POLS1009 Lecture Notes - Lecture 12: Selection Bias, Dependent And Independent Variables, Big Finish Productions
L12 THE BIG FINISH
1. What did we try and accomplish this year?
2. Can political science be scientific (in the physicist sense)?
3. What’s on the final exam?
Political Science?
- Political science is about the scientici study of political phenomena.
- This implies that something “scientific” is going on.
- The “science” in “political science” refers to the “method” (or the process) by which political
scientists go about learning about and explaining the political
oComparative method
oDescriptive statistics causation.
- Scientific statements are falsifiable (Popper)
oFalsifiability means that there must be some imaginable observation that can falisfiy or
refute the statement
oDoesn’t mean that it will be falsified, just that it can be falsified
oContrast with tautologies: statements that are true by definitions
“God created the world”
It may be true, but can’t be falsified because can’t be tested ergo not scientific.
Only evidence we find it not falsifiable.
- Statistical analysis are better for replication.
1. Questions
- Observe the political world and come up with a question or a puzzle
- Look for the unexpected or the surprising
- A surprise or a puzzle suggests that an existing theory is not holding up or requires more attention
2. Causal Theory
- Found a puzzle or something that needs explaining? Great!
- Now you need a theory or a “model” to explain it.
- Theory is a set of logicaly consistent statements that tell us why the things that we observe occur
find more resources at oneclass.com
find more resources at oneclass.com
oNeed logical consistency otherwise we can’t make predictions. If there are no predictions,
then there is no way that we can test the theory
- Models? Simplifications that help explain more complex phenomena.
oA model airplane cannot fly, but its shape can tell us whether the real thing has a chance of
flying
- PS are interested in the capacitive theory.
3. Hypothesis
- A hypothesis is a theory based statement about a relationship that we expect to observe
- For every hypothesis there is a corresponding null hypothesis. The null hypothesis is a theory-based
statement about what we should observe if there is no relationship between an independent and
dependent variable.
- Hypothesis testing is a process by which scientists evaluate systematically collected evidence to
make a judgement of whether the evidence favours the hypothesis or the corresponding null
hypothesis
4. Empirical Test
- Examine if the implications of the model are consistent with observation drawing on data from the
real world
- We are trying to falsify the theory. Repeated, failed attempts to falsify lead us to have greater faith
in the theory
- Always subject a model to the hardest test you can think up.
5. Evaluation of the hypothesis
- Evaluate the hypothesis relative to the corresponding null hypothesis
- Scientists always favour the null hypothesis
- Statistical techniques allows scientists to make probability – based statements about the empirical
evidence collected.
- Even when a hypothesis has an 80-20 edge over the null hypothesis, most scientists will favour the
null hypothesis.
oLooking for 95% confidence.
- Don’t want to make the chance of falsely rejecting the null hypothesis.
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Political science is about the scientici study of political phenomena. This implies that something scientific is going on. The science in political science refers to the method (or the process) by which political scientists go about learning about and explaining the political: comparative method, descriptive statistics causation. It may be true, but can"t be falsified because can"t be tested ergo not scientific. Only evidence we find it not falsifiable. Statistical analysis are better for replication: questions. Observe the political world and come up with a question or a puzzle. A surprise or a puzzle suggests that an existing theory is not holding up or requires more attention: causal theory. Now you need a theory or a model to explain it. Theory is a set of logicaly consistent statements that tell us why the things that we observe occur: need logical consistency otherwise we can"t make predictions.
