Political Science 5 Lecture 4, 2504

1
Political Science 5 Lecture 4, 2/5/04

• 1. Homework 1 due on Tuesday!
• 2. Class get-together Fog of War

2
Inglehart article Group Activity
Answer the following questions What is the
hypothesis? What is the unit of analysis? What is
the dependent variable? -- Conceptual definition
and operational definitions What is the
independent variable? --Conceptual and
operational definitions What is the causal
story? Are issues of validity or reliability
mentioned?
3
The NYT article

• Research cycle
• Theories?
• Hypotheses?
• Dependent and independent variables?
• Measurements?
• What is normatively distasteful about such
  research?

4
Case, or Unit of Analysis

• A case, or unit of analysis, is the item (person,
  city, nation, and so on) for which we have data.
• Different research questions require different
  units of analysis.
• Individual data vs. aggregate data
• Examples of each?

5
More on Aggregate vs. Individual Data

• The danger of using the wrong data to test a
  hypothesis the Bush/divorce example from the
  text.
• Can someone describe this?

6
Approaches to Data Collection

• Surveys
• How are they conducted?
• Experiments
• Not so common in poli sci, but increasingly used
• Example from text
• Direct observation
• Content Analysis
• Extracting Data from Public Records

7
Measurement II. Quantifying and Describing
Variables

• Four Levels of Precision
• Measures of Central Tendency
• Mode
• Median
• Mean
• Measures of Dispersion
• Variance, Standard Deviation

8
Four Levels of Precision For Measuring Variables

• Nominal Measure You can put cases into a
  category, but cannot specify an order or
  relationship between the categories.
• Example The variable religion can take on
  values such as Catholic, Protestant, Mormon,
  Jewish, etc.

9
Four Levels of Precision For Measuring Variables

• Ordinal Measure You can put cases into different
  categories, and order the categories.
• Example The variable strength of religious
  belief can take on values such as devoutly
  religious, fairly religious, slightly religious,
  not religious.

10
Four Levels of Precision For Measuring Variables

• Interval Measure Not only can you order the
  categories of the variable, you can specify the
  difference between any two categories.
• Example. The variable temperature on the
  Fahrenheit scale can take on values such as 32
  degrees, 74 degrees, 116 degrees.

11
Four Levels of Precision For Measuring Variables

• Ratio Measure You can order categories, specify
  the difference between two categories, and the
  value of zero on the variable represents the
  absence of the variable.
• Example. The variable annual income can take
  on the values of 0, 98,000, or 694,294,129.

12
Four Levels of Precision For Measuring Variables

• In general, better to have higher levels of
  measurement.
• In specific cases, more convenient to use lower
  levels of measurement.

13
What about multi-dimensional phenomena?

• Constructing an index.
• Examples
• Democracy
• Support for abortion

14
Measures of Central Tendency

• Shaquille ONeal 17.1 million
• Kobe Bryant 11.8 million
• Robert Horry 5.0 million
• Glen Rice 4.5 million
• Derek Fisher 4.3 million
• Rick Fox 4.2 million
• Travis Knight 3.1 million

• Ron Harper 2.1 million
• A. C. Green 2.0 million
• Devean George 1.0 million
• Brian Shaw 1.0 million
• John Salley 0.8 million
• Tyronne Lue 0.7 million
• John Celestand 0.3 million

15
Measures of Central Tendency

• Mode The most frequently occurring value.
• 1.0 million
• Median The midpoint of the distribution of
  cases.
• 1. Arrange cases in order
• 2. If the number of cases is odd, median is the
  value taken on by the case in the center of the
  list.
• 3. If the number of cases is even, median is the
  average of the two center values. 2.6 million

16
Measures of Central Tendency

• Mean is the arithmetic average of the values that
  all the cases take on. 4.1 million.
• Add up all the values
• Divide this sum by the number of cases, N.

17
Measures of Dispersion

• The variance is a measure of how spread out cases
  are, calculated by
• Compute the distance from each case to the mean,
  then square that distance.
• Find the sum of these squared distances, then
  divide it by N-1. 22.66 million.

18
Measures of Dispersion

• The standard deviation is the square root of the
  variance, 4.76 million.