What should we control for? Or, some 5ps on your essay. WK 4 Andrea Ruggeri Q Step, Year 2

Transcription

1 What should we control for? Or, some 5ps on your essay WK 4 Andrea Ruggeri Q Step, Year 2

2 Main Goals (re)thinking about confounders How can we decide about other Xs? How to tackle essays and exams DemocraJc Peace!!! Some thoughts on ComGov and PolSoc

3 Bivariate to muljvariate, why? Chocolate consumpjon and noble prize = + Controlling for GDP and Geography! Pipe smoking >CigareTe Smoking = death rate Controlling for age

4 Confounding variable A great strategy for essays or exam quesjons in which you assess evidence is to think about confounding variables. Two strategies for thinking of confounding variables: (a) think of things that affect y but are not in the regression [this is basically what Kellstedt and WhiTen say to do] and then ask yourself whether it might be related to x. (b) think of things that affect x (or things that are related to and precede x) but are not in the regression and then ask yourself whether it might affect y.

5 What makes a control variable important? Does the variable have a causal effect on Y? Is the variable correlated with those variables whose effects are the focus of the study? Absence of control variables à Spurious correlajon and omited variable bias Allison, Paul D. Mul$ple regression: A primer. Pine Forge Press, 1999.

6 Prof. Eggers Eggers WK 7, Poli5cal Analysis Year 1

7 Essay: Past & Future Past A lot of bivariate regressions and scater plots Less muljvariate!!! AlternaJve hypothesis? CompeJng explanajons? Factors influencing X? Future: Your 2 nd year essay should focus on these challenges Spurious? How do we idenjfy confounders? How do we control for them?

8 IR 214 Syllabus , page 15 Topic 3: Global Governance and Security Ques$on Two: Does the democrajc peace theory represent a challenge to Realism? OR The greater number of democrajc states in the world has made it a more peaceful place. Do you agree?

9 **Readings **Barkawi, Tarak and Laffey, Mark (eds.), Democracy, Liberalism and War: Rethinking the Democra$c Peace Debate (2001) **Brown, Michael, Lynn- Jones, Sean and Miller, Steven (eds.), Deba$ng the Democra$c Peace (Cambridge, MA, 1996) **Doyle, Michael, Kant, Liberal Legacies and Foreign Affairs, Philosophy and Public Affairs, vols. 12, 3 and 4 (Summer and Fall 1983) **Kant, Perpetual Peace (1795) in Hans Reiss (ed.), Kant s Poli$cal Wri$ngs (1991) **Layne, Christopher, Kant or Cant: The Myth of DemocraJc Peace, Interna$onal Security 19 (1994), 5-49 **Mansfield, Edward D. and Snyder, Jack, DemocraJzaJon and the Danger of War, Interna$onal Security (1995), 20, 5-38 **Owen, John, DemocraJc Peace Research: Whence and Whither? Interna$onal Poli$cs (2004), **Rosato, SebasJan, The Flawed Logic of DemocraJc Peace Theory, American Poli$cal Science Review 97 (2003),

10 Monadic DP Democra5c Peace Prelims and IR paper democracies more peaceful Dyadic DP democracies don t fight each others Structural DP more democracies stabilize internajonal system Cross- secjon Sample = N Times- Series =T TSCS= N*T Dyad- Years=((N(N- 1))/2)*T

11 DP as EvoluJon of Behavioral/ AnalyJcal Study of IR Kant's Zum ewigen Frieden The classical studies of war Richardson, Sorokin, Wright Babst (1964, 1972) Singer & Small (1976), Correlates of War Project Rummel (1983), Doyle (1983, 1986) Maoz & RusseT (1992, 1993), RusseT (1993)

12 Monadic DP Democra5c Peace Prelims and IR paper democracies more peaceful Dyadic DP democracies don t fight each others Structural DP more democracies stabilize internajonal system

13 What is the unit of analysis? Dyad- year. This means it looks at pairs of states over Jme. Dyadic research design has the benefit of shiying the focus onto the interacjon that exists between two states it s not state A s level of democracy that maters, but the relajonship between state A and state B s level of democracy.

14 Empirical PaTern vs ExplanaJon Democracies don t fight each others Why? NormaJve (Maoz & RusseT 1993) Structural (Maoz & RusseT 1993) Signals/InformaJon (Fearon 1994) Common Interests (Gartzke 1998)

15 Realists atack DemocraJc Peace It is about alliances (DEM s.s. à omizng an X) It is about Cold War period (Jme/sample) It is about capabilijes(dem s.s. à omizng an X) It is about geography DEM s.s. à omizng an X

16 Maoz, Zeev, and Bruce RusseT. "Alliance, conjguity, wealth, and polijcal stability: Is the lack of conflict among democracies a sta5s5cal ar5fact?. Interna9onal Interac9ons 17.3 (1992): ) Wealth Democracy à cost/befit war 2) Economic Growth Democracy à losing possible benefits 3) Common Alliance Democracy à same countries Here we use simple bivariate con$ngency tables to analyze the hypotheses. (page 251)

17 Maoz, Zeev, and Bruce RusseT. "NormaJve and Structural Causes of DemocraJc Peace, " American Poli9cal Science Review (1993):

18 KanJan Tripod Oneal, John, and Bruce RusseT. "TriangulaJng peace.". New York,: Norton (2001). Democracy Interdependence Interna$onal Organiza$on

19 CriJcisms the causes of war and of militarized disputes are different. à Different Xs for Different Ys; Y s opera$onaliza$on effect of economic interdependence, depend upon a par5cular, complex specificajon that is inappropriate. à Other Xs results indicate that the effects of the liberal (and realist) variables in their model differ when esjmated with informajon from 5me series rather than cross sec5onal varia5on. à Sample varia$on The pacific benefits of democracy and interdependence are spurious and their conflict- reducing effects are probably amributable to economic development à Omi[ed X Oneal, John R., and Bruce RusseT. "Rule of three, let it be? When more really is beter." Conflict Management and Peace Science 22.4 (2005):

20

21

22 ConJguity Trade Peace

23 Enemy Within!!! Capital(ist) Flows Democracy Peace It is not about democracy is about capitalism!

24 Capitalist Peace Gartzke, Erik. "The capitalist peace." American Journal of Poli9cal Science 51.1 (2007):

25

26

27

28 Islam & Authoritarianism Fish, M. Steven. "Islam and authoritarianism." World poli$cs (2002): 4-37.

29

30

31 Donno, Daniela, and Bruce RusseT. "Islam, Authoritarianism, and Female Empowerment: What Are the Linkages?." World Poli$cs (2004):

32 Do Women Need Women RepresentaJves?

33 What should we control for? Theory on data generajng process of Y (War/ Peace) AlternaJve theories ( knowing literature!) Realism versus democrajc peace Liberal peace / DemocraJc Peace/ Capitalist Peace BeTer measurement/proxy of alternajve theory Trade/ Democracy / Financial flows Theory on data generajng process of X (Democracy)

34 Some other important aspect to consider On Unit of analysis What theory assumes? Individuals, groups, state, dyad On Sample Geographic Sample Temporal Sample On OperaJonalizaJon Measurement of Y War vs Militarized Disputes Measurement of X LinguisJc FracJonalizaJon vs Horizontal InequaliJes

35 Ethnicity & Civil War Fearon, James D., and David D. LaiJn. "Ethnicity, insurgency, and civil war." American poli$cal science review (2003):

36 Re- operajonalizing an X

37 Cederman, Lars- Erik, Nils B. Weidmann, and KrisJan Skrede Gleditsch. "Horizontal inequalijes and ethnonajonalist civil war: A global comparison." American Poli9cal Science Review (2011):

38 Spurious CorrelaJons are Dangerous!!!

39 Next Weeks à Dave Kirk 5. Randomized experiments 6. AlternaJves to randomized experiments 7. Threats to internal and external validity 8. EvaluaJng research and evidence