created with NetLogo
WHAT IS IT?
This model explores how people form and change their political preferences. It is based upon Condorcet's theorem, which shows that when individuals have more than two choices, the outcome of an election depends upon the rules of election (institutions) rather than on the distribution of preferences. Condorcet showed that in the absence of such institutions, the collective choice may "cycle," or change regularly between possible societal choices. This model illustrates how such cycles may emerge from the decentralized and localized decision-making of individual voters.
HOW IT WORKS
The model simulates a Condorcet problem by generating a population of voters with randomly distributed preferences for three (nominal) policies: blue, green and yellow. It explores the conditions under which Condorcet-style cycling may occur. Rather than imposing an election rule, it simply calculates the social choice as the plurality. The model holds periodic elections, from every ten time periods (with the first election at t = 10) to every 100 time periods (first election at t = 100).
Voters in the model move randomly around the space and may change their preference order. Their decision to re-order their preferences depends upon two factors: (a) the preferences of their immediate neighbors (that is, all voters within a radius of one); and (b) the outcome of the previous election. The user can enact different decision rules for voters (called good-winners?, sore-losers?, and coalitions?) by using the swicthes and sliders in the "Voter Options" control area.
HOW TO USE IT
Click the "Setup" button to generate a population of voters whose preferences and locations are distributed randomly. You can adjust the number of voters using the "number" slider to create a denser or thinner network of voters. Click "go" to have the voters move and change their preferences according to their immediate neighbors.
When the "good-winners?", "sore-losers?", and "coalitions?" switches are all in the OFF position, a voter has no memory of the previous election. A voter will change its preferences with a fixed probability only if its neigbors' preferences are different. The user can change the probability of voters changing their preferences with the "voter-sensitivity" slider. If the slider is set to 0.20, for example, a voter will change its preferences 20 percent of the time (or alternatively, on average about one-fifth of the voters will change their preferences during any given step in the model). If the slider is set to 1.00, voters will evaluate and, if necessary, change their preferences during each step of the model.
The good-winners?, sore-losers? and coalitions? switches activate different strategies for a subset of the voters. Each switch allows the voters to remember the outcome of the previous election and to change (or choose not to) their preferences according to a specific decision rule. When activated, the good-winners? switch tells those voters whose preferences "won" the previous election to be good winners--that is, to be more sensitive to the preferences of their neighbors. Winners will be two to five times more sensitive than losers according to the value set by the "winner-sensitivity" slider.
The sore-losers? switch acts in a similar way, but makes those voters who lost the previous election perfectly insensitive to other voters. That is to say, after losing an election the losing voters will not change their preferences until the next election at the earliest.
The coalitions? switch simulates the possibility of losing voters cooperating. When selected, the switch allows losing voters to change their preference in response to other losing voters but not to the winners. That is to say, losing voters listen to each other (but not to winning voters) and change their preferences accordingly, while winners listen to everyone, both winners and losers.
THINGS TO NOTICE
Watch what happens when none of the Voter Options switches are enacted. Notice how one preference quickly "wins" an election, and how the society locks into this choice.
Watch what happens when you activate the "good-winners?" switch. Set up and run the model several times to see whether or not good winners make the system more or less "stable"--that is, whether or not society converges around one choice or election cycles emerge.
Watch what happens when you activate the "sore-losers?" switch only.
Watch what happens when you activate the "coalitions?" switch only.
THINGS TO TRY
Try running the model with the "coalitions?" and "good-winners?" switches activated at the same time.
Try running the model with none of the switches activated. Then, after four or five elections (40 or 50 on the time counter), activate the "coalitions?" switch (or alternatively the "sore-losers?" switch).
Try setting up the model so that one of the three options is "locked out." Under what conditions does the society settle into a binary social choice?
EXTENDING THE MODEL
Extensions of this model might include:
1. introducing new collective choice rules, other than the plurality rule.
2. introducing non-local networks of voters, so that voters' preferences depend upon voters who are distant in the model.
3. extending loser strategies to be probabilistic rather than deterministic, more strategic (for example, vote for second choice).
NETLOGO FEATURES
None.
RELATED MODELS
See other voting models in the Models Library, including Rudy Ruckers' voting model adapted by Uri Wilensky and the various Prisoners' Dilemmma models in the unverified folder.
CREDITS AND REFERENCES
For information on Condorcet's Theorem and election cycles, see:
James Adams, "Condorcet Efficiency and the Behavioral Model of the Vote," The Journal of Politics, 59, 4 (November 1997): 1252-1263.
David M. Estlund, Jeremy Waldron, Bernard Grofman, and Scott L. Feld, "Democratic Theory and the Public Interest: Condorcet and Rousseau Revisited," The American Political Science Review, 83, 4 (December 1989) 1317-1340.
Kurt Taylor Gaubatz, "Election Cycles and War," The Journal of Conflict Resolution, 35, 1 (June 1991): 212-144.
H. P. Young, "Condorcet's Theory of Voting," The American Political Science Review, 83, 4 (December 1988): 1231-1244.
Model developed by:
David C. Earnest, Ph.D.
Assistant Professor of Political Science
Old Dominion University
BAL 700
Norfolk, VA 23529
dearnest@odu.edu |
