The Scientific Process

1
The Scientific Process

• Research Question
• Theory and Hypotheses
• Research Design
• Operationalization (measurement)
• Empirical Observation and Analysis

2
The Research Question

• A personal choice
• Should be inspired by existing literature
• Something important
• Basic vs. Applied research

3
Theory Construction

• A set of assumptions involving a set of
  interrelated concepts from which a causal
  statement(s) can be derived. These assumptions
  and causal statement(s) constitute an explanation
  for the phenomenon under investigation.

4
Theory Construction

• A reasoned and precise speculation about the
  answer to a research question, including a
  statement about why the proposed answer is
  correct. Must have observable implications.

5
Good Theories

• Can be generalized far beyond the specific data
  being analyzed
• Are not TOO abstract
• Are falsifiable

6
Concepts and Theories

• Concepts are the building blocks of theories
• Def A concept is an abstraction representing an
  object, a property of an object, or a certain
  phenomenon.
• Examples Political participation, socioeconomic
  status, party competition, welfare generosity

7
Example of a (Simple) TheoryWhat is the effect
of negative advertising on political
participation?

• Higher levels of exposure to negative advertising
  lead to lower levels of political trust because
  voters gradually become disenchanted with all
  political candidates. As trust is eroded, we
  should expect to see lower levels of political
  participation.
• Neg. Advertising ? Trust ?Participation

8
Some Important Theories in Political Science

• Median Voter Theory
• The Race to the Bottom in State Welfare Policy
• Costs/Benefits of Voting
• Inter-Party Competition and Responsiveness
• The Democratic Peace

9
What explains the positions parties take during
an election? Median Voter Theory

• Assumptions
• Voter preferences can be summarized as falling
  somewhere on a liberal-conservative (left-right)
  ideological scale
• Voters choose the party which most closely
  reflects their policy preferences
• Parties are solely motivated to win elections

10
Median Voter Theory
11
Median Voter Theory

• Observable Implications
• In a 2-party system, the party that locates
  itself closest to the median position will win
• In 2-party systems, parties will compete with one
  another by converging toward the median position

12
Theories and Hypotheses

• A hypothesis is a testable statement of causal
  relationship between two variables, derived from
  theory
• For directional variables, relationships
  expressed in hypotheses may be either positive or
  negative

13
Concepts vs. Variables

• Concepts
• Highly abstract
• Can represent a variety of things
• Variables
• Generally more specific/observable
• Takes on at least 2 values/categories that vary
  across the units/cases in our analysis

14
Causal Relationship

• Change in one variable causes change in another
  variable

15
State Welfare Benefits and the Race to the
BottomMaximum AFDC Benefit for a Family of
Four, 1975-95
16
RTB - Theory

• Assumptions
• State policymakers seek to improve state economic
  health by keeping taxes low/increasing tax base
  (prefer to keep the poor from moving to their
  state)
• Poor people are aware of welfare benefits across
  states and (since 1970) make migration decisions
  in part on state welfare benefits

17
RTB - Hypotheses

• The generosity of a states welfare benefits
  (relative to other states) is positively related
  to the number of poor people who move to that
  state
• The generosity of the welfare benefits among a
  states neighbors is positively related to that
  states welfare benefits

18
Dependent vs. Independent Variable

• Dependent variable what we are trying to
  explain
• Independent variable explains variation in
  dependent variable

19
Hypotheses an example

• Theory (conceptual)
• Neg. Advertising ? Trust ?Participation
• Hypothesis (operational)
• As the number of hours of campaign commercials
  watched increases, the probability that an
  individual will vote decreases

20
Good Hypotheses

• Are the logical implication of the theory being
  tested
• Are stated in explicit, empirical terms
• Can be generalized to different contexts
• Are plausible
• Clearly specify a relationship between an IV DV

21
Directional Hypotheses

• Apply to cases where IV and DV are orderable
  (directional) variables
• Positive relationship
• As ones education increases, the probability of
  voting increases
• There is a positive relationship between ones
  education level and voting

22
Directional Hypotheses (contd)

• Negative relationship
• As the number of hours of negative ads watched
  increases, the probability that an individual
  will vote decreases
• There is a negative relationship between exposure
  to negative advertising and the probability that
  an individual will vote

23
Non-Directional Hypotheses

• Appropriate for variables that are not orderable
• Hypothesis describes comparison among categories
• Examples
• Men have greater levels of support for President
  Bush than do women
• Whites are most likely to be Republican, while
  African-Americans are most likely to be Democrat

24
The Null Hypothesis

• The null hypothesis states that there is no
  relationship between the independent variable and
  the dependent variable
• Hypothesis There is a positive relationship
  between exposure to negative ads and turnout
• Null hypothesis There is no relationship between
  negative ads and turnout

25
Specifying the Unit of Analysis

• Def The social entities whose characteristics
  are the focus of study.
• May include individuals, groups, programs,
  organizations and institutions, cities, states,
  nations, etc.
• Individual vs. Aggregate level of analysis
• Ecological fallacy

26
What is the unit of analysis?

• There is a negative relationship between exposure
  to negative advertising and the probability that
  an individual will vote
• There is a negative relationship between the
  number of negative ads broadcast in a state and
  the turnout rate in the presidential election

27
What is the unit of analysis?

• There is a negative relationship between the
  level of wealth in a country and the level of
  democracy
• There is a positive relationship between a
  basketball teams shooting percentage and its
  winning percentage