STRATEGIES FOR RESEARCH

1
STRATEGIES FOR RESEARCH

• Approaching the Paper Assignment

2
OUTLINE

• Part I Basic Steps
• Key Steps in Research
• Defining a Topic
• Example Political Radicalism
• Expressing the Question in Terms of Variables
• Imagining Potential Explanations
• Framing Hypotheses
• Clarifying the Model
• From Concept to Measurement

3
OUTLINE (cont.)

• Testing Hypotheses
• Challenges, Technical and Analytical
• Postscript On Skewness
• Part II What You Will Do in Your Paper!

4
Key Steps in Research

• Observing variation in a variable, and proposing
  a preliminary explanation
• Stating a hypothesis
• Testing the hypothesis

5
Defining a Topic General Theme Political
radicalism On definition What do we mean by
radicalism? General Question What are the
sources of political radicalism? Specific
Question Why is there more political radicalism
in some societies (or locales) than in others?
Or, in other words, what explains variation (or
variance) in levels of political radicalism in
societies around the world?
6
Expressing the Question in Terms of
Variables Radicalism dependent variable
Y Source or cause or explanation independent
variable X Conceptualization Y
f(X) or, with greater complexity, Y f(X1
X2 X3 .. Xk)
7
Imagining Potential Explanations

• Poverty
• Inequality
• Fundamentalist religious indoctrination
• Colonial suppression
• Loss of privileged status

8
Framing Hypotheses

• In comparing units of analysis, those having
  one value on the independent variable will be
  more likely to have one value on the dependent
  variable than will those having a different
  value on the independent variable. Or,
• The greater (or lesser) the value of X, the
  greater (or lesser) the value of Y.

9
Clarifying the Model Inequality might lead to
rage, which in turn leads to radicalism Poverty
might affect calculations of risk (since poor
people have less to lose than others), and that
calculation might bolster willingness to engage
in radical action Thus rage and risk
calculation would constitute intervening
variables
10
From Concept to Measurement The Problem of
Operationalization Political radicalism
demonstrations, riots, assassinations
(Y) Plausible independent variables poverty
(X1) inequality (X2) colonial suppression
(X3) extremist indoctrination (X4) loss of
privileged status (X5) authoritarian repression
(X6) Question How to find indicators that are
reliable and valid?
11
The Technical Challenge Analyzing variation in
levels of political radicalism in other words,
discerning a pattern within the variance Sample
hypothesis The higher the level of poverty (X),
the greater the degree of political radicalism
(Y) Empirical question Is there covariance
between these variables? Any kind of systematic
relationship? Do they vary together? If so, to
what degree?
12

• The Analytical Challenge(s)
• Establishing cause-and-effect
• Considering unobserved (lurking) variables
• the What else? question
• Assessing roles of other X variables
• Inventing alternative explanations for any
• observed relationship (how might we interpret
• a positive relationship between poverty and
• radicalism?)

13

• Looking Ahead and Beyond.
• What if all bivariate relationships support the
• the separate hypotheses? What then?
• Check relationships among the independent
  variables
• Are they measuring different things? Or the
  same?
• Can we identify the relationship between Y and,
  say, X1,
• while controlling for the effects of X2 and X3?

14
YOUR PROJECTS DATASETS

• NES2000.sav
• States.sav
• World.sav

15
GETTING STARTED

• Scan data set(s)
• Select a dependent variable (Y)
• Determine its level of measurement
• Assess variation
• Obtain (and present) univariate descriptive
  statistics
• Select 2 potential independent variables (X1
  Xk)
• Check levels of measurement, variation, and
  (ideally) explore descriptive statistics on each
• Explain concepts and ideas at each step

16

• INTRODUCTION
• Why is your question important?
• Concepts and definitions see Pollack page 10
• STATEMENT OF HYPOTHESES
• Operationalization of concepts
• Identification of dependent, independent,
  intervening variables
• Formal hypotheses see Pollack pp. 33-37 plus
  lecture notes
• PRESENTATION OF VARIABLES
• Levels of measurement
• Univariate descriptive statistics including
  histograms and/or bar grams, measures of central
  tendency and dispersion

17

• BIVARIATE RELATIONSHIPS
• Assess form, strength, and significance of
  relationship
• Crosstabulation tables with frequencies,
  percentages, appropriate
  measures of association, chi-square statistic
• Regression Each independent variable with
  dependent variable full equation with
  coefficients, correlation, scatterplot with
  regression line, r-squared values,
• F-statistic and standard errors
• MULTIVARIATE RELATIONSHIPS
• Cross-tabulation or multiple regression
• Examine difference(s) from bivariate results

18

• CONCLUSIONS
• Accept or reject null hypotheses
• Confirm, reject, or modify hypotheses
• Provide alternative explanations for findings
• Indicate which explanation you think is most
  appropriate
• Describe possible avenues for further research
• OVERALL LENGTH 4-6 pages, plus tables and graphs