Quick Quiz

1
Quick Quiz

• What does it mean if a statistic is biased?
• What kind of error causes bias?
• What does it mean if a distribution has high
  variability?
  
• What kind of error increases variability?

2
Designing Research
3
Where to Begin?
4
The inductive Approach Examples of things we
see in the real world

• Observed Patterns
• Incumbents get re-elected more often than not.
• Students from wealthy schools do better
  academically.
  
• Descriptive Statistics
• On average 55.6 of Texas students passed the
  aptitude test.
  
• 39 of americans 50-64 yrs. Old are actively
  avoiding carbs.
  
• Real World Needs
• We want to stop the spread of AIDS
• We want to reduce unemployment in america

5
The Deductive ApproachExamples of ideas we
already have about the world

• Math
• 22 4
• Area of a circle pr2
• Existing Theories
• The law of supply and demand.
• Theory of the median voter.
• Gravity
• Analogies
• Electrons
• Market

6
Places in the Real World for you to start your
research

• Read the News
• New York Times
• The Economist
• General Political Science Data
  psrm.cqpress.com/data_resources.htmldatasets
  
• Data for American Politics
• Public Opinion www.ropercenter.uconn.edu
• US economics www.economicindicators.gov
• Data for IR
• War www.umich.edu/cowproj/dataset.html
• Minorities at Risk www.cidcm.umd.edu/inscr/mar
  
• Data for Comparative Politics
• Developing Countries www.worldbank.org
• Labor Statistics www.ilo.org

7
Getting more specific
8
Forming the Research Question

• Deductive or Inductive, you must develop a
  question.
  
• What will happen if is a question experimental
  scientists often ask.
  
• For social scientists, Why questions are good
  since nature designs our experiments, we
  explain the results.
  
• Puzzles make the best social science research
  questions.

9
Research Question Examples

• Experimental Science questions
• Will this drug slow cancer?
• Does more testosterone increase aggression?
• Social Science Why questions
• Why dont democracies fight each other?
• Why are some former colonies poor, while others
  prosper?
  
• Why do African Americans usually vote Democrat?

10
PuzzlesThe best research questions.

• Why havent women supported John Kerry as much as
  they usually support the democratic presidential
  candidate?
  
• Why do some political parties boycott unfair
  elections, while others participate in those same
  elections?
  
• Why do countries go to war when they know they
  arent going to win?
  

11
After youve got the question, you need a Theory.
12
Constructing a Theory

• Your theory provides a GENERAL answer to your
  research question.
  
• Your theory tells a story of cause and effect.
• Theories can be inductive or deductive
• Inductive theories come from experience
• Deductive theories follow from principles

13
A Good Theory 2 Key Features

• Good ? True
• A theory is NEVER proven true. The best it can
  do is exhibit the two characteristics below
  
• Falsifiability MOST important
• You must be able to think of a way that your
  theory could be proven WRONG (falsified).
  
• If there is no way to potentially falsify your
  theory, it is not a good theory.
  
• Parsimony
• The theory is simple, easy to understand
• Analogies can be very useful and effective.

14
More on Falsifiability

• For a good theory, should be able to say
• If my theory is true, I expect to see xif my
  theory is NOT true, I expect to see y.
  
• If a theory is not falsifiable, it is not good.
• Examples of non-falsifiable theories
• Conspiracy Theories
• Marxs theory of History
• General rule of thumb
• If a theory explains everything, its probably
  non-falsifiable.
  

15
More on Parsimony

• The simpler the better
• A theory doesnt HAVE to be parsimonious to be
  good, but its preferable.
  
• Should be easy to understand
• Analogies can be powerful
• Insulin is like a key
• Party ID is like religion
• Examples of parsimonious theories
• Theory of the Median Voter
• Barrington Moore No Bourgeoisie, No Democracy
• Marxs theory of History

16
Coming up with a Good TheoryA few research
questions

• Why do Toy manufacturers advertise during
  Saturday morning cartoons?
  
• Why is CD packaging so hard to open?
• Why does my line at the grocery store always move
  the slowest?
  
• Why do some resource-rich countries remain poor?

17
Operationalization
18
Constructing Variables

• With operationalization, we turn the concepts of
  our theory into variables.
  
• To operationalize a theory, you must have, at
  least, 2 variables
  
• The Independent Variable (IV)
• Represents the cause in your story
• Can also be called treatment or explanatory
• The Dependent Variable (DV)
• Represents the effect you are trying to explain
• Other variables may also be necessary, but this
  depends on your research design.

19
Find the IV and DV for these theories

• Smoking causes cancer
• Masturbation causes blindness
• Home ownership is up because low interest rates
  make it easier for people to get home loans.
  
• Blondes have more fun because gentlemen prefer
  blondes.
  
• Tax cuts reduce unemployment by allowing
  businesses to hire more workers.