Lecture+6+-+Indexes%2C+Scales%2C+and+Typologies
Lecture Overview
Topic: Indexes, Scales, and Typologies
Objectives:
Discuss the benefits of indexes and scales.
Explain the logic behind index and scale construction.
Describe how typologies classify individuals across multiple variables.
Provide examples of indexes, scales, and typologies.
Indexes and Scales
Definitions:
Index: A composite measure summarizing and rank-ordering multiple observations that represent general dimensions.
Scale: A composite measure consisting of several items structured logically or empirically.
Key Similarities Between Indexes and Scales
Both are:
Ordinal measures of variables.
Composite measures based on more than one data item.
Importance of Indexes and Scales
Benefits:
Capture multiple dimensions of a concept more effectively than single indicators.
Possess improved validity.
Allow for a wider range of variation in ordinal rankings.
Logic of Index Construction
Example of Political Activism:
Different political actions signify varying degrees of activism (e.g., writing a letter, giving money).
An overall political activism index could assign points for each action taken by an individual.
Logic of Scale Construction
Activism Scale: Examples of political actions vary by intensity (e.g., running for office is more challenging than voting).
Indexes vs. Scales
Scales are typically superior to indexes as they consider the intensity of indicators reflecting the variable measured.
Indexes accumulate indicators simply without considering their weight or significance.
Steps in Index Construction
Item Selection
Select appropriate items to measure a variable.
Examination of Empirical Relationships
Index Scoring
Index Validation
Item Selection Criteria
Face Validity:
Items should logically represent the concept (e.g., morality composed of compassion, justice).
Unidimensionality:
Each composite measure must represent a single dimension of the concept.
Generality vs. Specificity:
Choose between general or specific items based on the index's focus.
Variance:
Ensure items divide participants into groups based on the variable.
Empirical Relationships
Relationship types include:
Bivariate: Interaction between two variables.
Multivariate: Interaction involving multiple variables.
Index Scoring Process
Determining Range:
Define the score range and balance it with case numbers per score.
Weight Assignment:
Decide on equal or differential weighting for items in the index.
Index Validation Methods
Item Analysis: Assess whether individual items independently contribute to the measure.
External Validation: Test the index against other presumed indicators of the same variable.
Handling Missing Data
Approaches include:
Excluding cases with missing data.
Treat missing data as valid responses.
Analyze and interpret missing data.
Assign average values for missing data.
Proportionally score the missing responses based on available data.
Scale Construction
Acknowledges that not all indicators of a variable are proportionately important.
Scales ensure ordinal measurement by capturing intensity relationships among indicators.
Common Types of Scales
Bogardus Social Distance Scale: Measures willingness for social relations with different groups.
Thurstone Scale: Measures varying indicators based on assigned weights using judges' rankings.
Likert Scale: Uses standard categories to measure degrees of agreement.
Semantic Differential Scale: Assesses responses along a continuum between opposing adjectives.
Guttman Scale: Summarizes various observations and identifies a structure where strong indicators imply weaker ones.
Examples of Scales
Bogardus Scale Example: Measures acceptance of sex offenders at varying proximity levels.
Thurstone Scale Example: Ranks indicators of aggression based on perceived strength.
Likert Scale Example: Rosenberg Self-Esteem Scale measures self-esteem through agreement-disagreement items.
Guttman Scale Example for Abortion Attitudes
Survey responses can be categorized to highlight pro-choice attitudes based on the situation (health risk, rape, unmarried status).
Typologies
Definition: Classifying observations based on attributes across two or more variables.
Example: Classifying opinions on capital punishment and abortion into four categories.
Newspaper Ideology Typology: Categorizes newspapers based on their foreign and domestic policies.