Political representationTrench warfareRational voting
Candidate positioningRecap
Mathematical vs. statistical models in socialscience
Andrew GelmanDepartment of Statistics and Department of Political Science
Columbia University
24 Oct 2005
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Themes
I Mathematical models in social science are cool . . .
I But they tend to give qualitative rather than quantitativepredictions
I Statistical modeling as an alternative
I Collaborations with Hayward Alker, Aaron Edlin, NoahKaplan, Gary King, and Jonathan Katz
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Themes
I Mathematical models in social science are cool . . .
I But they tend to give qualitative rather than quantitativepredictions
I Statistical modeling as an alternative
I Collaborations with Hayward Alker, Aaron Edlin, NoahKaplan, Gary King, and Jonathan Katz
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Themes
I Mathematical models in social science are cool . . .
I But they tend to give qualitative rather than quantitativepredictions
I Statistical modeling as an alternative
I Collaborations with Hayward Alker, Aaron Edlin, NoahKaplan, Gary King, and Jonathan Katz
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Themes
I Mathematical models in social science are cool . . .
I But they tend to give qualitative rather than quantitativepredictions
I Statistical modeling as an alternative
I Collaborations with Hayward Alker, Aaron Edlin, NoahKaplan, Gary King, and Jonathan Katz
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Themes
I Mathematical models in social science are cool . . .
I But they tend to give qualitative rather than quantitativepredictions
I Statistical modeling as an alternative
I Collaborations with Hayward Alker, Aaron Edlin, NoahKaplan, Gary King, and Jonathan Katz
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Examples
I Political representation
I Trench warfare
I Rational voting
I Moderation and vote-getting
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Examples
I Political representation
I Trench warfare
I Rational voting
I Moderation and vote-getting
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Examples
I Political representation
I Trench warfare
I Rational voting
I Moderation and vote-getting
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Examples
I Political representation
I Trench warfare
I Rational voting
I Moderation and vote-getting
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Examples
I Political representation
I Trench warfare
I Rational voting
I Moderation and vote-getting
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Part 1: political representation
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I The U.S. is a representative democracy
I The right to vote; # representatives per voter
I Procedures vs. outcomes: what if 90% of the voters get theCongressmember whom they want?
I How close are actual elections?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I The U.S. is a representative democracy
I The right to vote; # representatives per voter
I Procedures vs. outcomes: what if 90% of the voters get theCongressmember whom they want?
I How close are actual elections?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I The U.S. is a representative democracy
I The right to vote; # representatives per voter
I Procedures vs. outcomes: what if 90% of the voters get theCongressmember whom they want?
I How close are actual elections?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I The U.S. is a representative democracy
I The right to vote; # representatives per voter
I Procedures vs. outcomes: what if 90% of the voters get theCongressmember whom they want?
I How close are actual elections?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I The U.S. is a representative democracy
I The right to vote; # representatives per voter
I Procedures vs. outcomes: what if 90% of the voters get theCongressmember whom they want?
I How close are actual elections?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Congressional elections in 1948 and 1988
0.0 0.2 0.4 0.6 0.8 1.0Democratic share of vote for Congress
U.S. Congressional districts in 1948
0.0 0.2 0.4 0.6 0.8 1.0Democratic share of vote for Congress
U.S. Congressional districts in 1988
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Comparing to votes for President
0.0 0.2 0.4 0.6 0.8 1.0
010
3050
Dem share of Congressional vote in 1988
0.0 0.2 0.4 0.6 0.8 1.0
020
4060
Dem share of Congressional vote in 1948
0.0 0.2 0.4 0.6 0.8 1.0
020
4060
80
Dem share of Presidential vote in 1988
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I Equal votes, satisfaction with outcomes, having your votepotentially matter
I Are your political views represented?
I Do your representatives look like you? Data from 1989?Proportion of
Proportion of seats in HouseU.S. population of Representatives
Catholic 0.28 0.27Methodist 0.04 0.14Jewish 0.02 0.07Black 0.12 0.09Female 0.51 0.06Under 25 0.37 0
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I Equal votes, satisfaction with outcomes, having your votepotentially matter
I Are your political views represented?
I Do your representatives look like you? Data from 1989?Proportion of
Proportion of seats in HouseU.S. population of Representatives
Catholic 0.28 0.27Methodist 0.04 0.14Jewish 0.02 0.07Black 0.12 0.09Female 0.51 0.06Under 25 0.37 0
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I Equal votes, satisfaction with outcomes, having your votepotentially matter
I Are your political views represented?
I Do your representatives look like you? Data from 1989?Proportion of
Proportion of seats in HouseU.S. population of Representatives
Catholic 0.28 0.27Methodist 0.04 0.14Jewish 0.02 0.07Black 0.12 0.09Female 0.51 0.06Under 25 0.37 0
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
What does it mean to be “represented”?
I Equal votes, satisfaction with outcomes, having your votepotentially matter
I Are your political views represented?
I Do your representatives look like you? Data from 1989?Proportion of
Proportion of seats in HouseU.S. population of Representatives
Catholic 0.28 0.27Methodist 0.04 0.14Jewish 0.02 0.07Black 0.12 0.09Female 0.51 0.06Under 25 0.37 0
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Seats and votes in a legislature
I Proportional representation in Europe
I No proportional representation in U.S.
I Wasted votes
I Small changes in votes
I Pinball analogy based on vote changes between election years
I No way to mathematically derive the “best” system
I Paradox of voting power and decisive votes
I Paradox of voting for native Australians
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Research problem: unequal representation across the world
I Small states overrepresented in U.S. Senate and electoralcollege
I Small states in U.S. get more than their share of gov’t funding
I Look at other countries: small states/provinces are generallyoverrepresented
I Small states/provinces get more than their share of funds
I Larger consequences?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Research problem: unequal representation across the world
I Small states overrepresented in U.S. Senate and electoralcollege
I Small states in U.S. get more than their share of gov’t funding
I Look at other countries: small states/provinces are generallyoverrepresented
I Small states/provinces get more than their share of funds
I Larger consequences?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Research problem: unequal representation across the world
I Small states overrepresented in U.S. Senate and electoralcollege
I Small states in U.S. get more than their share of gov’t funding
I Look at other countries: small states/provinces are generallyoverrepresented
I Small states/provinces get more than their share of funds
I Larger consequences?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Research problem: unequal representation across the world
I Small states overrepresented in U.S. Senate and electoralcollege
I Small states in U.S. get more than their share of gov’t funding
I Look at other countries: small states/provinces are generallyoverrepresented
I Small states/provinces get more than their share of funds
I Larger consequences?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Research problem: unequal representation across the world
I Small states overrepresented in U.S. Senate and electoralcollege
I Small states in U.S. get more than their share of gov’t funding
I Look at other countries: small states/provinces are generallyoverrepresented
I Small states/provinces get more than their share of funds
I Larger consequences?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Research problem: unequal representation across the world
I Small states overrepresented in U.S. Senate and electoralcollege
I Small states in U.S. get more than their share of gov’t funding
I Look at other countries: small states/provinces are generallyoverrepresented
I Small states/provinces get more than their share of funds
I Larger consequences?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Part 2: trench warfare
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Trench warfare: the live-and-let-live system
I Front-line troops in World War I avoided fighting (Ashworthbook)
I Informal agreements across no-man’s-land
I How to understand this?
I Prisoner’s dilemma
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Trench warfare: the live-and-let-live system
I Front-line troops in World War I avoided fighting (Ashworthbook)
I Informal agreements across no-man’s-land
I How to understand this?
I Prisoner’s dilemma
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Trench warfare: the live-and-let-live system
I Front-line troops in World War I avoided fighting (Ashworthbook)
I Informal agreements across no-man’s-land
I How to understand this?
I Prisoner’s dilemma
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Trench warfare: the live-and-let-live system
I Front-line troops in World War I avoided fighting (Ashworthbook)
I Informal agreements across no-man’s-land
I How to understand this?
I Prisoner’s dilemma
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Trench warfare: the live-and-let-live system
I Front-line troops in World War I avoided fighting (Ashworthbook)
I Informal agreements across no-man’s-land
I How to understand this?
I Prisoner’s dilemma
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Prisoner’s dilemma for trench warfare
I Payoffs in the “game” (Axelrod book)
I No motivation to cooperate in single-play game
I Cooperation in repeated-play game
I Cool mathematical model
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Prisoner’s dilemma for trench warfare
I Payoffs in the “game” (Axelrod book)
I No motivation to cooperate in single-play game
I Cooperation in repeated-play game
I Cool mathematical model
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Prisoner’s dilemma for trench warfare
I Payoffs in the “game” (Axelrod book)
I No motivation to cooperate in single-play game
I Cooperation in repeated-play game
I Cool mathematical model
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Prisoner’s dilemma for trench warfare
I Payoffs in the “game” (Axelrod book)
I No motivation to cooperate in single-play game
I Cooperation in repeated-play game
I Cool mathematical model
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Prisoner’s dilemma for trench warfare
I Payoffs in the “game” (Axelrod book)
I No motivation to cooperate in single-play game
I Cooperation in repeated-play game
I Cool mathematical model
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Refuting the prisoner’s dilemma for trench warfare
I Look more carefully at payoffs
I No motivation to fight! Shooting poses a risk, whether or notthe other side shoots
I Commanders manipulate the “game” to get soldiers to fight
I Hidden assumption of conventional roles of soldiers onopposing sides
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Refuting the prisoner’s dilemma for trench warfare
I Look more carefully at payoffs
I No motivation to fight! Shooting poses a risk, whether or notthe other side shoots
I Commanders manipulate the “game” to get soldiers to fight
I Hidden assumption of conventional roles of soldiers onopposing sides
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Refuting the prisoner’s dilemma for trench warfare
I Look more carefully at payoffs
I No motivation to fight! Shooting poses a risk, whether or notthe other side shoots
I Commanders manipulate the “game” to get soldiers to fight
I Hidden assumption of conventional roles of soldiers onopposing sides
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Refuting the prisoner’s dilemma for trench warfare
I Look more carefully at payoffs
I No motivation to fight! Shooting poses a risk, whether or notthe other side shoots
I Commanders manipulate the “game” to get soldiers to fight
I Hidden assumption of conventional roles of soldiers onopposing sides
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Refuting the prisoner’s dilemma for trench warfare
I Look more carefully at payoffs
I No motivation to fight! Shooting poses a risk, whether or notthe other side shoots
I Commanders manipulate the “game” to get soldiers to fight
I Hidden assumption of conventional roles of soldiers onopposing sides
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Why was the prisoner’s dilemma model appealing?
I “The evolution of cooperation”
I Using game theory to solve the “tragedy of the commons”
I Axelrod’s theory: politically liberal or conservative?
I “The norm of self-interest” (Miller)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Why was the prisoner’s dilemma model appealing?
I “The evolution of cooperation”
I Using game theory to solve the “tragedy of the commons”
I Axelrod’s theory: politically liberal or conservative?
I “The norm of self-interest” (Miller)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Why was the prisoner’s dilemma model appealing?
I “The evolution of cooperation”
I Using game theory to solve the “tragedy of the commons”
I Axelrod’s theory: politically liberal or conservative?
I “The norm of self-interest” (Miller)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Why was the prisoner’s dilemma model appealing?
I “The evolution of cooperation”
I Using game theory to solve the “tragedy of the commons”
I Axelrod’s theory: politically liberal or conservative?
I “The norm of self-interest” (Miller)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Why was the prisoner’s dilemma model appealing?
I “The evolution of cooperation”
I Using game theory to solve the “tragedy of the commons”
I Axelrod’s theory: politically liberal or conservative?
I “The norm of self-interest” (Miller)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Toward the future
I How to defuse future conflicts?
I Axelrod’s logic: set up repeated-play structures to motivatelong-term cooperation
I Alternative strategy: set up immediate gains fromcooperations and watch out for outside agents who coulddisrupt the cooperation
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Toward the future
I How to defuse future conflicts?
I Axelrod’s logic: set up repeated-play structures to motivatelong-term cooperation
I Alternative strategy: set up immediate gains fromcooperations and watch out for outside agents who coulddisrupt the cooperation
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Toward the future
I How to defuse future conflicts?
I Axelrod’s logic: set up repeated-play structures to motivatelong-term cooperation
I Alternative strategy: set up immediate gains fromcooperations and watch out for outside agents who coulddisrupt the cooperation
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Toward the future
I How to defuse future conflicts?
I Axelrod’s logic: set up repeated-play structures to motivatelong-term cooperation
I Alternative strategy: set up immediate gains fromcooperations and watch out for outside agents who coulddisrupt the cooperation
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Part 3: rational voting
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Rational model for voting
I Utility of voting = pB − c :I p = probability that a single vote will be decisiveI B = net benefit from your candidate winningI c = met cost of voting (whether or not your candidate wins)
I Paradox of voting: p is very small, so even for large values ofB, there is no “instrumental” benefit to voting
I In presidential elections, p is about 1 in 10 million
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Rational model for voting
I Utility of voting = pB − c :I p = probability that a single vote will be decisiveI B = net benefit from your candidate winningI c = met cost of voting (whether or not your candidate wins)
I Paradox of voting: p is very small, so even for large values ofB, there is no “instrumental” benefit to voting
I In presidential elections, p is about 1 in 10 million
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Rational model for voting
I Utility of voting = pB − c :I p = probability that a single vote will be decisiveI B = net benefit from your candidate winningI c = met cost of voting (whether or not your candidate wins)
I Paradox of voting: p is very small, so even for large values ofB, there is no “instrumental” benefit to voting
I In presidential elections, p is about 1 in 10 million
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Rational model for voting
I Utility of voting = pB − c :I p = probability that a single vote will be decisiveI B = net benefit from your candidate winningI c = met cost of voting (whether or not your candidate wins)
I Paradox of voting: p is very small, so even for large values ofB, there is no “instrumental” benefit to voting
I In presidential elections, p is about 1 in 10 million
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Rational model for voting
I Utility of voting = pB − c :I p = probability that a single vote will be decisiveI B = net benefit from your candidate winningI c = met cost of voting (whether or not your candidate wins)
I Paradox of voting: p is very small, so even for large values ofB, there is no “instrumental” benefit to voting
I In presidential elections, p is about 1 in 10 million
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Rational model for voting
I Utility of voting = pB − c :I p = probability that a single vote will be decisiveI B = net benefit from your candidate winningI c = met cost of voting (whether or not your candidate wins)
I Paradox of voting: p is very small, so even for large values ofB, there is no “instrumental” benefit to voting
I In presidential elections, p is about 1 in 10 million
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Rational model for voting
I Utility of voting = pB − c :I p = probability that a single vote will be decisiveI B = net benefit from your candidate winningI c = met cost of voting (whether or not your candidate wins)
I Paradox of voting: p is very small, so even for large values ofB, there is no “instrumental” benefit to voting
I In presidential elections, p is about 1 in 10 million
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Possible explanations for voting
I Utility of voting = pB − cI “Benefit” of voting or “civic duty”
I Does not explain higher turnout in close elections and moreimportant elections
I Poor estimation of pI Estimation would have to be really poor for p to be large
enough
I Is voting irrational?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Possible explanations for voting
I Utility of voting = pB − cI “Benefit” of voting or “civic duty”
I Does not explain higher turnout in close elections and moreimportant elections
I Poor estimation of pI Estimation would have to be really poor for p to be large
enough
I Is voting irrational?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Possible explanations for voting
I Utility of voting = pB − cI “Benefit” of voting or “civic duty”
I Does not explain higher turnout in close elections and moreimportant elections
I Poor estimation of pI Estimation would have to be really poor for p to be large
enough
I Is voting irrational?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Possible explanations for voting
I Utility of voting = pB − cI “Benefit” of voting or “civic duty”
I Does not explain higher turnout in close elections and moreimportant elections
I Poor estimation of pI Estimation would have to be really poor for p to be large
enough
I Is voting irrational?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Possible explanations for voting
I Utility of voting = pB − cI “Benefit” of voting or “civic duty”
I Does not explain higher turnout in close elections and moreimportant elections
I Poor estimation of pI Estimation would have to be really poor for p to be large
enough
I Is voting irrational?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Possible explanations for voting
I Utility of voting = pB − cI “Benefit” of voting or “civic duty”
I Does not explain higher turnout in close elections and moreimportant elections
I Poor estimation of pI Estimation would have to be really poor for p to be large
enough
I Is voting irrational?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Possible explanations for voting
I Utility of voting = pB − cI “Benefit” of voting or “civic duty”
I Does not explain higher turnout in close elections and moreimportant elections
I Poor estimation of pI Estimation would have to be really poor for p to be large
enough
I Is voting irrational?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Voting to benefit others
I Utility of voting = pB − cI B = Bself + αNBsoc
I Bself = individual benefit from candidate A winningI Bsoc = (your perception of) avg. benefit of others from
candidate A winningI α (probably less than 1) discounts benefits to othersI N = number of persons affected by the election
I It can now be rational to vote!
I Decoupling rationality from selfishness
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Example: a close election
I Each candidate expected to get between 47% and 53% ofvote
I Vote differential in range ±6%I Pr (your vote is decisive) ≈ 1/(0.12n)
I Suppose the selfish benefit to you is $10,000
I If n = 1 million, then expected selfish benefit is less than 10cents
I Now consider a “social voter”I Suppose n/N = 1/3 and suppose that the benefit to others (as
you perceive it) is $10 eachI The effect of your vote on their expected gain is
$10N/(0.12n) = $250I Voting is like making a $250 charitable contribution
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Supporting evidence for the theory
I Small contributions to national campaigns
I Declining response rates in opinion polls
I Turnout is higher, not lower, in large elections
I Turnout is higher in close elections
I Strategic voting
I Voting on issues without direct instrumental benefits(abortion, All-Star game, Academy awards, . . . )
I Ask people why they vote
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Empirical tests
I Are altruistic people more likely to vote?
I Is turnout higher in U.S. Senate elections in small states?
I Is turnout higher in NYC when there is heavy snow in Buffalo?
I Studying corporate contributions (Ansolabehere)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Empirical tests
I Are altruistic people more likely to vote?
I Is turnout higher in U.S. Senate elections in small states?
I Is turnout higher in NYC when there is heavy snow in Buffalo?
I Studying corporate contributions (Ansolabehere)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Empirical tests
I Are altruistic people more likely to vote?
I Is turnout higher in U.S. Senate elections in small states?
I Is turnout higher in NYC when there is heavy snow in Buffalo?
I Studying corporate contributions (Ansolabehere)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Empirical tests
I Are altruistic people more likely to vote?
I Is turnout higher in U.S. Senate elections in small states?
I Is turnout higher in NYC when there is heavy snow in Buffalo?
I Studying corporate contributions (Ansolabehere)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Empirical tests
I Are altruistic people more likely to vote?
I Is turnout higher in U.S. Senate elections in small states?
I Is turnout higher in NYC when there is heavy snow in Buffalo?
I Studying corporate contributions (Ansolabehere)
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Part 4: candidate positioning
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Candidate positioning
The “median voter theorem” (Hotelling, 1928):
Left-wing Right-wingD RM
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Median voters and Newt Gingrich
I In the 1994 election, the Republicans gained about 50 seats inCongress
I The Democrats who lost were mostlymoderate-to-conservative
I The liberal Democratic congressmembers were reelected
I Democrats should be liberal and be proud?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Median voters and Newt Gingrich
I In the 1994 election, the Republicans gained about 50 seats inCongress
I The Democrats who lost were mostlymoderate-to-conservative
I The liberal Democratic congressmembers were reelected
I Democrats should be liberal and be proud?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Median voters and Newt Gingrich
I In the 1994 election, the Republicans gained about 50 seats inCongress
I The Democrats who lost were mostlymoderate-to-conservative
I The liberal Democratic congressmembers were reelected
I Democrats should be liberal and be proud?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Median voters and Newt Gingrich
I In the 1994 election, the Republicans gained about 50 seats inCongress
I The Democrats who lost were mostlymoderate-to-conservative
I The liberal Democratic congressmembers were reelected
I Democrats should be liberal and be proud?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Median voters and Newt Gingrich
I In the 1994 election, the Republicans gained about 50 seats inCongress
I The Democrats who lost were mostlymoderate-to-conservative
I The liberal Democratic congressmembers were reelected
I Democrats should be liberal and be proud?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Looking at the 1994 election more carefully
x
x
x
x
x
x
x
xx
xx
xx
x
x
x
x
xx
x
x
x
xx
xx
x
x
x
x
x
x
x
x
x
x
x
x x
xx
xx x
xx
x x
x x
xx
x
x
x
x
x
xx
x
x
x
x
x
xx
x
xx
x
x
x x
x
xx
xxxx
x
x
x
x
x
xx
x
x
x
xx
x
x
x
x
x
xx
x
x
x
x
xx
xx
x
xx
x
x
x
x x
x
x
x
x
x
x
x
x
x
xx x
x
x
x
x x
xx
x
x
x
−1.0 −0.8 −0.6 −0.4 −0.2 0.0
0.4
0.6
0.8
Democrats running for reelection in 1994
(liberal) ideology (moderate)
vote
sha
re in
199
4
x
x
x
x
xx
x
xx
x
x
xxx
x
x
xxx
x
x
x
xx x
x
x
x
x
x
x
xx
x
x
x
xx
x
xx
x
x
x
x
x
x
xx
x
x
x
xx
x
x
x
x
x
x
x
x
x
x
x
x
xx
xx
x
x
x
x
x
x
xx
x
xx
x
x
x
xx
x
x
xx
xx
x
x
x
x
x
xx
x
x
x
x
x
xx
x
x
xx
x
xx
xx
x
x
x
x
x
xx
x
x
xx
xx
xx
x
x
xx
xxx
−1.0 −0.8 −0.6 −0.4 −0.2 0.0
−0.
150.
000.
10
(liberal) ideology (moderate)vote
in 1
994
com
pare
d to
exp
ecte
d
o
oo o
o
o
o
o o
o
o
o
o
oo
o
o
oo
oo
o
o
o
ooo
o
oo
o
oo
o
o
o
oo
o
o
oo
o
o
o
oo ooo
o
o
o
o
o
o
o
o ooo
o
o
o
o
o
o
o
o
o
oo
o
o
o
o
o o
o
o
o
o
oo
o
o
o
o
o
o
0.0 0.2 0.4 0.6 0.8 1.0
0.55
0.65
0.75
Republicans running for reelection in 1994
(moderate) ideology (conservative)
vote
sha
re in
199
4
o
o
oo
ooo
o
o
o
o
o
o
oo
oo
o
o oo
o
o
o
o
oo
o
o
o
o
ooo
oo
oo
o
o
o
o
o
o
o
oo ooo
o
o
o
o
o
o
o
oo
oo
o
o
o
o
o
o
o
oo
o
o
o
o
ooo o
o
oo
o
o
o
o
o
oo
oo
0.0 0.2 0.4 0.6 0.8 1.0
−0.
15−
0.05
0.05
(moderate) ideology (conservative)vote
in 1
994
com
pare
d to
exp
ecte
d
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Congressmembers’ ideologies and median voters
0.1 0.2 0.3 0.4 0.5 0.6 0.7−1.
0−
0.5
0.0
0.5
1.0
normal vote in district
(libe
ral)
id
eolo
gy s
core
(co
nser
vativ
e)
xx
x
x
x
x
x
x
x
xx
x
x
x
x
xx
x
x
xx
x
xx
x
x
x
x
x
x
x
x
xx
x
x
x
x
x
x
x
x
x
x
x
x
xx
x
x
x
x
xx
xx
x
x
x
xx
xx
x
x
x
x
xx
x
x
x
x
x
x
xx
x
x
xx
x
x
x
x
x
x
xx
x
x
xx
x
x
x
x
x
x
x
xx
x
xx
xxx
xx
xxx
x
xx x
x
x
x
x x
x
x x
x
x
x
x
x
x
x
x xx
xxx
x
x
x
x
x
x x
x x
x
x
x
x
x
x
xx
xx
x
x
x
x
x
xx
x
x
x
x
xx
x
xx
xx
x
x
xx
x
x
x
x
xx
xx
x
x
x
x
x
x
x
x
x
x
xx
x
o
o
o
o
o
o
oo
o
o
o
o oo
o
oo
o
o
o
o
o
o
o
o
oo
o
o
o
oo
oo
ooo
oo
o
o
oo
o
o
o
o
o
oo
oo
o
oo
o
o
o
o
o
o o
o
o
o
o o
o
o
o
o
o
o
o
o
o
o
o
oo
o
o
oo
o
o
o oo
o
o
o
o
o
o
o
o
o
o
oo
o
o
oo
o
o
o
o
o
o
o
oo
o
o
oo
o
o o
o
o
o
o
o
o
o
o
o
oo
o
o
oo
o
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Estimating the electoral benefits of moderation
I Look at districts where Congressmembers are running forreelection
I Predict their vote share given their “ideology score”I Also control for Presidential vote in previous election
I Noisy estimate in any particular year, so plot estimates overtime
I Also look at Nixon, Clinton impeachments
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Estimating the electoral benefits of moderation
I Look at districts where Congressmembers are running forreelection
I Predict their vote share given their “ideology score”I Also control for Presidential vote in previous election
I Noisy estimate in any particular year, so plot estimates overtime
I Also look at Nixon, Clinton impeachments
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Estimating the electoral benefits of moderation
I Look at districts where Congressmembers are running forreelection
I Predict their vote share given their “ideology score”I Also control for Presidential vote in previous election
I Noisy estimate in any particular year, so plot estimates overtime
I Also look at Nixon, Clinton impeachments
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Estimating the electoral benefits of moderation
I Look at districts where Congressmembers are running forreelection
I Predict their vote share given their “ideology score”I Also control for Presidential vote in previous election
I Noisy estimate in any particular year, so plot estimates overtime
I Also look at Nixon, Clinton impeachments
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Estimating the electoral benefits of moderation
I Look at districts where Congressmembers are running forreelection
I Predict their vote share given their “ideology score”I Also control for Presidential vote in previous election
I Noisy estimate in any particular year, so plot estimates overtime
I Also look at Nixon, Clinton impeachments
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Estimating the electoral benefits of moderation
I Look at districts where Congressmembers are running forreelection
I Predict their vote share given their “ideology score”I Also control for Presidential vote in previous election
I Noisy estimate in any particular year, so plot estimates overtime
I Also look at Nixon, Clinton impeachments
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Estimated effects of moderation for reelection vote
analysis based on poole’s dwnom1 score
year
Est
imat
ed b
enef
it fr
om b
eing
a m
oder
ate
1960 1970 1980 1990 2000
−2%
0%2%
4%
Democratic incumbentsavg of both partiesRepublican incumbents
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Return to the median voter theorem
I Is the median voter theorem “true”?
I No, and yes . . .
I Systematic differences between Democrats and Republicans,even in comparable districts
I Moderation is worth about 2% of the vote: some motivationto be in the median, but not a lot
I Bush’s gamble in 2001–2004 (and Truman’s in 1945–1948):how does ideology map to policy?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Return to the median voter theorem
I Is the median voter theorem “true”?
I No, and yes . . .
I Systematic differences between Democrats and Republicans,even in comparable districts
I Moderation is worth about 2% of the vote: some motivationto be in the median, but not a lot
I Bush’s gamble in 2001–2004 (and Truman’s in 1945–1948):how does ideology map to policy?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Return to the median voter theorem
I Is the median voter theorem “true”?
I No, and yes . . .
I Systematic differences between Democrats and Republicans,even in comparable districts
I Moderation is worth about 2% of the vote: some motivationto be in the median, but not a lot
I Bush’s gamble in 2001–2004 (and Truman’s in 1945–1948):how does ideology map to policy?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Return to the median voter theorem
I Is the median voter theorem “true”?
I No, and yes . . .
I Systematic differences between Democrats and Republicans,even in comparable districts
I Moderation is worth about 2% of the vote: some motivationto be in the median, but not a lot
I Bush’s gamble in 2001–2004 (and Truman’s in 1945–1948):how does ideology map to policy?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Return to the median voter theorem
I Is the median voter theorem “true”?
I No, and yes . . .
I Systematic differences between Democrats and Republicans,even in comparable districts
I Moderation is worth about 2% of the vote: some motivationto be in the median, but not a lot
I Bush’s gamble in 2001–2004 (and Truman’s in 1945–1948):how does ideology map to policy?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Return to the median voter theorem
I Is the median voter theorem “true”?
I No, and yes . . .
I Systematic differences between Democrats and Republicans,even in comparable districts
I Moderation is worth about 2% of the vote: some motivationto be in the median, but not a lot
I Bush’s gamble in 2001–2004 (and Truman’s in 1945–1948):how does ideology map to policy?
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Political representationTrench warfareRational voting
Candidate positioningRecap
Summary: mathematical models in social science
I 4 examples where mathematical models gave “normative”conclusions:
I Proportional representation is fairI Cooperation is a good strategy in the repeated prisoner’s
dilemmaI Voting is irrational (unless you find it intrinsically enjoyable)I Politicians want to be at the median
I Each theory had big holes
I Each theory’s predictions were essentially qualitative
I Statistical models take the next step
I Similar ideas hold in psychology, sociology, economics, . . .
Andrew Gelman Mathematical vs. statistical models in social science
Top Related