Choosing Research Designs II

1
Choosing Research Designs II

• Nonexperimental Methods

2
The Purpose of Control Variables

• We use control variables to account for possible
  alternative explanations we can think of.
• For example, when I examined whether democracies
  are generally more peaceful than autocracies I
  included several control variables.

3
Explaining Pacifistic Democracy

• Peace (Y) Democracy (X1) State Power (X2)
  Development (X3) of Bordering States (X4)
• In the model above, I have more confidence that
  Democracy is related to peace considering I
  control for the other variables that may skew my
  test.

4

• We need to take care that our theory is not
  missing other factors that may undermine the
  validity of our theory and tests.
• Our inferences will be flawed if we are actually
  capturing other processes through our variables.
• This means that the validity of our measures
  would be undermined.

5

• Several possible problems arise that are related
  to model misspecification and spurious
  relationships.
• Thus, we need to control for confounding factors
  and alternative explanations!!!

6
Model Misspecification and Spuriousness

• Antecedent variable A variable that indirectly
  affects the relationship between two other
  variables.
• For example, Ivy league education increases
  income.
• However, parental wealth and legacy admissions
  affect Ivy league education. Thus, income of
  graduates from Ivy League schools may not be
  random.

7

• Here Ivy League Parents is an antecedent variable
• Ivy League Parents Ivy League Kids
  high income kids
• Hence, admission to Ivy schools clearly not
  random or pure merit-based, and thus the income
  earned by these people.

8
Model Misspecification and Spuriousness

• Intervening Variable These may be spuriously
  related to another relationship.
• How can states fight each other if they are not
  contiguous with each other? Only the strongest,
  with large navies, bases, etc., could do so.
• Hence, geographic contiguity or distance is an
  intervening variable. States may or may not be
  more peaceful, but it is hard to avoid conflict
  when it is on your borders.

9
Model Misspecification and Spuriousness

• Alternative Variables We also want to control
  for variables that would bias our results if
  omitted.
• In this case, the X variables in a model would
  produce biased estimates, undermining their
  validity and producing error that leads to
  inaccurate inferences.

10
Here is a spurious relationship from my research

• IGOs conflicts
• Powerful states
• Powerful states both in more IGOs and conflicts,
  but these two variables not directly related but
  a function of state power.

11
Classic Spurious Case
???
Ice Cream Consumption
Crime



Summer Temperatures
Hence we see that despite the fact that ice cream
consumption is correlated with crime, the real
cause is that summer temperatures increase both
ice cream consumption and crime.
12
Veronica Says, Beat Marshall!!!
Go Miners!!! UTEP Fight! UTEP Win! Im going
to Homecoming, Are you?
13
Non-Experimental Designs

• These studies use data collected or aggregated
  from surveys, history, or government indicators
• Cross sectional studies
• Panel (cross sectional over a few time points)
• Longitudinal (time series and pooled
  cross-sectional time series)
• Case studies and focus groups

14
CROSS SECTIONAL Designs

• Statistical or case studies that compare
  individuals or subjects across several variables
• Surveys comparing peoples political views
• Comparison of countries, groups, organizations
  along different dimensions, such as countries
  with different levels of development (low,
  medium, high) relative to other factors.

15
Non-Experimental Designs

• These studies use data collected or aggregated
  from surveys, history, or government indicators
• Cross sectional studies
• Panel (somewhat rare)
• Longitudinal (time series and pooled
  cross-sectional time series)
• Case studies and focus groups

16
CROSS SECTIONAL Designs

• Statistical or case studies that compare
  individuals or subjects across several variables
• Surveys comparing peoples political views
• Comparison of countries, groups, organizations
  along different dimensions, such as countries
  with different levels of development (low,
  medium, high) relative to other factors.

17
Cross-Sectional Data
ID State Abortions/1,000women Bush04 Conservative score for House delegation
1 Alabama 15 62.5 73
3 Arizona 19.1 54.8 67
4 Arkansas 11.1 54.3 48
5 California 33.4 44.4 41
6 Colorado 18 51.7 67.8
7 Connecticut 23 44 37.6
8 Delaware 34.4 45.8 40
10 Georgia 21.2 58 63.7
12 Idaho 5.8 68.4 90
13 Illinois 25.6 44.5 48.9
14 Indiana 10.6 59.9 69
15 Iowa 9.8 49.9 64.6
16 Kansas 18.3 62 75
18
Example of a Panel Study
State Democracy Illiteracy HDI Islamic
Argentina91 7 4.3 0.81 0
Argentina95 7 3.7 0.832 0
Argentina00 8 3.3 0.854 0
Armenia91 7 2.57 0.751 0
Armenia95 3 2 0.708 0
Armenia00 5 1.69 0.754 0
Australia91 10 0 0.892 0
Australia95 10 0 0.932 0
Australia00 10 0 0.942 0
Azerbaijan91 -3 3 . 1
Azerbaijan95 -6 3 . 1
Azerbaijan00 -7 3 0.746 1
Bangladesh91 6 65 0.417 1
Bangladesh95 6 61.9 0.445 1
Bangladesh00 6 59.2 0.497 1
Belarus91 7 0.7 0.785 0
Belarus95 0 0.5 0.752 0
Belarus00 -7 0.5 0.775 0
Belgium91 10 2 0.897 0
19
Time Series

• Observations are made over time, which can
  provide descriptive information or used to test
  hypotheses.
• If testing hypotheses, we track data for a
  dependent variable and at least one independent
  variable over time (based on some measure e.g.
  days, weeks, months, or years)

20
Example of a Time Series Presidential Approval