POL322: Introduction to Quantitative Reasoning
How to Study Politics Scientifically: September 16, 2013
Scientific Study of Politics
What do we study?
NOT merely facts and historical cases in Canadian politics, foreign countries or international relations
BUT more importantly the scientific knowledge of political phenomena, generalizable beyond specific
Science = a systematic method of knowledge creation.
what does it mean to study politics scientifically?
the goal is to develop our knowledge of the causal process behind various phenomena
science is arguably the most successful way to obtain and refine knowledge of the nature and our
theory (causal theory, theoretical model)
a tentative conjecture about the causes of some phenomenon of interest.
a theory-based statement about a relationship that we expect to observe.
a process in which scientists evaluate collected evidence systematically to make...
First we need a theory...
for a theory you need:
1) dependent variable (Y) - a phenomenon of interest (i.e democratization)
2) independent variable (X) - or explanatory variable -- the cause/factor of the phenomenon (i.e level of
3) explanation/mechanism/process that links your independent and dependent variables (why X causes
Y -- causal explanation).
all theories should have these 3 components.
economic voting example:
voters try to hold the government party accountable for the consequences of its economic policy.
they punish the government party for poor economic performance at the national level and reward it for
II. Then we need to derive a hypothesis from the theory...
operationalize these concepts into something measurable and testable.
i.e hypothesis of economic voting theory would predict that real GDP growth (economic performance
operationalized) would lead to higher incumbent party’s vote share.
when we make a hypothesis we should be able to come up with an expectation of what we would see if
the theory were true and if it were not true.
need to have a null hypothesis and falsifiability-- expectation that you should observe if your
hypothesis is wrong (0 correlation coefficient/negative correlation).
III. Then we test our hypothesis empirically...
regression analysis to find line of best fit, helps support of discredit your hypothesis empirically.
a process in which scientists evaluate systematically collected evidence to make a judgement of whether the evidence favours their hypothesis.
Process of producing scientific knowledge:
theory ---> hypothesis ---> empirical test ---> evaluation of theory and evaluation of hypothesis based
on empirical test.
when we look at slide of Canada federal election
multiple hypothesis and tests --
Alternative Hypothesis: incumbent government party’s vote share increases as unemployment
we should be able to point out what our null hypothesis would be.
evaluate institutional differences as alternate explanation -- i.e difference between Canadian
parliamentary and US presidential system.
empirical tests using data at different levels:
we have tested the theory of economic voting using aggregate data (=observations are aggregated at
what’s behind the aggregate...
individual voter’s perception of the state of national economy improves, this voter is more likely to
vote for the incumbent government party’s candidate.
individual’s perception of economy --influences--> individual’s vote choice.
its always tentative -- we remain open to the possibility that there is sti