Improving the Presentation of Introduction Practice in Political - - PowerPoint PPT Presentation

improving the presentation of
SMART_READER_LITE
LIVE PREVIEW

Improving the Presentation of Introduction Practice in Political - - PowerPoint PPT Presentation

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Improving the Presentation of Introduction Practice in Political Quantitative Results in Political Science Science Principles


slide-1
SLIDE 1

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Improving the Presentation of Quantitative Results in Political Science

John Kastellec and Andrew Gelman

Columbia University

February 12, 2009

slide-2
SLIDE 2

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Overview

  • What do political scientists do?
  • Why tables?
  • Why graphs?
  • Some basic principles
  • A graphing template
  • Some examples
slide-3
SLIDE 3

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Overview

  • What do political scientists do?
  • Why tables?
  • Why graphs?
  • Some basic principles
  • A graphing template
  • Some examples
slide-4
SLIDE 4

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Overview

  • What do political scientists do?
  • Why tables?
  • Why graphs?
  • Some basic principles
  • A graphing template
  • Some examples
slide-5
SLIDE 5

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Overview

  • What do political scientists do?
  • Why tables?
  • Why graphs?
  • Some basic principles
  • A graphing template
  • Some examples
slide-6
SLIDE 6

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Overview

  • What do political scientists do?
  • Why tables?
  • Why graphs?
  • Some basic principles
  • A graphing template
  • Some examples
slide-7
SLIDE 7

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Overview

  • What do political scientists do?
  • Why tables?
  • Why graphs?
  • Some basic principles
  • A graphing template
  • Some examples
slide-8
SLIDE 8

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Overview

  • What do political scientists do?
  • Why tables?
  • Why graphs?
  • Some basic principles
  • A graphing template
  • Some examples
slide-9
SLIDE 9

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

The use of tables vs. graphs in political science

  • Examined 5 journals in 2006 (Kastellec and Leoni

2007)

  • Coded tables/graphs, and purpose of each
  • Overall: 150 tables, 89 graphs
  • 10

20 30 40

Mathematical Other Non−numeric Predicted Values Estimates and Uncertainties Summary Statistics Percentage of Graphs and Tables Combined, by Category

  • 25

50 75 100

Percentage of Graphs Within Each Category

slide-10
SLIDE 10

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Tables?

  • Tables:
  • Much easier to produce
  • Standard in teaching, presentation and publishing
  • Can aid replication studies
  • Graphs:
  • Takes a lot of work
  • Belief it’s not possible to present info graphically
slide-11
SLIDE 11

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Tables?

  • Tables:
  • Much easier to produce
  • Standard in teaching, presentation and publishing
  • Can aid replication studies
  • Graphs:
  • Takes a lot of work
  • Belief it’s not possible to present info graphically
slide-12
SLIDE 12

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Tables?

  • Tables:
  • Much easier to produce
  • Standard in teaching, presentation and publishing
  • Can aid replication studies
  • Graphs:
  • Takes a lot of work
  • Belief it’s not possible to present info graphically
slide-13
SLIDE 13

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Tables?

  • Tables:
  • Much easier to produce
  • Standard in teaching, presentation and publishing
  • Can aid replication studies
  • Graphs:
  • Takes a lot of work
  • Belief it’s not possible to present info graphically
slide-14
SLIDE 14

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Tables?

  • Tables:
  • Much easier to produce
  • Standard in teaching, presentation and publishing
  • Can aid replication studies
  • Graphs:
  • Takes a lot of work
  • Belief it’s not possible to present info graphically
slide-15
SLIDE 15

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Graphs?

  • Better at communicating empirical results
  • Process of graph creation a feature, not a bug
  • Most data and results can be presented graphically
slide-16
SLIDE 16

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Graphs?

  • Better at communicating empirical results
  • Process of graph creation a feature, not a bug
  • Most data and results can be presented graphically
slide-17
SLIDE 17

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Why Graphs?

  • Better at communicating empirical results
  • Process of graph creation a feature, not a bug
  • Most data and results can be presented graphically
slide-18
SLIDE 18

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion Coefficient Variable (Standard Error) Constant .41 (.93) Countries Argentina 1.31 (.33)∗∗B,M Chile 93 (.32)∗∗B,M Colombia 1.46 (.32)∗∗B,M Mexico .07 (.32)∗∗A,CH,CO,V Venezuela 0.96 (.37)∗∗B,M Threat Retrospective egocentric .20 (.13) economic perceptions Prospective egocentric .22 (.12)♯ economic perceptions Retrospective sociotropic

  • .21 (.12)♯

economic perceptions Prospective sociotropic

  • .32 (.12)*

economic perceptions Ideological distance from

  • .27 (.07)**

president Ideology Ideology .23 (.07)** Individual Differences Age .00 (.01) Female

  • 0.03 (.21)

Education .13 (.14) Academic sector .15 (.29) Business sector .31 (.25) Government sector

  • .10 (.27)

R2 .15 Adjusted R2 .12 n 500 **p < .01, *p < .05, ♯p < .10 (two-tailed)

slide-19
SLIDE 19

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

  • −2

−1 1 2

Government sector Business sector Academic sector Education Female Age Ideology Distance from president Prospective sociotropic Retrospective sociotropic Prospective egocentric Retrospective egocentric Venezuela Mexico Colombia Chile Argentina R2 = 0.15 Adjusted R2 = .12 n = 500

slide-20
SLIDE 20

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Some principles

  • All graphs are comparisons
  • Graphs aren’t just for raw data; they’re for inferences

too

  • Communication to self as well as others
  • Graph + caption is a unit
slide-21
SLIDE 21

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Some principles

  • All graphs are comparisons
  • Graphs aren’t just for raw data; they’re for inferences

too

  • Communication to self as well as others
  • Graph + caption is a unit
slide-22
SLIDE 22

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Some principles

  • All graphs are comparisons
  • Graphs aren’t just for raw data; they’re for inferences

too

  • Communication to self as well as others
  • Graph + caption is a unit
slide-23
SLIDE 23

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Some principles

  • All graphs are comparisons
  • Graphs aren’t just for raw data; they’re for inferences

too

  • Communication to self as well as others
  • Graph + caption is a unit
slide-24
SLIDE 24

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Template

  • Figure X shows ...
  • Each point (or line) indicate ...
  • Before making this graph, we did ... which didn’t

work because ...

  • A natural extension would be ...
slide-25
SLIDE 25

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Template

  • Figure X shows ...
  • Each point (or line) indicate ...
  • Before making this graph, we did ... which didn’t

work because ...

  • A natural extension would be ...
slide-26
SLIDE 26

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Template

  • Figure X shows ...
  • Each point (or line) indicate ...
  • Before making this graph, we did ... which didn’t

work because ...

  • A natural extension would be ...
slide-27
SLIDE 27

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Template

  • Figure X shows ...
  • Each point (or line) indicate ...
  • Before making this graph, we did ... which didn’t

work because ...

  • A natural extension would be ...
slide-28
SLIDE 28

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

slide-29
SLIDE 29

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

slide-30
SLIDE 30

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Age and voting

Year Republican share of the two−party vote 1988 1992 1996 2000 2004 2008 30% 40% 50% age 18−29 age 30−44 age 45−64 age 65+ The youth vote and everybody else

slide-31
SLIDE 31

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Peacekeeping

5 10 15

The Efficacy of Post−Cold War−era Peacekeeping Operations

Pre−treatment measure of problems in country in conflict Time at peace (years) Very Bad Less Bad

  • Peacekeeping−−Peace holds

Peacekeeping−−War resumes No Peacekeeping−−Peace holds No Peacekeeping−−War resumes

slide-32
SLIDE 32

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Ideology of sports fan versus non-fans

slide-33
SLIDE 33

Multiple regression models: table

Dependent Variable = County-Level Turnout Excluding Excluding counties Full sample counties Full sample Full w/ partial w/state-year Full w/ partial w/state-year sample registration dummies sample registration dummies (1) (2) (3) (4) (5) (6) % of county

  • 0.039**
  • 0.036**
  • 0.051**
  • 0.037**
  • 0.034**
  • 0.050**

registration (0.003) (0.003) (0.003) (0.003) (0.003) (0.003) Law change

  • 0.020**
  • 0.018**
  • 0.023**

(0.005) (0.005) (0.006) Log population 0.048** 0.036** 0.017 0.047**

  • 0.035**

0.016 (0.011) (0.012) (0.010) (0.011) (0.021) (0.010) Log median

  • 0.133**
  • 0.142**

0.050**

  • 0.131**
  • 0.139**
  • 0.049**

family income (0.013) (0.014) (0.013) (0.013) (0.014) (0.013) % population with 0.071* 0.070* 0.011 0.072* 0.071* 0.013 h.s. education (0.028) (0.029) (0.024) (0.028) (0.029) (0.024) % population

  • 0.795**
  • 0.834**
  • 0.532**
  • 0.783**
  • 0.822**
  • 0.521**

African American (0.056) (0.059) (0.044) (0.055) (0.059) (0.044) Constant 1.47** 1.70** 0.775** 1.45** 1.68** 0.819** (0.152) (0.171) (0.124) (0.152) (0.170) (0.127) R2 0.91 0.91 0.94 0.91 0.91 0.94 N 3572 3153 3572 3572 3153 3572

  • Note. *p < .05, **p < .01. Huber-White standard errors in parentheses. Year dummies and

state-year dummies are not reported.

slide-34
SLIDE 34

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

  • −0.04

−0.02 0.02 0.04 % of county registration

  • −0.04

−0.02 0.02 0.04 Law change

  • −0.1

−0.05 0.05 0.1 Log population

  • −0.2

−0.1 0.1 0.2 Log median family income

  • −0.2

−0.1 0.1 0.2 % population with h.s. education

  • F

u l l S a m p l e E x c l u d i n g c

  • u

n t i e s w . p a r t i a l r e g i s t r a t i

  • n

F u l l s a m p l e w . s t a t e y e a r d u m m i e s −1 −0.5 0.5 1 % population African American

  • With law

change dummy

  • Without law

change dummy

slide-35
SLIDE 35

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Conclusion

  • Software needs to be improved so that graphs are

automatic

  • Make data more available (aid replication)
  • Change incentives: encourage people to use graphs
slide-36
SLIDE 36

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Conclusion

  • Software needs to be improved so that graphs are

automatic

  • Make data more available (aid replication)
  • Change incentives: encourage people to use graphs
slide-37
SLIDE 37

Improving the Presentation of Quantitative Results in Political Science John Kastellec and Andrew Gelman Introduction Practice in Political Science Principles Examples

Coefficients & standard errors Graphs for methodological research Age and voting Peacekeeping Sports fans Multiple regression models

Conclusion

Conclusion

  • Software needs to be improved so that graphs are

automatic

  • Make data more available (aid replication)
  • Change incentives: encourage people to use graphs