The document discusses various concepts and strategies related to measurement in the social sciences, including defining different levels of measurement such as nominal, ordinal, interval and ratio scales. It also covers developing operational definitions and valid and reliable measures for concepts through careful consideration of indicators and data collection methods. Strategies for assessing validity and reliability of measures like test-retest analysis are presented.

1. GEOG 420 / POL 420 / SOC 420
Fall 2019

4.  Measurement
 Systematic observation and representation using
scores or numerals of variables under investigation
 Numeral: Symbol in form 1, 2, 3, or I, II, III, etc.
 Numeral with quantitative meaning = Number
 Numbers: Description, Explanation, Prediction

5.  Numbers assigned to objects or events
 Example: DummyVariable for Gender
▪ 1 = Females, 0 = “Otherwise” (Males)
 A rule specifies a process used to assign numerals
or numbers to objects and events
 Example: Freedom House

6. POLITICAL RIGHTS
 Electoral Process
 Political Participation
 Functioning of
Government
CIVIL LIBERTIES
 Freedom of Expression
 Associational /
Organizational Rights
 Rule of Law
 Personal Autonomy and
Individual Rights
Free = 1.0 to 2.5
Partly Free = 3.0 to 5.5 Not Free = 5.5 to 7.0

7. FREE PARTLY FREE NOT FREE
Americas 24 (69%) 10 (28%) 1 (3%)
Asia-Pacific 16 (41%) 15 (38%) 8 (21%)
Central /
East Europe
13 (45%) 9 (31%) 7 (24%)
Middle East /
North Africa
1 (6%) 4 (22%) 13 (72%)
Sub-Saharan
Africa
9 (18%) 21 (43%) 19 (39%)
Western Europe 24 (96%) 1 (4%) 0 (0%)

8.  Restating concept so it can be tested
 How will we measure concept?
 Operational Definition: Deciding what kinds
of observations should be made to measure
occurrence of attribute or behavior

9.  Step #1: Thinking through what concept
means and how will we define it
 Step #2: Decide which variables we will use
to measure concept
 Step #3: Propose specific indicators of concept
 Step #4: Select data sets or instruments to
measure indicators

12. INSTRUCTIONS
Devise a measurement strategy
for a concept in your field.

14. ORDERABLE
 Ordinal Level
 Internal Level
 Ratio Level
NON-ORDERABLE
 Nominal Level

15.  Classification of observations into categories
 Examples:
 Religious Faith
▪ 1 = Christian, 2 = Jewish, 3 = Muslim
 Race
▪ 1 =White, 2 = African-American, 3 =Asian
▪ 4 = Native American, 5 = Pacific Islander

16.  Variable X has values X1, X2, and X3
 X1 < X2 < X3
 Uncertainty of Equality Between Values
 Example: Olympic Performance
 X1 = Bronze Medal
 X2 = Silver Medal
 X3 = Gold Medal

17.  Variable X has values X1, X2, and X3
 X1 > X2 > X3
 Equality
 NoTrue Zero
 Example:
Temperature

18.  Interval variable with true zero point
 Examples: Income,Years of Education

20.  Dichotomous
 Variable that can take on only two values
 Example: Gender (Either Male or Female)
 Discrete
 Orderable variable that can take on limited values
 Examples: 1, 2, 3, 4; 1979, 1980, 1981, 1982
 Continuous
 Orderable variable that can take on limitless set of values
 Example: Decimals Pi = 3.14159…

23.  Extent to which a measurement procedure
measures what it intends to measure
 Valid measure provides true and accurate
picture of something

24.  Does the measure appear to be valid “on face?”
 To properly assess face validity, need to know:
▪ Meaning of the concept being measured
▪ Whether information collected is germane to concept
 Examples: Political Ideology; IQTests

25.  Extent to which measuring instrument
consistently measures what it is measuring
 Consistent results across individuals / times

26. Type of Reliability How to Measure
Stability orTest-Retest
Give the same assessment twice,
separated by days, weeks, or
months. Reliability is stated as the
correlation between scores atTime
1 andTime 2.
Alternate Form
Create two forms of the same test
(vary the items slightly). Reliability
is stated as correlation between
scores ofTest 1 andTest 2.
Internal Consistency
Compare one half of the test to the
other half.

27.  Two researchers independently find same
results when looking at data
 Why would this be useful?