vote about a proposal hides the degree of support for the resulting vote. Sometimes the level of enthusiasm for a decision is tepid, or a minority of those voting are adamantly opposed. Serious problems can surface later when it’s time for implementation.

A better approach is to publicly discover the degree of consensus about a proposal. If doubts or opposition are uncovered, addressing them is possible before a final decision is adopted.

Testing consensus using Roman voting

My favorite solution for gauging the strength of consensus is Roman voting, as described by Esther Derby in Self-facilitation Skills for Teams by Esther Derby:

The Romans indicated their will in the gladiator’s arena with a thumbs-up or a thumbs-down. A modern modification of Roman voting helps groups arrive at a decision.

Thumbs-up = “I support this proposal.”

Thumbs-sideways = “I’ll go along with the will of the group.”

Thumbs-down = “I do not support this proposal and wish to speak.”

If all thumbs are down, eliminate the option. On a mixed vote, listen to what the thumbs-down people have to say, and recheck agreement. Be cautious about choosing an option if the majority are thumbs-sideways: This option has only lukewarm support.

This technique generates consensus. Consensus doesn’t necessarily mean complete unanimity. Consensus means that everyone must be willing to support the idea, even if it’s not his personal first choice.

Sooner or later, you’ll have a situation where one person withholds support for any option. Manage this situation before it happens. At the start of the consensus process, set a time limit:

“We’ll work really hard to reach consensus until the end of this meeting. If we don’t have agreement by that time, we will turn the decision over to _________, or take a vote, or __________ (a technical expert, coach, manager) will decide.”

Most people don’t hold out to be obstinate; they are responding to a deeply held value or belief. Often the lone holdout will move on, but not at the cost of relinquishing an important belief. Respect the belief, use your fallback decision-making method, and move forward. However, when a group seldom reaches consensus, but instead relies on voting or deferring to authority, it’s a sign there are deeper issues at play.

Silence when someone asks, “Do we all agree?” does not signal consensus! Instead, try testing consensus using Roman voting. Do you test consensus on decisions made at your meetings?

Photo attribution: Flickr user vegaseddie

15 thoughts on “Testing consensus using Roman voting

  1. Any ideas on translating this to larger groups? Might take a while to count the thumbs, but I like the idea of having a “third” option in voting. I guess one possibility would be to use polling software with a three stage ranking…

    1. Good question, Brandt! Yes, it’s hard to see thumb positions for a large group. Instead, use card voting. Give each attendee three colored cards (say green [agree], yellow [go along, with reservations], and red [opposed]) and have then hold up the relevant card to vote. Much easier to tally. I’ll be including card voting as a participation technique in my new book.

    2. A nice visual way is to hand out colored stickies and having everyone ‘vote’ by putting the stick on a whiteboard or chart. It creates a very visual look. You can then ask for people to express some of their concerns.

      For really large groups, you can even break people up by their card color and have them do a group brainstorm (method of your choice) separately to come up with their ‘most important’ concerns that they would like to address to make the resulting discussion more productive and pointed.

      1. Thanks Nicholas. Both the techniques you describe—dot voting and what I’d call affinity grouping—are described in my new book out in June, in a long section on “participative voting”. I also include “Idea Swap”, which is a great way to get anonymous opinions on contentious or sensitive topics. I cover the detailed use of these useful techniques there so I won’t go into a lot of detail in this comment. Suffice it to say that Roman voting is best used in a group where it’s safe to to publically express opinions, there has been significant discussion on an issue, and it’s time for a quick method to discover whether a broad concensus exists or not.

        1. OO awesome, I didn’t know that I’d been using dot-voting all along. I’ll keep an eye out for your book, sounds like exactly the kind of thing I’d like to read.

    3. One method I have seen is to divide large groups into groups of 4 people and make a vote. Then get one person from each group into new groups of 4 and make a new vote. Continue until there is one group left. Works even better if you are brainstorming ideas, so that you have multiple outcomes, instead 3 choices like the Roman Vote.

      1. What you’re describing, Piotr, is not testing consensus but old-fashioned yes/no voting, which can be done in one step by traditional hand-voting. Yes, small-group work can be useful for idea generation, but that’s not what Roman voting is for.

  2. The only problem with this method, is that some people are hesitant to speak up in a public forum and that tends to bias a vote. At scale it creates an an environment for group-think. Fostering openness is not enough to counteract this tendency – some people are just naturally reserved, and others are more poignant/intimidating.

    1. The problem you’re describing, Nicholas, is not the fault of the voting method but of the process before the vote. If people don’t feel safe to express their views openly, public voting is not going to work. Possible solutions in order of preference:
      1) do the groundwork for people to feel safe about displaying their vote
      2) use a semi-public voting method, like dot-voting
      3) use anonymous voting (e.g. clickers or Idea Swap)

      1. I agree with you that the process needs to be modified.

        I don’t think #1 is a viable solution. and option #2 blurs one and three together.

        I address my perspective on #3 first, and then later on talk about issues surrounding #1 (and partially 2).

        I’d like to point out that I think specifically number #3 is quite effective, however #3 goes against the nature of the public voting method strictly called for by roman voting.

        I do think the bias is a direct ‘fault’ of the public voting method. I’d argue that this effect of the voting is intentional and basically by design. It was used to enforce a sort of group shame dynamic used intentionally to enforce a form of groupthink.

        Groupthink has benefits when you are working as part of a team whose lives depend on each other such as in the midst of a battle or as part of a strategic maneuver that requires all parties to follow through.

        Using the voting tactic directly – with thumbs up visibility occurring in the open, assumes that people are all coming to the voting table with the same mental fortitude. This dynamic does not help a group come to the most creative solution – simply one that they can all agree on.

        In the roman military one could assume that all the leaders had been vetted as part of their training regime – they were encouraged to be forceful in their attitude and would fit right at home among the harsh criticism in a board room or within the ranks of Amazon. There is a lot of be said about a praetoriate who were biased toward action.

        This brings up option #1: Even if everyone trusts everyone else and feels safe, there are always going to be natural variances among a group that will bias decisions. In the normal world, you can’t necessarily tell if the person sitting across from you was raised by a hurtful parent (Narcissistic or Borderline parents are good at training their kids to be meek). Or victim of other social force (bullied?) that has made them inherently weaker and less willing to vote their conscious. This takes Option #1 off the table if you are using it in a diverse group.

        Requiring everyone in your decision making circle have a fully working ‘spine’ is going to limit the ability of the group to truly tap the diversity of creative thinking that you really want.

        Using Roman voting to discover group consensus can backfire in these cases.

        There are some interesting parallels among group-decision making for patient care that have come up in recent studies. I recommend taking a look at some of the articles here: https://scholar.google.com/scholar?q=related:gC1K4VTVnBoJ:scholar.google.com/&hl=en&as_sdt=0,46 – specifically the concerns outlined in “Systematic biases in group decision-making: implications for patient safety”

        I also had an interesting back and forth with @buffer on twitter: https://twitter.com/ultimape/status/568213715230834688
        https://twitter.com/ultimape/status/568218069212008448 – their managment theory seems to strike a nice balance between fast decision making, creative input, and group concensus.

        1. Somehow, Nicholas, I didn’t see your long thoughtful reply until now. I’m sorry I overlooked it.

          It’s important that Roman voting is only used after there’s been “sufficient” (always a judgement call) discussion. Given this, I think that Roman voting is most useful when there’s been enough discussion that it’s clear there’s some minority unease with a majority-supported decision. When this occurs, it’s helpful to clarify the distinction between those who are prepared to go along with the majority viewpoint and those that are still opposed and want to discuss more . A Roman vote under these circumstances simply provides more information — it doesn’t require minority members to change their minds, but just to state where they currently stand.

          Yes, there may be group members who are meeker than others, more likely to go along with whatever the majority wants. I don’t know any way to account for this; anonymous voting won’t help because, if there’s been enough discussion, it will be already clear what the majority wants to do, and the meek person will go along with it. (If you are wondering if it will make a difference, you can vote anonymously first and see whether there is indeed a significance majority preference. if not, either there hasn’t been enough discussion or the decision is truly contentious and consensus under current conditions will be unlikely.)

          Finally, I think it’s important to bear in mind that, most of the time, the value of attempting to reach “consensus” is in the process of getting somewhere, rather than the end result. Informed, even though reluctant consent to a course of action that some don’t personally agree with is a core facilitation goal in a world where true unanimous agreement on almost anything is rare to non-existent.

        1. Jason, I agree that anonymous voting is a great tool to discover whether there is, in fact, majority agreement on a course of action. Roman voting then becomes useful as I describe in my reply to Nicholas below.

  3. Two comments:
    1) I have spent many hours observing preschoolers acting as a group. What I have seen is 1-2 members within the group take control, 1-2 leave the group (sometimes in tears) and the rest fire off ideas/directions at a rate no one can follow. This is not collaboration, this is directional chaos. It does work but there is a lot of wasted energy.
    2) I would like to see the results of the experiment if it were given to a group of adults that score high in spacial ability (none of those mentioned are known for this). There you might see a combination of trial and error and scientific method practiced.

Leave a Reply

Your email address will not be published.