math 1010 ‘mathematical thought and practice’

28
Math 1010 ‘Mathematical Thought and Practice’ An active learning approach to a liberal arts mathematics course

Upload: vilina

Post on 13-Jan-2016

30 views

Category:

Documents


0 download

DESCRIPTION

Math 1010 ‘Mathematical Thought and Practice’. An active learning approach to a liberal arts mathematics course. Nell Rayburn David Cochener. Department of Mathematics Austin Peay State University Clarksville, Tennessee. Decisions. How do we make choices?. Types of Decisions. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Math 1010 ‘Mathematical Thought and Practice’

Math 1010‘Mathematical Thought and

Practice’An active learning approach to

a liberal arts mathematics course

Page 2: Math 1010 ‘Mathematical Thought and Practice’

Nell RayburnDavid Cochener

Department of Mathematics

Austin Peay State University

Clarksville, Tennessee

Page 3: Math 1010 ‘Mathematical Thought and Practice’

Decisions

How do we make choices?

Page 4: Math 1010 ‘Mathematical Thought and Practice’

Types of Decisions

• Individual—our opinion is our decision.

• Group—Individual opinions are expressed by voting(at least in a democratic society)and some procedure is used to combine these individual preferences for a group decision.

Page 5: Math 1010 ‘Mathematical Thought and Practice’

Two Basic Questions

• What type of election decision procedure should we use to combine individual decisions (preferences) into a group decision?

• How can we be sure that what is decided is really what the group wants?

Page 6: Math 1010 ‘Mathematical Thought and Practice’

Desirable Properties of Majority Rule

(Two Alternative Case)• All voters are treated equally. (Swapping

marked ballots gives no change)

• Both alternatives are treated equally. (If all votes are reversed, so is the winner.)

• If a new election were held and a single voter changed from a vote for the previous loser to the previous winner, then the outcome would be the same as before.

Page 7: Math 1010 ‘Mathematical Thought and Practice’

May’s Theorem

• If the number of voters is odd, and if we are interested only in voting procedures that never result in a tie, then majority rule is the only voting system for two alternatives that satisfies the conditions listed on the previous slide.

Page 8: Math 1010 ‘Mathematical Thought and Practice’

Kenneth O. May (1915-1977)

• Mathematician, Political Activist

• PhD, Univ. of California, 1946

• Mathematics plays a crucial role in social science!

Page 9: Math 1010 ‘Mathematical Thought and Practice’

Ordinal Ballots

Preference List

Page 10: Math 1010 ‘Mathematical Thought and Practice’

Ordinal Ballots

• List your choices in order with the favorite on top and ‘least favorite’ on bottom

Ballots must beComplete (you must rank all candidates)Linear (no ties)Transitive (If you prefer A to B and B to C,

then you must prefer A to C.)

Page 11: Math 1010 ‘Mathematical Thought and Practice’

Preference Schedule

18 12 10 9 4 2

Allen Evert Davis Baker Childs Childs

Baker Childs Evert Davis Evert Davis

Childs Baker Childs Childs Baker Baker

Davis Davis Baker Evert Davis Evert

Evert Allen Allen Allen Allen Allen

Page 12: Math 1010 ‘Mathematical Thought and Practice’

Five Election Decision Methods

• Plurality

• Standard Runoff

• Sequential Runoff

• Borda Count

• Condorcet Winner Criterion

Page 13: Math 1010 ‘Mathematical Thought and Practice’

Plurality

• Whoever has the most votes wins!

• Problems? Remember Jesse Ventura??

Page 14: Math 1010 ‘Mathematical Thought and Practice’

1998 Minnesota Governor’s Race

• Jesse ‘The Body’ Ventura: 38%

• Hubert Humphrey III: 33%

• Norm Coleman: 29%

• The latter two were highly experienced, but somewhat dull compared to Ventura.

Page 15: Math 1010 ‘Mathematical Thought and Practice’

Standard Runoff

• If there is no majority, the two candidates receiving the most votes compete head to head.

Page 16: Math 1010 ‘Mathematical Thought and Practice’

Sequential Runoff

• If no one has a majority, eliminate the candidate(s) with the fewest first place votes, and count again. Continue in this way until someone has a majority.

• Also known as Hare elimination.

Page 17: Math 1010 ‘Mathematical Thought and Practice’

Borda Count

• If there are n candidates, assign n – 1 points to a first place choice, n – 2 points to a second place choice,…, 0 points to a last place choice. Sum the points for each candidate – the one with the most points wins.

• Problems? Borda Count can violate majority rule!

Page 18: Math 1010 ‘Mathematical Thought and Practice’

A surprising result!

• Borda count may violate majority rule!

A has a majority, but B wins Borda count by 21 – 19 over C, with A getting 18 points and D getting 8 points.

6 2 3

A B C

B C D

C D B

D A A

Page 19: Math 1010 ‘Mathematical Thought and Practice’

Condorcet Winner Criterion

• Conduct head – to – head contests between each pair of candidates. If any candidate can beat each of the others, he is a Condorcet candidate. Under this method a Condorcet candidate is declared the winner.

• May not always be decisive! (produce a winner)

Page 20: Math 1010 ‘Mathematical Thought and Practice’

Fairness Criteria• 1. Majority Criterion. If there is a choice that has a majority of the first-

place votes, then that choice should be the winner of the election.• 2. Condorcet Criterion. If there is a choice that is preferred by the voters

over each of the other choices (in a head-to-head matchup), then that choice should be the winner of the election.

• 3. Monotonicity Criterion. If choice X is a winner of an election and, in a re-election, all the changes in the ballots are favorable to X, then X should still be a winner.

• 4. Independence-of-Irrelevant-Alternatives Criterion. If choice X is a winner of an election, and one(or more) of the other choices is disqualified and the ballots recounted, then X should still be a winner. (Also called Binary Independence.)

• IIA can also be stated as: It is impossible for an alternative B to move from non-winner to winner unless at least one voter reverses the order in which he/she had ranked B and the winning alterative.

• 5. Pareto Criterion. If every voter prefers alternative X over alternative Y, then the voting method should rank X above Y.

Page 21: Math 1010 ‘Mathematical Thought and Practice’

Fairness Criteria• **Every method we have studied can violate one or more of

these!• Plurality: violates Condorcet, IIA• Standard Runoff: violates Monotonicity, Condorcet• Hare Elimination(Sequential Runoff): violates Monotonicity,

Condorcet and IIA• Coombs’ Method: violates Monotonicity and Condorcet• Sequential Pairwise Runoff: violates Pareto• Condorcet’s Method: May not even produce a winner!• Borda Count: violates Majority (and hence Condorcet) and IIA• Is there any election decision procedure we could devise which

satisfy these fairness criteria if we have 3 or more candidates and use ordinal ballots to rank the candidates?

Page 22: Math 1010 ‘Mathematical Thought and Practice’

Arrow’s Theorem

The search for the perfect election decision procedure

Page 23: Math 1010 ‘Mathematical Thought and Practice’

The story so far…

• We have studied several election decision procedures designed to produce one or more winners from a slate of 3 or more candidates.

• Each procedure has had some desirable features and some undesirable ones (‘quirks’).

• We’ve even seen that these methods can give different winners using exactly the same set of ordinal ballots!

Page 24: Math 1010 ‘Mathematical Thought and Practice’

Enter Kenneth Arrow

• Arrow, an economist, wanted to find a completely ‘fair’ election decision procedure.

• He began by making a list of a few basic properties that he believed any good election decision procedure should have:

Page 25: Math 1010 ‘Mathematical Thought and Practice’

Arrow’s Properties

• Universality—The decision procedure must be any to process any set of ordinal ballots to produce a winner, and must be able to compare any two alternatives.

• Non-dictatorship (no one voter can determine the outcome)

• Independence-of-Irrelevant-Alternatives Criterion (Binary Independence)

• Pareto Criterion

Page 26: Math 1010 ‘Mathematical Thought and Practice’

IIA or Binary Independence

• It is impossible for an alternative B to move from non-winner to winner unless at least one voter reverses the order in which he/she had ranked B and the winning alternative.

• In other words, whether A or B wins should depend only on how the voters compare A to B, and not on how other alternatives are ranked relative to A or B.

Page 27: Math 1010 ‘Mathematical Thought and Practice’

Pareto Criterion

• If every voter prefers alternative X to alternative Y, then the decision procedure should rank X above Y.

Page 28: Math 1010 ‘Mathematical Thought and Practice’

Asking the Impossible

• In 1951 Arrow published a book Social Values and Individual Choice in which he proved that there does not exist an election procedure which ranks for society 3 or more candidates based on individual preferences and which satisfies the fairness criteria we have listed.