Abilene Paradox

When organizations blunder into the AbileneParadox, they take actions in contradiction to what they really want to do and therefore defeat the very purposes of what they try to accomplish. It is a symptom of the inability to manage agreement - not the inability to manage conflict. Groups that suffer from the AbileneParadox display a number of characteristics: This AbileneParadox is described in TheAbileneParadox and other meditations on management (ISBN 0787902772 ) by Jerry B. Harvey.

A sunny afternoon, a family playing cards on a terrace. One of them thinks they should move - not that he cares, on the contrary, but he thinks the others want to - so he proposes a trip to Abilene. No fun, hot, bad food. Back home one of them admits that he had preferred to stay home. Everybody would have liked that, only they did not admit to it when it was still time to enjoy the afternoon.

-- MartineDevos

Could someone please bring back (or rewrite) the more-detailed version of the story that explains the name?


Here's a re-write of a sort:

It is 104 degrees and the sun is hammering Coleman, Texas with its merciless heat. Two couples sit on the porch of a house sipping lemonade and watch a fan with a bent blade pointlessly stirring the hot air to bring them no relief. The younger couple have invited her parents to visit, now it's so hot that playing dominoes no longer takes their minds from their discomfort. Breaking the rhythm of their misery, Dad coughs into the stoic, contemplative silence and suggests driving to Abilene to get lunch.

Like automata they drag themselves into the ancient pickup and drive the 53 scorched miles to Abilene without the benefit of the air-conditioning system that broke years ago. Lunch is not fun. The food is poor, the beer warm and the service is lifeless. So they return silent, exhausted by their excursion.

Back on the grumpy porch as afternoon heat gives way to sweltering, airless night they discover that nobody actually wanted to go to Abilene; everybody thought that everyone else did...

... and somewhere in the distance, a wild coyote calls ...


Applications to Voting

TacticalVoting can sometimes produce that sort of result. E.g., your true preference order is Nader,Gore,Bush, but you think most other people will vote for for Gore or Bush, so you decide to vote for Gore on the grounds that a vote for Nader would be "wasted". -- DaveHarris
To solve this problem, PreferentialVoting should be used where everyone explicitly communicates their ordered list of preferences and resolution of the votes is done as follows: suppose you have N positions to be filled by candidates in a riding, then first-choice votes are assigned to the candidates until they acquire 1/N votes. Any spillover votes, which are unnecessary and would be wasted normally, fall back to their second choice and then their third choice and so on. If this does not resolve the election then the candidate with the least first-choice votes is eliminated and his votes fall to their second choice. Complications can be made to this voting scheme where instead of arbitrarily deciding which votes for a candidate are spillover votes, one looks at the second-choice on all the votes for that candidate and preserves this proportion in the assignment of used versus spillover votes.

Example, suppose a candidate is elected by 50 votes and Green received 100 votes, then we can either take out 50 votes from the 100 at random or we can look at all the second-choices on those 100 votes, notice that 70 second-choice votes were cast for Blue and 30 for Red and take out 35 2nd Blue votes and 15 2nd Red votes.

Suppose there is only one position and everyone votes for A, B or C as a first choice and for D as a second choice. Then the above resolution scheme would eliminate D from the running entirely even though that is the most sensible choice. This problem is unsolvable and it is known to be impossible to create a voting process with all the qualities that are generally agreed upon to be good for a voting process to possess. This is known as ArrowsTheorem.

Preferential Voting is used in Australia and information on its use around the world can be gleaned from a web search for the term. -- RichardKulisz


The paradox of ArrowsTheorem applies all voting systems, including AcceptanceVoting (also called ApprovalVoting).


To solve this problem the voters should exchange information on their priorities and then form voting blocks; equivalence classes of voters where sameness is "having the same priorities" (within some delta). Then the voting blocks' managements assign votes to members (voters) in each class so that the overall class' voting percentages (how many for A, for B, for C) model the priorities of voters in that voting block. -- PhillipHankins


Oftentimes, at least in a small group, this is completely illusory. This past summer, the group I was with then decided it was time for a reorg. Two different proposals were put forth, and we were asked to vote on them. I saw neither as being beneficial, because it seemed the group's primary problem was that most of the people hated supporting their software. I made and voted for a third choice. Of course, I was the only person who did vote for it (later, privately, several other people admitted that they'd have voted for it except for the AbileneParadox), and the only people who wanted the second plan were in the group I was in (reasoning - hey, this is a reorg to do almost exactly what we're doing now. Admittedly, the other team isn't anywhere close, but this would at best only bring them up to our speed.) I was then given the option to changing my vote to choice B, so that I could be united with the rest of my team in our loss.

Most of the people who were against it were against it because they were afraid they would be placed where they didn't want to be, in support, when the whole point was to get those of us who actually liked support to take over that work from them. I think this is another type of paradox, but I'm not sure about a name for it.

-- EdGrimm
Is PreferentialVoting the same as AlternativeVoting?? They seem to be related to the SingleTransferableVote, if only distantly.

The problem seems to be that people don't vote for the candidate they most want overall, but for the candidate they most want in the set of candidates who they think are likely to win; this is the ancient problem of TacticalVoting.

Personally, I favour AcceptanceVoting. Here, voters simply tick the boxes of the candidates who they would accept, and the candidate with the most votes wins. This eliminates TacticalVoting and the AbileneParadox. A more refined version allows people to give each candidate a score (eg out of 10), and then gives victory to the candidate with the highest score.

-- TomAnderson

AcceptanceVoting does not eliminate TacticalVoting. For example, say you have two options, and the real preferences are that six people find the first option acceptable and five people find the second option acceptable. If two people who think both options are acceptable but prefer the second option vote only for the second, then the second option wins the vote (4 to 5). -- Martin Shobe

Applications to Other Ways of Group Decision Making

One algorithm I've seen for choosing a restaurant in a group goes: walk about until you find somewhere you all agree is good enough, then carry on looking until you find somewhere better (or you're all exhausted) pretending that you can always go back to the first place. I've seen it work, although I don't know why. It may be something to do with taking some of the heat out of the decision, or setting an initial benchmark. Of course, you have to be in a place with more than 1 place to eat...

-- SteveFreeman
A personal decision-making algorithm I use when faced with two alternatives is to choose one, then see how I feel about it. If I don't like it, I take the other alternative. It's only now I've seen this page I realize I may have been using my own personal way to avoid a sort of minor AbileneParadox. -- SteveHolden

I HaveThisPattern but I find it's better to use a coin or a die to represent the choice. I assign a value to each face of the coin (or to each number on the die) and then roll (or flip). If I don't like the answer, I roll again. In some instances however, this practice defeats itself when I just keep on rolling (then I've got to find another way to make my decision).

Yeah. When faced with a decision, flip a coin. Say it comes up heads. If you decide to go for "best of three", then obviously, you wanted tails. Choice made. Flipping a coin doesn't so much make the decision for you as make you think about the options more clearly.
moved from GroupIdiocy

Everyone's stupid about something. Me, I like to bite off more than I can chew and stay up too late at night. You - well, I know what you're stupid about, but I won't embarrass you by revealing it in public. The point is that, together, we're stupider than separately. If we can get enough stupes like us into a room, we'll become so idiotic we won't be able to dress and feed ourselves.

All development groups expand until they become idiotic. If you're in a pleasant, productive, functional group now, just wait around a little while.
Why is agreement and reaching it equated with the probability of success being higher?


In the small, RomanEvaluation can be effective to surfacing ambivalence about a proposal. But RomanEvaluation doesn't scale easily past small groups. -- DaveSmith
We've been using the video - Abilene Paradox - in our company for year. You may enjoy sharing it with your colleagues. We just discovered the original video version is no longer available, but purchased the Abilene Paradox 2 from Trainer's Toolchest. -- Lorene Hanser www.trainerstoolchest.com or http://www.trainerstoolchest.com/show_product.php?idnum=573
See also: GroupThink
CategoryPsychology CategoryManagement

EditText of this page (last edited February 15, 2009) or FindPage with title or text search